Background Guide to Setting Source Options

Why am I writing this guide?
Simple! There has been a great deal of confusion about the settings and ‘calibrations’ of source devices. Well-intentioned users commonly misunderstand the functions of some adjustments, and end up unintentionally degrading their image quality. This FAQ /guide is an attempt to filter through some of the mess and assemble a decent background in one place, so that you can do your best to make sure that your playback chain is not degrading your video unnecessarily.

The layout:
I’ve organized this guide into two parts with background sections first, in the hope that everyone will want to read the background first before going on to adjust settings and calibrate. For you throw-away-the-manual adventurers, you can skip ahead to the adjustment section that applies to your system. The numbers match between the two parts, so if you’ve skipped ahead to the goodies of finding the right setting and it’s confusing, circling back to the background explanation may help!

Say, who’s behind this fiasco of a FAQ? What’s the conspiracy! :)
I’ve written and compiled this material with the help of Guy Kuo, Keith Jack, Stacey Spears, Joe Murphy Jr, Bob Pariseau, and the indirect help of many other knowledgeable folks, posts, books, alternate personalities etc. Thanks guys! Much of this background has already been discussed on the forums, but is scattered and obscured by popular misinformation. My hope is to unite a solid background that is sorely lacking.

Wait a minute, you’re wrong about _____!
I’ve taken great care to make sure the information is trustworthy by having some of the biggest video super-heroes check to make sure the content is solid, so you might actually find that if you read through the background you’ll discover that what you read in the past may not have been as correct as you thought. But if I have left something out or made a mistake, let me know and I’ll fix it! I have used some simplified terms, for instance I label non-linear R’G’B’ simply RGB, etc, for simplicity and clarity: this is not likely to cause any confusion since I can’t fathom any HT environment dealing in any non-corrected linear RGB values!

What tools do I need?
If you are a Home Theater enthusiast, I should certainly hope you already own a copy of Avia or Digital Video Essentials. If you’re a nut like me you own both, and perhaps also Avia PRO! For testing Studio versus PC levels Avia PRO or DVE are necessary.


1)Analog input/outputs from digital sources
2) Digital outputs and digital levels
3) Upconverting players and possible colorspace issues
4) HTPCs


BACKGROUND

1)Analog input/outputs from digital sources

Say, what the heck is an IRE?
As often as the term IRE is thrown around when discussing video, relatively few actually know what it means!
An IRE simply a representation of volts: there are with 140 IRE units in one volt. This simplified expressing the ~700mV excursion of an analog video waveform, with white represented using 100 IRE instead of an odd 714mV. As I will describe, IRE units are ambiguous at describing the intended image content unless you know whether or not the 7.5 IRE setup pedestal is present. Many do not grasp that IRE is merely another way to represent volts, and as such confuse IRE into some mysterious value that magically describes the actual image information, which cannot do without caveats.

North American NTSC standards historically used a level for the analog output of black that was 53.6mV (aka 7.5 IRE) above the blanking voltage of 0mV (0 IRE), and we still do. Other standards like HD, Japanese NTSC, etc use 0mV to output black, and lack the 53.6mV (7.5 IRE) ‘setup pedestal’ that North America uses. This IRE setup option allows you to switch between these two voltage output levels for black. Neither will necessarily provide any universal increase in picture quality: it merely alters the voltage output that is used to represent black. With a correctly designed player and a good display, upon recalibrating your display to the new source outputs, your picture would be identical. As we all know the ideal player is rare and displays vary, so in some systems one setting may be preferred over the other to prevent a degraded picture. But neither setting is universally preferred. I’ve explained how to examine the behavior of your specific system to determine which setting(if any) is preferred in the second half of this FAQ.

Test patterns from test DVDs are (unfortunately, in my opinion) often labeled or described in IRE units. These are merely labels, and they usually assume the presence of setup, thus the source outputs black at 7.5 IRE. If there is no setup then the source is outputting black at 0IRE and the labels on the test patterns are now wrong for that system. The absolute IRE levels will change depending on the behavior of the source device, and the DVD has no control over this. These charts accurately show that the encoded digital values that are the stimulus do not change, while the absolute mV outputs (hence IRE values) will change depending on the voltage for black output:

**
For Avia and Avia PRO--
7.5 IRE output:

Labeled Percent Output Output
IRE Stim IRE mV
7.5 0.0 7.5 53.6
10 2.7 10.0 71.4
20 13.5 20.0 142.9
30 24.3 30.0 214.3
40 35.1 40.0 285.7
50 45.9 50.0 357.1
60 56.8 60.0 428.6
70 67.6 70.0 500.0
80 78.4 80.0 571.4
90 89.2 90.0 642.9
100 100.0 100.0 714.3


0 IRE output:

“Players with a 0 IRE representation for black and 100 IRE for 100% white will output IRE levels that don't match the on disc labeling would do the following. Notice that the relationship between labeled IRE and percent stimulus remains the same as for players with 7.5 IRE setup. However, the output IRE and mV change.” -GK


Labeled Percent Output Output
IRE Stim IRE mV
7.5 0.0 0.0 0.0
10 2.7 2.7 19.3
20 13.5 13.5 96.5
30 24.3 24.3 173.7
40 35.1 35.1 251.0
50 45.9 45.9 328.2
60 56.8 56.8 405.4
70 67.6 67.6 482.6
80 78.4 78.4 559.8
90 89.2 89.2 637.1
100 100.0 100.0 714.3


**
(From Guy Kuo)

You can see that in a system that outputs black at 7.5IRE, the labels are accurate. In a system that outputs black at 0IRE, the labels must be recalculated: actual IRE levels when setup is not present = (labeled IRE-7.5) * 100/92.5. You can see that an Avia pattern labeled 10 IRE will end up about being output at a voltage about 2.5IRE (more accurately 2.7 IRE or 19.29mV) in a system without setup. As long as the display is correctly calibrated to the source, the pattern should be imaged correctly.

If that last portion seemed like too many numbers, don’t worry! The most important point you need to take away is that IRE units are merely another way of expressing mV.

It should be obvious already that DVDs are not analog. They contain digitally encoded video. As such, there are no voltages read from a DVD(it’s an optical media!). By extension, there are of course no absolute IRE units represented on a DVD. A DVD uses 8-bit digital levels from 1-254 to represent the image data. Only when that data must be output in an analog form do voltages, hence IRE, even come into play. That’s right: if you are using a digital output, there is no analog video signal, thus IRE units do not come into play in any form (cautious note: the adjustment may unintentionally affect digital output, but if designed correctly it should not.).

Because many have confused IRE units into some mysterious creature of a unit, they’ve confused digital levels and IRE as somehow being the same: i.e. that IRE are encoded on DVDs, or that DVDs only go down to 7.5 IRE while HD has better black capabilities because HD goes down to 0 IRE.

This is rubbish.

Both HD material and DVDs use the same digital levels. It is the ANALOG output standards which can vary. The encoded levels represent black just as black. The dynamic range is the same. Video encodes black at digital 16. Period. That’s black, and whether it is output at 0 IRE or 7.5 IRE the resultant image should be equally black. If there is any difference in the visible levels of black, your video chain is miscalibrated. If you connect sources that output at various voltages, you have to calibrate to EACH of these sources. Users will often connect a DVD player that outputs black at 7.5 IRE, then connect an HD source that outputs black at 0 IRE, discover that the HD source is darker, and mistakenly conclude that HD has improved black capabilities. This discovery of changed blacks is merely that you have different sources sending different signals. You must calibrate your display to each individually.

If you’ve followed this far, then you’ll see that the following types of statements(some repeated often) are complete nonsense:

‘Movie studio A encodes their DVDs at 7.5 IRE, while studio B uses 0 IRE.’
‘Test disc X is at 7.5 IRE while test disc Y goes down to 0 IRE.’
‘You must always switch any DVD player to 0 IRE for: darker blacks/better -contrast/blacker-than-black.’
‘HD can provide better blacks because it goes down to 0 IRE.’

To reiterate, all the previous statements were nonsensical and meaningless. They represent confusion on the facts surrounding analog voltage outputs(expressed in mV or IRE units) as opposed to what is digitally encoded on the DVD.

The first two are essentially the same statement. Again, if you understand that IRE is just an expression of volts, you realize that DVDs don’t have volts on them at all. DVDs are an optical format, read digitally by a laser, not by anything magnetic/electric. Whether a picture from a DVD is output in analog form with black represented at 7.5 IRE or 0 IRE is purely a function of the player. A DVD is encoded in digital levels. Black is encoded at digital 16 and this is fixed. Whether that is then output at 0 IRE or 7.5 IRE is beyond the control of the DVD itself.

The third statement is also incorrect. In an ideal system, as long as you recalibrate between 0 IRE and 7.5 IRE, there will be no change at all in the final image. The final image will have the exact same visible black, white, and contrast in both cases. Using the 0 IRE option does not in any way output “more” data that is magically “below 7.5 IRE.” Again, data is not encoded in IRE at all. Stating that there is somehow “extra data below 7.5 IRE on a DVD—you need to set your player to 0 IRE to retrieve this data” is like saying that your car has better fuel mileage because your speedometer is labeled in MPH instead of Kilometers. The two things are completely unrelated and the statement is, well, total baloney.

The last statement is also wrong. Black is black. If switching between sources you discover that your black levels in the image are different, this means that at least one of your sources is miscalibrated. If individually calibrated to one source at 0IRE and another source at 7.5 IRE, black will appear *identical* in the final image, along with identically bright whites, and contrast.

On to the whole blacker-than-black fiasco!

Once again, this issue has been confused and entangled with IRE. If you’ve read up to now, you hopefully know that IRE units are just a representation of analog volts. The DVD is written using digital levels, not IRE at all. A common wrong statement that I see goes along these lines:


‘Blacker-than-black is data below 7.5 IRE. You have to use the 0 IRE setting to get this ‘superblack’ data.’

WRONG! Blacker-than-black data has nothing to do with whether a source outputs at 0 IRE or 7.5 IRE. You know already that IRE is just mV, and absolute mV are not encoded in any way on the DVD. The DVD is encoded in digital values from 1-254. Reference black is encoded at digital 16, with nominal reference white at 235. Codes 0 and 255 are illegal for image data. Codes 1-15 represent data that is below reference black, hence it is called blacker-than-black (BTB). Codes 236-254 are peak whites. I will refrain from calling them ‘whiter than white’ because this implies that they shouldn’t normally be present or visible in the final picture. They should be, unlike BTB(BTB should not regularly be visible in the final calibrated image)! It is important to maintain the full range of encoded data through your video chain for the best image. When the digital image data is converted into an analog waveform at an analog output, only then does IRE enter the picture. BTB data will simply fall below whatever the IRE output level for black is. In a system that outputs black at 7.5 IRE, BTB data will be output at voltages slightly below 7.5 IRE. In a system that outputs black at 0 IRE, BTB data will be output at voltages slightly below 0mV (simply negative volts). If you’ve digested that correctly, you realize that BTB data can be maintained in BOTH situations.

Unfortunately, in reality DVD players are often designed poorly and may clip BTB data. Some will clip BTB in both settings, others will clip it in one or the other. A correctly designed DVD player will maintain BTB data at analog output regardless of whether it is set to output black at 0 IRE or 7.5 IRE.

“Why does data below black even exist? That makes no sense! What can be blacker than black!?”
In video, headroom and footroom is important for a number of reasons. The most basic is that mastering can be less than perfect, so some “fudge-room” has always existed. However, even with ‘perfect’ mastering, data regularly extends outside reference black/reference white. Peak white data allows for highlight details to be maintained, which is common in clouds and other bright objects. BTB data helps prevent image anomalies from hard clipping of the analog waveform at black when converted to analog. BTB can also sometimes become visible as the actual black level on a CRT display floats up and down with image content because black level retention on CRTs is not perfect. The mastering engineer is viewing on a CRT display and actively changing the encoded levels so that they are imaged correctly on the display. This reverse-float compensation in black is allowed with BTB footroom. This reverse float compensation should be the only times data encoded below black is visible in the final image. If you are using a PLUGE pattern with BTB bars to calibrate, you should calibrate so that the BTB data is not visible. BTB data also helps define dithering duty patterns on DLP projectors. Lastly, BTB and peak white data is quite useful for any image processing/scaling etc applied to an image. For all these potential reasons, video engineers advocate preserving full BTB and peak white data whenever possible. Lastly, this data is helpful for any image processing that is applied on an image, such as sharpening, scaling, etc.

Now you should understand better what an IRE is, what the 7.5 IRE/0 IRE difference does NOT entail, and what blacker-than-black data really is.

For more reading, see posts in these threads:
http://archive.avsforum.com/avs-vb/showthread.php?s=&postid=4030461#post4030461
http://archive.avsforum.com/avs-vb/showthread.php?postid=3307309#post3307309

This thread has some useful attached pictures that label patterns with their actual encoded values in digital levels, and not in ambiguous IRE units. The digital levels written to the disc itself are fixed and absolute even though the analog voltages vary. Knowing the digital values tells you exactly what a portion of the pattern should look like. Parts encoded 16 are unambiguously black, always!:
http://archive.avsforum.com/avs-vb/showthread.php?s=&threadid=418084


2) Digital outputs and digital levels

I explained IRE in background section 1, so you should already recognize that IRE only describes analog waveforms, and is ambiguous as to the actual intended image content. When using digital transmission formats like SDI, DVI, HDMI, analog measurements like IRE units or mV are useless for describing the image data. As briefly explained before, digital video uses 8-bit encoding, which creates 256 unique steps from 0-255.

Video from DVDs or other digital sources, follows Studio RGB standards which encodes reference black at level 16, and nominal reference white at level 235. Outside these bounds is the footroom and headroom that contain blacker-than-black(BTB) and peak white data. Footroom and headroom is necessary for the best possible video reproduction for the reasons discussed briefly in background section 1:

Mastering fudge room
Highlight details and undershoot
Minimizing analog waveform anomalies at D/A converters
Compensation for the inability of CRTs to clamp black levels perfectly
Determining dithering patterns on DLPs
Any image processing applied

Enter the PC.
If you’re a computer graphics person, you may recognize the 0-255 scale of digital levels. But you may be confused because with computer graphics black is at 0, white is at 255, and that’s it! There is no footroom or headroom for BTB or peak white data. Since computers and video often have to be dealt with at the same time, this creates a problem between PC levels and Studio levels.

Projectors with digital inputs have to be able to handle both video applications and computer graphics applications. They should be able to calibrate or switch their white and black levels to accommodate both Studio levels and PC levels, but some don’t. Because video sources and graphics sources may be used simultaneously in some applications, video source manufacturers tend to include an option to leave the digital outputs at Studio levels, or to re-map them to PC levels.

Mapping Studio levels to PC levels can be done a few ways. Sometimes all the levels are just shifted down 16 steps, thus clipping off BTB data, but not introducing banding/contouring or clipping highlight details. If there are other PC level sources fed to the display, their whites will be substantially brighter. The clipping of BTB data is undesirable and the brightness mis-match is also noticeable, however note that there will be no banding problems. Usually the levels are expanded: digital 16 (black) is shifted down to 0, and 235 (white) is shifted up to 255 (or sometimes a value slightly lower than 255) thus expanding the numerical range between black and white to match PC levels. Note that this doesn’t improve contrast in the final image. In this case, re-mapping Studio levels to PC levels will destroy BTB and peak white image data, and introduce banding/contouring artifacts because of the expansion. This is also undesirable.

Always try to maintain Studio levels whenever possible in your system. You should use Avia PRO or DVE to test whether or not levels are being clipped. Both discs contain patterns with both BTB and peak white data. Note that the current consumer Avia does not, it only contains data in the range or 16-235. Thus if BTB or peak white data is being clipped, you won’t see it with Avia! If you are clipping the black or white bars in Avia, then your system is doing even more severe damage to the video by clipping above black shadow details and below-white detail: very bad! Note that Avia PRO also has some very useful ramp patterns with levels encoded at equal width so as to be completely banding-free. These are very useful for observing banding/contouring caused by the playback system.

For more reading, see this thread, especially the expert posts by Don Munsil and Stacey Spears:
http://archive2.avsforum.com/avs-vb/showthread.php?s=&threadid=416292

3) Upconverting players and possible colorspace issues

This issue is still a fairly obscure one, and at this point I am including it only for the sake of being comprehensive. My concern is that people will read this portion about colorspace problems, and blame all color problems on this issue when it may have nothing to do with what they are seeing.

Note that this specific discussion ONLY applies to component video outputs: analog YPbPr or digital HDMI YCbCr.
This issue may arise in other narrow situations, but is quite unlikely.

(I’ve labeled ITU-R Recommendations BT.601 and BT.709 simply SD 601 and HD 709 so readers can follow which is which as they read)

If you are using YPbPr/YCbCr output, you are outputting component video that is actually encoded on the DVD. Your display is then transforming this component video into a form that it can use to drive the display (usually always RGB). The problem that can arise is that Standard Definition material uses one set of equations (SD 601) to move between RGB and YPbPr/YCbCr, while High Definition material uses a slightly different equation (HD709). A DVD is a digital SD source, as such the YCbCr on it was created using the SD 601 equations and should be decoded into RGB using the same SD 601 matrix or color errors will result.
Your display will likely apply either the SD 601 or the HD 709 decoding depending on the resolution it is seeing at input. An upconverting DVD player is sending an HD signal that your TV will respond to by decoding using the HD 709 decoding(usually). Unfortunately, your upconverting DVD player may or may not have properly re-coded the component video into HD 709. If it did not, then you have a mismatch: SD 601 encoded material is being incorrectly decoded with HD 709 equations. This color encode/decode mismatch will lead to color errors.
Note again that this section is included for the sake of being comprehensive. The actual source material is unpredictable, and may or may not be encoded using the proper color matrix. In my opinion, it is worthwhile for enthusiasts and video purists to consider and fix this colorspace problem if present, but because the problem is fairly limited to mis-designed upscaling DVD players at this point, I don’t see it as a concern for the majority of users. Hopefully, manufacturers will design upconverting players correctly to twist the color space into HD 709 before output.
See especially Stacey’s images attached in post 2 of this thread, along with the discussions nearing the end of the thread:
http://archive2.avsforum.com/avs-vb/showthread.php?s=&threadid=486428&perpage=20&pagenumber=1
edit: it seems the attachment of the colorbar error is no longer present in the thread linked, however Ron (dr1394) has kindly generated a better pattern with all the bars in one image which is attached in post 5 of *this* thread.

I am also adding a link that I had not previously included due to ambiguous and confusing image captions at the time which Keith Jack has kindly changed to clarify, which show the effect of 601/709 matrix mixmatches in real images. At the bottom of the page is a series of images you can click through, each has an iteration of a correct image, and then a wrongly decoded image in both iterations, now unambiguously captioned:
http://www.sdesigns.com/public/Support/chromaticity.html


4) HTPCs
Because HTPCs are a huge discussion area, I can’t hope to cover all the possible problems that can arise with all the software, drivers, hardware, settings, etc that can come into play. In any case, I just reiterate that it is important to design your HTPC and settings purposefully so that you can maintain Studio RGB for video playback, and calibrate your display to this and not to the graphics’ PC RGB levels.

See the HTPC Forum (http://www.avsforum.com/avs-vb/forumdisplay.php?s=&daysprune=&forumid=26) ;)




******** ON TO THE SETTINGS! ********

1) For DVD players when using Analog outputs (Composite, S-video, Component video):

When using analog outputs, many DVD players have an option to change the IRE ouput level for black between 7.5 IRE and 0 IRE. There have been many misleading posts suggesting that 0 IRE output universally provides “deeper” blacks, is necessary to maintain blacker-than-black data, increases the picture’s contrast, or otherwise will improve your picture. These recommendations are not correct. To understand why, read the explanations in background #1 that differentiate between analog output voltage levels represented by IRE, and the actual digitally encoded values on the DVD.

Q: “So which IRE output setting should I use, and since it varies between players and systems how do I know what’s right for my system?”
A: There are two main performance considerations to take into account:

a) First is the adjustment range of your display. The most important thing to do is to make sure that your display has enough black level adjustment range(brightness) to accommodate either 0 IRE or 7.5 IRE output for black. Most displays should have plenty of adjustment range to handle both, but if you find that no matter how far you raise or lower your black level setting that your black levels remain either submerged (too dark) or elevated (too bright: grayish) with one of the output settings, make sure you use the other setting that allows the combination of your DVD player’s output and your display’s adjustment range to show black properly. If you find yourself in this situation, keep this setting, and do not continue on to the other considerations, as proper black level in the image is the most important. In most cases, though, you will find that your display will have enough adjustment range to handle both output voltages properly. If both are handled properly, then the next thing to consider in choosing between 0 IRE and 7.5 IRE is blacker-than-black (BTB) data:

b) Next to consider is the preservation of blacker-than-black data. This issue has been confused a lot, so if my recommendations don’t agree with what you’ve been told in the past, please read the explanations behind them (background #1).

Ideally, a DVD player will preserve the full range of digital data contained on a DVD when outputting an analog signal in both the 0IRE setting, and the 7.5 IRE setting. In practice, many times this is not the case. To find out whether a specific player will pass blacker-than-black data, and in which setting, it must be tested. Reviewers sometimes describe whether or not a player passes BTB data, but unless the settings and connection method for this test are revealed, the information is not very useful. Some players will correctly pass BTB data in both settings, while many others will only pass BTB data in one of the settings, and some will clip BTB data in BOTH settings. (Others will perform still worse and clip more than just BTB data, but also clip above-black shadow details; hopefully you do not have a player like this!) To test for BTB data, you must use a test disc that contains such data, namely Avia PRO, or DVE. Note that consumer Avia does not contain BTB data. If you are clipping the black bars in Avia, your system is clipping even more than just BTB data: it is clipping above-black shadow details as well. Using the PLUGE patterns on DVE is the most likely scenario for the average enthusiast, but if you do have access to Avia Pro I recommend using it.

Using DVE, I particularly like Title 12, Chps 13-15. Chp 13 has a 20% center bar that is not blindingly bright when you raise your black levels to see the three bars beside it. If you’re not familiar with DVE, the two pairs of inner bars are above black, the background is black (digital 16) and the outermost pair of bars are below black. If your system is preserving BTB data, you will see all three bars when you raise your black level at your display to observe. If BTB data is being clipped, you will only see the inner two pairs of bars: the outermost below-black bars will appear black and match the background, which is black. Chp 14 is a useful full-range ramp marked by dots at the points for reference white and black. Chp 15 has a ramp in the lower portion of the pattern that runs from slightly above black to below black. Black crosses at the center.

After you’ve tested your DVD player to see whether it passes BTB, and in which setting(s), you should choose one that maintains BTB data. Note again that this is secondary to being able to calibrate black in your system properly in the first step. I hope users will share their player-specific findings on the preservation of BTB data in both of these settings to aid others who have the same player!

If you have a well-designed player that passes BTB data in BOTH the 0IRE and 7.5IRE output modes, then you can use either one. Always note that you have to calibrate to your chosen setting. If you do not recalibrate, your picture will be wrong: the black level output on your DVD player will not match the display’s calibration which leads to elevated or submerged blacks in the final image.

Extra: The last consideration is only if you have a player that correctly passes BTB data in both settings AND your display can be calibrated properly to either output: convenience. If you have enough inputs on your display, and per-input calibration adjustments, then you can go ahead and calibrate each input to a particular source, and it won’t matter which setting you choose. However, many times multiple sources are run through switchers, receivers, etc, or there are no per-input calibrations on your display. If you have sources conforming to different IRE standards, you would need to recalibrate your display each time you switched between them. This is annoying! You may want to set your DVD player to match the other sources more closely, which should eliminate the need to recalibrate between sources. While sources always will vary slightly in analog outputs and it is ideal to have individual calibrations for each source, when using a switcher, etc, this is often not possible. In my opinion, these slight differences are a worthy sacrifice for convenience, as few will want to recalibrate between different sources at 7.5 and 0IRE each time, and not recalibrating in this situation will result in very incorrect black levels and a significantly degraded picture for the odd source that is using a different output level. Where possible, always calibrate for each individual source, as voltage output precision isn’t always perfect! Where not possible, using the same output standards can be an acceptable compromise instead of needing to recalibrate the display each time you switch sources. How acceptable a compromise this is depends on how accurately the sources hit the standard voltages. If they vary widely enough as to be objectionable with a single calibration, you may just need to live with adjusting your display calibrations frequently.


2) For DVD players when using a digital output (DVI, HDMI).

When using digital outputs, your primary concern is to get the DVD’s image data output as undisturbed as possible. As simple as this might seem, design/price constraints, sloppy engineering, and ‘features’ can get in the way of transferring the digital data from the DVD intact. Always try to avoid using any image-altering features such as picture controls. These adjustments will usually have negative impacts on picture quality.

When using digital outputs, the major adjustment option you have is the one for digital levels. As is common in consumer labeling, the labels for this can be confusing. The most common labeling will at least hopefully show that you are making an adjustment to the DVI/HDMI digital outputs, and usually the options will read ‘Normal/Enhanced’ or ‘Normal/Expanded’ or ‘Video/PC,’ or some such label. The last pair of labels is clearer, as this adjustment is choosing whether or not the digital image data is correctly output using Studio (also called Video) levels, or is incorrectly re-mapped to PC levels. You should choose to maintain Studio levels by checking to make sure this option is properly set, usually by selecting the equivalent of "video" or "normal." Usually the default setting will correctly choose the option for Studio levels, but not always. Check to make sure.

Note that the Studio/PC level option will only work for DVI and HDMI RGB. If you are using HDMI YCbCr, Studio levels should be preserved correctly(and as far as I’m aware sources aren’t screwing this up, yet…) and the option change will either have no effect or be unavailable.

Because you are using digital outputs, adjustments for analog outputs shouldn’t be a consideration. Unfortunately again, due to cost-saving designs sometimes analog output adjustments, such as those for IRE setup, are implemented digitally even though they have nothing to do with digital outputs. Here your concern is still to ensure that the digital image data is being left as undisturbed as possible. In a well-designed player, the IRE option will have no effect whatsoever when using digital outputs. If this setting causes any change in the image when using the digital outputs, you should use test patterns to see which option leaves data un-clipped. In this instance, the degree of clipping or image alterations may be severe, so Avia, DVE, or any good test disc will come in handy. In Avia, look for the moving near-black and near-white white bars; in DVE use the ramps (Title12:Chp14) and check for clipping (if you have access to Avia PRO, the Deep Ramps are excellent tests!). Use the IRE setup setting that maintains as much data as possible through to your display. The default setting may more likely be the preferred setting. Default is usually the 7.5IRE setting. This was reported (http://www.avsforum.com/avs-vb/showthread.php?s=&threadid=477052) to be the case on the Denon 3910; users should test this on their players and share model-specific observations to aid others!

With the hope that digital levels and IRE are conceptually disentangled, you should understand that the IRE setup option is really for the analog output, and shouldn’t have to be discussed when using digital outputs. Unfortunately, instead of implementing these analog adjustments in the analog domain, cost-conscious designs seem to be altering the digital values, which necessitates double-checking this option to make sure your digital levels are not being tampered with.

3) Colorspace issues with upconverting DVD players

This is written only for those who are using component video, either analog YPbPr or digital HDMI YCbCr output. If you are using analog RGB, DVI(which is digital RGB)or HDMI RGB output this section does not apply to you! I omitted composite and s-video because of course you cannot output HD video via either.

A DVD is a digital SD source, as such the YCbCr on it was created using the SD 601 equations and should be decoded using the same matrix or color errors will result. Upconverting players should be modifying the colorspace to match the HD 709 coding, but some do not. Most displays will simply apply SD 601 decoding to any SD inputs, and HD 709 decoding to any HD inputs. If an upconverting player left the SD 601 colorspace encode intact, a display that applies HD 709 decoding to any HD inputs will unknowingly apply the wrong decode, leading to color errors. I place the blame with the DVD player.

You can check for this color problem by examining the color decoder test pattern on Avia(found under the section Special Tests: Color Decoder Check). If the player is upconverting to an HD resolution and not twisting the colorspace into 709, you will likely see color decoding errors in the Avia test pattern, with the most significant one being green depression(note that this is not a grayscale error but a color decoding error, and will not depend on you having correct grayscale: you will see this problem on the color decoder check pattern regardless!). To avoid these color decoding errors, your display may allow you to change between 709 and 601 manually. If it does not, using HDMI RGB or DVI RGB (analog RGB is a rarity, but also would avoid this problem) should eliminate this problem. If your only output option is analog component video YPbPr and your display does not have its own upscaling capabilities, then you may have to live with these color errors.

4) HTPCs

If you are using an HTPC, I will dodge the complex arena that is computer video, and merely recommend strongly that you do your best to maintain Studio levels and prevent your computer from expanding the video to PC levels. To check this, examine deep ramps on Avia PRO, or use the ramps and PLUGE patterns on DVE. Note that consumer Avia does not contain data outside the bounds of reference black/white (16-235) so you will not be able to observe the clipping of data outside these bounds using Avia. Moving to PC levels will usually clip data 1-15 and 235-255(or significant portions of this data); this data is present in the ramps and PLUGE patterns on DVE and in many patterns on Avia PRO.

Remember that proper playback of video requires Studio levels to maintain the full range of image data. This includes data outside the ‘bounds’ of the reference black/white points. Achieving this on a PC can be difficult sometimes, and it also means that your desktop and your video cannot both look correct at the same time. Expanding your video to match PC levels (0-255) will make your desktop and your video ‘agree’ and eliminate the need to recalibrate, but you’ve also negatively impacted video playback by clipping useful image data and introducing banding/contouring artifacts. This is undesirable, you should strive to maintain Studio levels if at all possible. Since this guide is directed at HT enthusiasts, I am assuming that accurate video playback is priority #1. I acknowledge that for users in other environments(multi-use, graphics etc) the problems introduced by expanding to PC levels might be an acceptable compromise for convenience. But wherever you strive for the best possible video quality, Studio levels are fundamental.

CZ

Edited 08/15/06: Fixed outdated links, should now accurately link to the threads many of which are now archived. Should they break again, let me know and I'll try to fix them as quickly as possible. Also, you can try archive.avsforum... or archive2.avsforum... and the rest of the URL and you should be able to find the intended thread. I have also added a link to a sigma designs site in section 3 that illustrates very well the colorspace decode issue between 601/709 in real images.

Edit 1/4/06: Slight language change for clarity in settings section 2.

Edit 8/13/09: Fixed Sigma Designs link.

Outstanding. Chris, ask the mods if they'd consider making this thread a sticky. :)

Thanks Bandit!

FYI:
I've decided to leave this thread open so that users can discuss and help each other with the material presented and what might be confusing for some at first.

If anyone feels something is *factually* incorrect, or needs elaboration I'd prefer you pm me about it and we can discuss, as I've tried to be very careful to be factually solid on this post. I Don't want to be defending this material constantly.
Thanks for the consideration!

As for how to calibrate, that is a long discussion in and of itself that I didn't include, if someone would like to add substantially to that area for those who may not be as initiated, let me know!

--Chris

Excellent post Chris. Definitely sticky-worthy.

I've created a color bars image that's a little easier to view than Stacey's,
with all four YCbCr to RGB conversion permutations in one frame.

Ron

Superb guide.. The one thing I have a problem with is digital outputs to digital projectors.

We all know that contrast is one issue with digital projectors that we like to improve upon. My feeling is that by preserving blacker than black signals you have to raise 'base' black level above lowest black that the projector can create (there is no way out of this logically)..

So in order to preserve a signal on some discs (badly authored discs IMHO) that present a final signal in the 0 - 7.5 IRE (0 - 16) range you end up watching the rest of your entire collection at a higher than lowest black signal. With this higher than best black you in turn have a drastically reduced real contrast ratio. It strikes me as odd that people will pay 1000's more for projector A over projector B due to improved CR and then not use it to its full abilities.

I dont see either choice as being necessarily 'wrong' (but I guess a ISF tech would have to have a favorite) merely one of those compromises you have to choose between. Either you have a system that displayed the best CR and black level on most of your titles (how I tune my setup) and miss some below black info, or you have a system that retains all the signal and displays below black info for a minority of titles but has a higher black level displayed for many titles mastered at 7.5 / 16 = black.

I dont know why but I have less of an issue with the white level side of the system range or clipping (probably as I have a fairly bright projector). Thinking about it I guess that people with a non dedicated room (perhaps people that allow some light in thier theater or have light walls) may mind the issue less as the raised black level may be less percievable in that kind of setup.

Perhaps there is something wrong in my thinking of this issue as I have always had this 'problem' over 0 = black or 16 = black while others have tried to explain to me that it is simple and I am wrong.. Any corrections to miss understood ideas welcome.

My feeling is that by preserving blacker than black signals you have to raise 'base' black level above lowest black that the projector can create (there is no way out of this logically)..

"Preserving" BTB does not mean that you calibrate such that it is viewable. Be clear that you do not raise your black level calibration(thus reducing contrast) to see the BTB material. This is wrong. When you calibrate, you calibrate black (16) to be black on the display. The BTB data will *not* be normally visible. Raising the black level to test whether your system is maintaining BTB data to the display is only temporary to see whether the BTB data is being clipped or not. After checking (and making changes if needed) then you re-calibrate, and correctly calibrated you will *not* see the BTB data.

final signal in the 0 - 7.5 IRE (0 - 16) range

These are not synonomous things at all. They have no direct relation to each other.

If you read through my explanations (I realize they are long and laborious, but there's really no painless way to try to completely re-explain things like this) you'll see that I am very clear in explaining that the ANALOG voltages (IRE) as being *totally* distinct from the digital levels.

If you read through carefully, you'll see why your statement arises from the confusion of IRE and digital levels as being intertwined:

or you have a system that retains all the signal and displays below black info for a minority of titles but has a higher black level displayed for many titles mastered at 7.5 / 16 = black.

Expressely stated in the background explanations are rebuttals for statements like these. DVDs are not mastered in voltages(like 7.5 IRE or 0 IRE or whatever). It's impossible. There are no volts on a DVD at all. Black is defined and standardized at digital level 16 for DVD, PERIOD. *ALL* dvds place black at 16 if they are mastered correctly.

Agreed I am competely miss using IRE (7.5 for black or 0 for black) incorrectly.. Really its a case of struggling to find terms for 'slightly higher shade of grey than absolute black' etc. Lack of terminology and being lazy in my descriptions.

I did read it all once.. Will go over again and see if I can get my head round it.. Still dont understand what a plunge or blacker than black signal is doing (or what use it has) at output if you clip it at display ?? Either you should see it and it should be in the outputted image or it should be clipped out (how I have worked).. I can understand possible uses for it in the signal processing side but not at oputput.

I think its comments like highlight room and footroom that confuse me because surely these are counter to the idea that all correctly mastered DVD's stop all (outputted) image detail at studio black (16) and studio white (235).. If so what is this issue of footroom ?? That should not be seen in the final image (IMO) ?? Why would peak white (above 235) ever be outputted ?? To show 'cloud detail' just seems odd to my ears.

Anway will read and digest.

Phat Phreddy,
This is the nub of people's confusion on this, along with the idea that the 0 or 7.5 IRE choice has something -- anything -- do to with the way the producer intended the material to be seen.

Specifically, the Blacker than Black data is *NOT* intended to be seen. If your display is calibrated correctly (and its Black levels don't float around) you won't see it. It will all merge into one, perfect, identical, absolute "Black".

Yet it's still a good idea to have it there. Why? Because it "protects" the Black and Above data in the presence of additional signal processing that might happen before the pixels light up, and because it preserves the actual dynamic range that was used when the producer originally recorded the data. The sorts of signal processing that might happen include things like scaling, and gamma correction, plus, plus, plus. It's also good to have it JUST IN CASE your display has Black levels that float a bit depending upon what's currently being displayed.

Now the video stream is necessarily an incomplete representation of a continuous real world image. The producer's choice of what source light level he wants to identify as "Black" is essentially arbitrary. It is a reference point based on the low light level details he wants to remain visible, and is used as a basis to insure that quality is maintained throughout the reproduction chain for that light level and above.

But a video signal goes through the signal processing mangle many many times before your pixels light up. And each time that happens there is the chance that artifacts will be introduced because the data has been arbitrarily clipped -- has a hard filter clamped on it -- cutting off anything below what the producer chose to consider as "Black". So to protect against that happening, the video data includes information BELOW the light level the producer arbitrarily chose to represent as "Black". Such Blacker than Black data should stay in the stream all the way through to when your pixels light up. Again, if things are set up right it won't become visible. But it's presence in the video stream preserves the quality of the Black and Above data through any final signal processing that occurs.

--------------------------------------------------------------------------

Meanwhile the 0 vs 7.5 IRE choice is something the DVD content producer has no part in. He doesn't care. It's a choice that only matters for an analog video stream, and thus only matters between two devices that are connected with analog video cables. All the producer cares about is that his DIGITAL video stream will be reproduced with complete fidelity REGARDLESS of whether two analog video devices choose to use a 0 or 7.5 IRE voltage to represent "Black".

With properly working video devices, the digital data from the DVD can be represented IDENTICALLY well with either the 0 or 7.5 IRE choice for analog hookups between any two such devices, and the resulting image will be IDENTICAL in appearance. I.e., with good hardware, the choice to use 0 or 7.5 IRE levels for any particular analog video hookup is completely arbitrary and won't alter the resulting image as long as both devices agree -- which means if you make a change on one device you have to adjust the OTHER device AS WELL to keep them in agreement, i.e., you need to RE-calibrate.

But not all analog video hardware works that well.

So for example, your hardware might work best with a 0 IRE (voltage) analog video hookup between the DVD player and an external scaler but with a 7.5 IRE (voltage) analog video hookup between the external scaler and your projector. It doesn't matter. As long as the full range of the signal is properly preserved in both hookups -- including the Blacker than Black data -- the picture from your projector will be correct.

The confusion here arises because SOME devices will only work properly with one analog voltage setting and not the other, and people have got it in their heads that that means the video signal CAN'T be represented completely correctly by ANY analog video hookups except by always using just one of those two choices.

It is further complicated by the fact that some devices prefer one setting or the other based on things you would think wouldn't matter, such as whether you decide to use S-video vs. Component cables, or whether you decide to pass a 480i signal vs. a 480p signal. And choosing a voltage level your particular hardware doesn't like can have unexpected side effects as well -- such as clipping off the Blacker than Black data which is properly passed FOR THAT HARDWARE only if the other voltage level is used for analog hookups.

So you have to check your calibration levels according to each hookup choice you make and you have to make sure the Blacker than Black data (and Peak White data) is getting through.
--Bob

Still dont understand what a plunge or blacker than black signal is doing (or what use it has) at output if you clip it at display ??

This part is explained in the guide, and is also explained in the linked thread about BTB data by Stacey Spears and Don Munsil. They haven't really said anything in that thread that I didn't include, though.

Second, it's not "clipped" at the display. Rather it is calibrated so as to be not directly visible. It's affects are subtle, and used for processing, black level float, DLP dithering, etc all the reasons I explained.

surely these are counter to the idea that all correctly mastered DVD's stop all (outputted) image detail at studio black (16) and studio white (235)

You have the wrong idea. Correctly mastered DVDs still use this footroom and headroom. Some DVDs are clamped at some point in the master and may clip headroom/footroom anyway (this isn't good). In any case, the bits certainly weren't intended to be wasted or video would just follow PC standards and use all 8-bits of the range not allowing for any footroom or headroom. This isn't the way video was designed, and for good reason. If everything was hard-limited to 16-235 with no data beyond that, that would be a strange way to waste precious bits.

All this theory is giving me a headache!

Has anyone got, or ever seen, a real world example of a DVD that actually looks different when viewed with blacker than black being passed, but not displayed, AND blacker than black not being passed at all?

(ie Identical setup in every other respect)

Bill

btiltman,
The problem is that the artifacts are different depending upon the type of processing that happens AFTER the BTB data is clipped, and if the data is clipped inside the DVD player that means the issue is also significantly dependent upon exactly where it happens -- e.g., at the video output stage, or earlier in the processing.

Some folks with "simple" projector systems (little or no signal processing) that are made well enough to have no "floating" of black levels, and who are using a DVD player that clips BTB data right at the video output stage, see no problem when BTB is clipped in their video -- once they've got black levels properly calibrated of course. There's not enough processing happening after BTB is clipped to make a difference.

But the most common symptom that people see when BTB is clipped is "black crush" to varying degrees. This is a feeling that low black detail is being lost no matter how careful you are at trying to set black levels properly. This is due to two things. First, if the black levels on your TV float up and down depending on what's being displayed (usually based on the average brightness of the image) then when the BTB range temporarily floats up to become visible there's no detail there -- it's all been rounded to one value. Second, if the image processing in your TV is sensitive to (i.e., depends upon or takes advantage of) the BTB data, then near-black details ABOVE "Black" may get incorrectly rounded to the same level, and/or the absolute "Black" level may get rounded up a bit yielding dark gray instead of black on screen.

So the most common reported symptom associated with improper clipping of BTB data is loss of detail in low light scenes on various DVDs. Usually if you've got this problem, it will be seen in most ALL of your DVDs that feature detail-rich, low light scenes.

Now "black crush" can come from other errors in processing and calibration than just improper clipping of Blacker than Black data. But if you've got visible "black crush", checking for proper passing of BTB data is one of the first things to do.

The second most common symptom of BTB clipping is "noise" in low light level scenes. Again this is due to signal processing that happens after the BTB data is clipped. But in this case, instead of improperly rounding true detail to the same level (thus blurring the detail), random variations are introduced in that detail. Think of it as a type of "ringing" as the signal processing tries to deal with the fact that the light sampling has a sharp cutoff at "Black" -- something that wouldn't occur in "real" images.

The commonality of this symptom is harder to quantify since the more usual problem of people seeing noise in low light detail is that they've mistakenly mis-calibrated their black levels to make BTB data visible when it *IS* present. BTB data is inherently lower quality since the entire video reproduction chain is geared primarily at preserving the fidelity of Black and Above data. So if you make the common newbie mistake of, "Keep turning up the Brightness control until new detail stops appearing," then you will have made the BTB data visible when it shouldn't be and it will be more noisy.

[In addition, DVD players that exhibit the Macroblocking Bug, will typically exhibit this bug as what looks very much like noise in low light level areas of the image -- particularly in "backgrounds".]

So folks who have made this newbie mistake might find that their mis-calibrated image (with BTB improperly visible because black levels are set too high) becomes LESS noisy if they mis-adjust their player or TV to CLIP the BTB data!

Given all this, it's much harder to pin down just how often noisy blacks really are related to improper clipping of BTB in modern equipment. And when it really does occur, it is likely to be very dependent on just what you are watching -- i.e., less common across different DVDs than the "black crush" symptom.

Finally there's the issue of PC vs. video style digital encoding. One way that BTB data gets clipped is if the video source device sends out PC style digital video over a DVI connection. PC style digital video encodes "Black" as digital 0, so there's no room to send BTB data below that, whereas video style digital signals encode "Black" as digital 16 and (usually) pass BTB correctly in the steps below 16.

Some older TVs and projectors only know how to handle a PC-style digital video signal via their DVI inputs because they were designed expecting that the DVI input would be used only to hook a computer to the display. And thus DVD players with digital outputs usually have some sort of user option somewhere to format the output signal that way so they can still be used with these older displays. Since such TVs and projectors aren't EXPECTING BTB data on their digital inputs, they are usually careful to avoid the most serious sorts of problems that might occur if they used signal processing that was designed to take advantage of BTB data.

But you still aren't out of the woods as regards image defects. Since the data range for the DVD video stream has been EXPANDED (at both ends) when converted to the PC style encoding, the original digital steps have to be increased in number. This means the player and TV need to decide how to extrapolate the additional steps. Often this is not done very cleverly and the result is false gray scale transitions -- steps that are too wide in one place and too narrow in another. And THAT translates into false contouring -- which could happen all at one end of the gray scale or scattered throughout the gray scale. This is not an issue with clipping of BTB per se, but if you test your video and confirm that BTB data is properly present, you can be sure you haven't accidentally set your system up to use PC-style digital video encoding with a TV that properly handles video-style encoding. And that's a good thing.
--Bob

Has anyone got, or ever seen, a real world example of a DVD that actually looks different when viewed with blacker than black being passed, but not displayed, AND blacker than black not being passed at all?

Bill, please heed my request to keep debates about this off this thread. If it continues it will have to be left as a closed-thread sticky, as this can be debated elsewhere (and has been already at length in the linked thread that I provide in my guide: http://www.avsforum.com/avs-vb/showthread.php?s=&threadid=416292).

Joe Kane, Guy Kuo, Stacey Spears etc are all video experts and demand that Studio levels for video, and BTB and peak white data be maintained correctly. Joe Kane is especially stringent on this. You can feel free to disagree with them elsewhere(a losing proposition) if you'd like, just not here please.

Thank you,
--Chris

Originally posted by ChrisWiggles
Bill, please heed my request to keep debates about this off this thread.
Thank you,
--Chris

Hi Chris,

I have no strong feeling either way on this issue so I am somewhat confused as to your response. I was asking a simple question as to whether the effect was demonstratable and, if so, what it might look like. Perhaps you misunderstood?

Bob,

Thanks for your response, that explanation really answers my question almost as well as an actual example. I appreciate your effort, thanks!

Bill

Chris, your detailed explanations have shown me what IRE really is. Thank you for taking the time to bring the rest of us up the AV learning curve!

I have some questions...

Question 1. How best to setup a TV for non-DVD sources, like an OTA HD tuner?
I have been using DVE through my DVD player, while still at the factory default settings, to set the TV's picture settings. Some digital video, mostly SD but some HD, looks very right in terms of color, while some HD seems to stray from a familiar color palette. I realize much may be deliberate -- on my set Cold Case has a blue bias, CSI has a yellow one. Others look like something's wrong -- PBS's European Travels and occasional HDNet demo channel programs seem to have a slight magenta tinge on my set. This leads to question the current calibration. How do I know what the "right" settings are for the OTA box? To do any better is specialized calibration equipment required?


Question 2 When is gamma correction a viable step to further refine black and white levels?
I had difficulty understanding how best to set up my Sony GWII (LCD RPTV, which has an inherent limitation in contrast) with my Denon 2900 via component connects. At first, I set brightness as high as possible without seeing any color bias in a 100% white area (my set's white goes cyan at the highest levels). Then, I set black using the DVE three bar pattern such that the BTB bar was slightly visible below the surrounding field. This seemed to give me crushed whites -- highlight areas appeared unnaturally large and really stood out. That was clearly wrong.

The white crush drove me to use the DVE ramp to ensure that I could discern a difference between the highest and lowest bars in the pattern. That made for a better picture. However, my recollection is that the DVE gray ramp appears asymmetric through my system. That is, the amount of "crush" is different at the lower end and higher ends of the scale. For now, there's a bit of crush at the black end.

My DVD player does have a gamma correction curve. There are ~16 points to adjust along a graph that I think shows output vs. input values. Is messing with gamma a viable next step or did I miss something fundamental with setting the black and brightness levels?

Excellent questions, let's see if we can point you in the right directions:

Q 1) OTA stuff is difficult, because the channels can (and will) vary all over the map. As long as you've calibrated for that input type by using patterns in your box, or a signal generator, or (tentatively) a DVD player that's set to the same output levels as your box (make sure you know what you're doing here!), then that's sort of the best you can hope for. If some channels look really off, it might just be their fault, and some individual tweaking is ok.

Q 2) For setting black level and white level with LCD, look for clipping and colorshift, as you mentioned. First, the BTB bar will be calibrated to be *NOT* visible, it will look as black as the black background, and you will still see the above black bars when using the DVE pluge patterns. When using Avia, calibrate so that the above-black bars are just barely visible above black. Avia Pro also provides a BTB bar which can be handy.

For whites, you should be looking at not just a 100% white screen, because then you cannot look for clipping as well. A DLP or LCD may clip data at high white levels, and it may also colorshift. You should be just below whichever of these things occurs first. Black and white level settings may interact a little bit, so you may have to go back and forth between them to arrive at the correct setting.

>>>All this theory is giving me a headache!<<<

No theory here -- just fact. Disagreement regarding below black and above white signal preservation is a losing proposition, as Chris has already stated. Perhaps those of us who contributed to this FAQ should have been a bit more emphatic about not questioning this "set-in-stone" fact, indeed the most important part and heart of this work. Please accept my apologies for thinking that links to discussions, information and explanations by industry leaders would prevent this.


If you'll accept a rather crude analogy of the movie experience and the affects of passing/clipping black levels, maybe this will help. Let's liken the experience to a wrestling match. In one corner, representing the actual image recorded on film, we have Andre the Giant. In the other corner, representing the image on a DVD, we have Andy Kaufman. To illustrate the difference in displaying the complete signal vs the compromised signal, we'll use two different wrestling rings. My ring first.

Andre grabs Andy by the arm and swings him to the side of the ring. Andy goes flying toward the ropes and, arms flailing, meets up with his rubberized friends. Only he doesn't stop when they gently couple him, as the ropes have give and flexibility and allow him to stretch their welcome a few feet. Andy is spared the pain and is allowed to return to Andre in the middle of the ring. Gladly, Andre grabs Andy's other arm and sends him reeling to the other side of the ring where, again, he stretches the ropes and saves himself from any real pain. Now, your ring.

The Giant swings his flippant instigator toward the ropes at an amazing rate of speed. But just before Andy gets to the edge of the ring (a padded wall, instead of rubberized ropes), he cringes and tries desperately to stop his impending doom. Nonetheless, he slams into the wall with a great thud and, although there is a bit of padding, he feels the impact of the bricks on the other side of that wall. Not quite getting the message, he challenges the human building and runs toward the champ. Again his arm is grabbed and he is flung in the other direction to meet his earlier fate: lots of cringing and screaming before that loud thud.

Now, while watching Andy Kaufman get slammed into a wall may seem like a ideal wrestling event, I'm sure you'll agree that the ringed mat will make for a more interesting and enjoyable match. Like I said, crude, but it should get the point across.


Since you asked for an example, try this one. If you have a display that is capable of displaying below black signals, here's an easy, yet again crude, way to see what happens when below black signals are clipped. Use a DVD player that you know passes below black signals in one mode (as in progressive and interlaced), but not in the other. Use a calibration disc (Digital Video Essentials) and calibrate the display in the mode that correctly passes below black using the PLUGE pattern. Now play the scene in Jurassic Park III where the T-Rex and "bigger/newer-saurus" dinosaur get into a scuff. Watch the scene in the mode that passes below black and then watch it in the mode that clips below black video (the player should be able to switch between interlaced and progressive modes on the fly for a more direct comparison). You should easily see more details and better saturation when in the below black mode vs the clipped black mode, which will seem grayed out and cloudy/pasty looking -- regardless of whether the mode passing below black is progressive or interlaced.

The above experiment (though not necessarily that scene) should cement ones opinion that clipping black is definitely undesirable into a fact.

I was not 'questioning' any fact. My system is already passing blacker than black etc. and I agree entirely with what is being presented.

I merely wanted an example scene to play with that will readily demonstrate the difference as not all DVDs will benefit from the adjustment.

I will do as you suggest, except I will calibrate black for each mode separately and compare. This is purely for my own education.

I am really finding the tone of this thread to be very aggressive and will no longer refer to it. I think you should make this a closed thread as it does not encourage participation anyway.

Chris, thanks for your PM apology.

Probably catch you all around the traps.

Bill

I merely wanted an example scene to play with that will readily demonstrate the difference as not all DVDs will benefit from the adjustment.

Here's an image I created from "The Talented Mr. Ripley". I've changed all
the below black pixels to white, so that they can be easily identified.

http://archive.avsforum.com/avs-vb/attachment.php?postid=2765749&fullpage=1

Ron

Very interesting...

If I am still stuck miss understanding then just say so but if a display is perfectly calibrated would all those values not be outside of the dynamic range of the display.. Eg all that below black would be 'clipped' at the display ??

Thats why I dont understand the advantage of below black in the outputted signal.. If I am supposed to tune black so that below black is not visible this information is then clipped at display ??

I should point out that my HT experience has been solely with HTPC (and projectors) I have never even owned a DVD player etc.. My HTPC output a full 0 - 255 signal with below black data. I then tune at the display to use the dynamic range of 16 - 235.. Perhaps this lack of experience with CRT based systems or systems and sources that do clip is where / why I am having such a hard time understanding. That said I still have a problem with the following...

Say I have my setup as is.. Outputting full signal with below black and peak white with the display adjusted so that below black is not visible etc.

Then say there was a secondary system that had the 16 - 235 data and outputted this (note I say output only) using full 0 - 255 range. the display would then be calibrated to use the full dynamic range again so that below black is again not visible and peak white is again not visible.

To my mind both of those outputs should look exatly the same provided that high quality signal processing is applied and only the output to the display is effected. I understand that the signal processing if poorly implemented could introduce banding or incorect hard cutoffs (leaving muddy undefined areas) but lets just assume that signal processing IS done correctly and this is merely a function of output signal, if the display can accept PC gamma and video gamma and both can be adjusted to display the correct range fully I dont logically see the difference.

Phat Phreddy,
Yes, if everything is working right the BTB data will be invisible -- i.e. it will look just the same as "Black". It is detail that has been recorded BELOW the light level that the DVD producer intended you to see.

But some TVs have "floating" black levels which vary according to what's actually being displayed at the moment. If you have such a TV and the BTB data floats up into the range of visibility despite your care in setting black levels properly, then your image will be damaged if the BTB data was clipped -- i.e., all rounded to the same value -- before it got to your TV. The portions that now float up into visibility will have no detail in them any longer.

And other TVs do signal processing (such as scaling or gamma correction) which takes advantage of the proper presence of BTB data in the image to do a better job adjusting the Black and Above data just above it. If the BTB data is improperly clipped -- i.e., all rounded to the same value -- these algorithms will introduce artifacts in the Black and Above data that you are SUPPOSED to see. What I think you are missing is that there are different ways to accomplish certain types of signal processing. Algorithms that know they can depend upon the proper presence of BTB data are capable of doing a better job -- but only if that data is really there!

Read any writeup on digital filtering for example and you will find a discussion of the various bad things that can happen (often called "ringing" or "aliasing") if you put an artificial hard filter on data that's supposed to represent a continuous, real-world event. What the film or video camera sees doesn't stop all of a sudden at what the cinematographer or film digitizing editor decides to identify as "Black". So you record some light levels BELOW that arbitrarily identified "Black" AS WELL just to help protect the Black and Above data in the face of various forms of subsequent processing.
--Bob

Fine, thats my understanding also (though I had believed some argued that below black should be visible.. That perception was wrong)..

However.. Given the previous example of Jurrasic park (I niether own this particular DVD or anyway of presenting a signal that does not pass below black) and given, for arguements sake, a perfectly clear display device that does not float, does not do any scaling, does not postprocess the image, pretend we have the perfect transparent display :) !!

Why will it be that if both inputs (the below black passing and below black clipping) are set up correctly to have black at 0 light and below black clipped, then why will one we have the following ?? You should easily see more details and better saturation when in the below black mode vs the clipped black mode, which will seem grayed out and cloudy/pasty looking

Given we have an equal display, once both input are calibrated (and lets assume source performs scaling correctly and only 'clips' at the very output stage) to our imaginary perfect display (that performs no scaling) what causes this ??

I realize that having perfect scaling and perfectly transparent displays is unrealistic but I am trying to seperate out the below black issue from any other effects.

Phat Phreddy,
If you have such an idealized display that makes no attempt to use BTB data in any way whatsoever then of course the BTB data won't be used and its presence or absence is irrelevant. This is tautology.

Real world displays for home theater usually don't work that way. Think of how you would implement a Brightness control on a display, for example, that wouldn't care whether adjacent pixels were black, vs. one step below black, vs. many steps below black. And how does gamma correction interact with that. Or automatic gain controls/limiters. And what of the processing that adjusts pixel fire timing in a plasma for example or other specialized, display technology specific processing. In the end, real world displays are physical devices and so what appears in any given pixel is partially a function of the signal around that pixel.
--Bob

Phreddy, again, please don't let this argument lead to thread closure.

Your questions have already been adequately addressed in my guide:

“Why does data below black even exist? That makes no sense! What can be blacker than black!?”
In video, headroom and footroom is important for a number of reasons. The most basic is that mastering can be less than perfect, so some “fudge-room” has always existed. However, even with ‘perfect’ mastering, data regularly extends outside reference black/reference white. Peak white data allows for highlight details to be maintained, which is common in clouds and other bright objects. BTB data helps prevent image anomalies from hard clipping of the analog waveform at black when converted to analog. BTB can also sometimes become visible as the actual black level on a CRT display floats up and down with image content because black level retention on CRTs is not perfect. The mastering engineer is viewing on a CRT display and actively changing the encoded levels so that they are imaged correctly on the display. This reverse-float compensation in black is allowed with BTB footroom. This reverse float compensation should be the only times data encoded below black is visible in the final image. If you are using a PLUGE pattern with BTB bars to calibrate, you should calibrate so that the BTB data is not visible. BTB data also helps define dithering duty patterns on DLP projectors. Lastly, BTB and peak white data is quite useful for any image processing/scaling etc applied to an image. For all these potential reasons, video engineers advocate preserving full BTB and peak white data whenever possible. Lastly, this data is helpful for any image processing that is applied on an image, such as sharpening, scaling, etc.


Please also see this thread, in which Stacey and Don cover this in their discussions:

http://www.avsforum.com/avs-vb/showthread.php?s=&threadid=416292

Including this post on page two of that thread:

It is, in fact, real-world and you can absolutely see the below-black detail, if you're using a CRT (or a fixed-pixel device that properly emulates a CRT).

First off, the below-black samples affect the waveform output from the D/A converter. In the same way that hard clipping of audio samples produces unfortunate frequency-domain artifacts that sound bad, clipping the video waveforms produces visible artifacts.

Secondly, no CRT has perfect black-level retention, and when there is a bright image on the screen, the black level goes up, revealing some of the information that is ostensibly "below black." The CRTs that telecine operators use have exactly this characteristic as well, so when the operator adjusts the shadow detail down so it's just barely visible it inevitably shifts data into the below-black range. Think of the 16 level not as a hard barrier but as a target level. The real black level floats around quite a bit.

This is easily verified (and I have done so). I have clips right here on my desk that have lots and lots of below-16 and above-235 values in them. These are not weird clips or unusual clips. I didn't seek them out as test material - they're just clips I happen to have around. I've actually never seen a video clip that stayed entirely within the 16-235 range, with the exception of test patterns. In my experience, real video ALWAYS goes outside the "nominal" range.

Similarly, I can and do demonstrate that clipping at 16 and 235 is visible. It's not just a little bit visible, it's awful. Your white detail goes all to hell. Clouds look like white plastic cutouts in the sky. Snow scenes lose all their detail. People wearing dark clothing look like silhouettes.

The real lesson here is that computer graphics should have been allowing for headroom and toeroom all along. Video engineers know much more about how to display images on a CRT than computer engineers, but they were not consulted when the computer RGB standards were created. The result is this flawed 0-255 standard, which was not really designed around showing high-end video and images in the first place.

Going forward, it's clear that the video concepts of headroom and toeroom will seep into all computer graphics interfaces, and eventually all graphics will use "video" levels or at least a format that allows for head and toeroom. Until then, we have these two standards to deal with. I wish we didn't, but there you are.

Don


If you want to continue nitpicking with imaginary situations, please move it to private messaging in deference to others who may benefit from keeping this thread open to added explanations. This has been debated for over 400 posts and 20 pages in that thread. I refuse to allow this thread to rehash that repeatedly.

Chris,

You say a few times that it is important to maintain studio RGB levels, and I follow the reasoning behind all of that.

The question that I still have, is how does the non-ISF person calibrate their HTPC/display to properly display studio RGB levels?

There are quite a few test pattern generator tools out there in HTPC land that allow you to setup your display, but presumably all of these output PC levels. I'm thinking specifically of the Philips and Nokia monitor test patterns.

Is there a way to use these, or perhaps a different generator, to properly set up sRGB?

In deference to Chris' concerns that this thread be focused on practical matters of just HOW to achieve proper gray scale and color rendition in the presence of real-world, home theater equipment, as opposed to whether or not the long-established standards for representing moving images in analog and digital video (plus the newer, evolving versions of those for high-res video) are actually "a good thing" or not, I won't be responding to any further questions -- in this thread -- on WHY BTB data should be preserved, or just how significant a problem it can be if it isn't preserved.

The theory and practice on this stuff is solid, but the math and signal processing theory that underlies it is just too much for a practical thread, and the simplified analogies and thought experiments that have been used so far to try to convince folks are obviously not doing the trick (based on at least a year's worth of threads on this topic -- not just the recent discussion in this thread).

I'd be happy to keep trying, up to a point, IN SOME OTHER THREAD, but in a perhaps forlorn effort to forestall that, let me just say:

TRUST US ON THIS!
--Bob

jvincent: I use PC test software mainly to get high-resolution patterns for this reason. They will allow you to calibrate to graphics (PC) levels only, usually. You can manually make patterns with data at the levels you desire, but the easiest way is just to use Avia or DVE. Calibrating this way will calibrate you properly to whatever your video software is doing. If you manually calibrate to Studio levels, and it turns out your video is being expanded to PC levels, then you'll be off. Just use a test DVD and you can be sure that if you calibrate to a test DVD, other DVDs will look correct.

jvincent,
If you've got your HTPC system set up to play commercial DVDs then you will have no problem checking this. Get the Digital Video Essentials (DVE) calibration disk and check that Blacker than Black portions of the appropriate test charts can be made visible *TEMPORARILY* by cranking up the Brightness on your display. For example, on the DVE pluge chart the outermost black bars on each side are encoded at a "below black" level and that data SHOULD be making it through to your display. If you can make three bars visible on each side by temporarily cranking up Brightness, then you know the data is getting to your display. Now remember to finish by re-setting the black levels PROPERLY by reducing Brightness again. When you are done, the outermost of those bars will vanish (merge back into the black background).

If on the other hand you can only make two bars visible no matter how high you crank up the Brightness then you have one of two problems. Either the BTB data isn't getting to your display -- which could happen if you are sending PC-style digital video, or alternatively if you are sending Studio-style digital video but with BTB improperly clipped -- OR your display doesn't have enough calibration range for Brightness adjustments to show them -- which could happen for example if you are sending an analog video signal at the 0 IRE level when your display only has enough calibration range to handle a 7.5 IRE signal (or the other way around). In either case you need to try to fix this problem.

Once you know that BTB data is getting to your display AT ALL, you can then use the DVE gray ramp charts to check this to a finer level -- including also confirming that Peak White data is getting to your display while also refining your black and white level settings (Brightness and Contrast respectively).
--Bob

Chris and Bob,

Thanks for the replies. I don't have DVE, just AVIA (non-pro), but I can use the THX optimizer that comes with some DVDs since it has a BTB pattern.

Just as a sanity check, what I have been doing to date is using the PC tests to set up my display (CRT HDTV connected via component) such that the 1% black bar is only just barely visible and 100% white is, well white. I then set up my DVD player (TT using VMR9) using AVIA such that the left black bar is only just barely visible.

Now if I understand correctly, what I have just done is set both my display and DVD player/ HTPC to 0 IRE black levels and assuming that it still sends BTB to the display (I need to check this) I should be OK. Is this right?

UPDATE: So I just checked with the THX optimizer, and I am indeed sending BTB as seen by cranking up the brightness on the display so I guess I am good to go.

It was also worth noting that as I lowered the brightness setting in TT the BTB settings started to get clipped.

jvincent,
The trick here is that you are fiddling with both ends of the signal and so you may be causing problems at one end that you are then trying to bandaid fix at the other end.

Normally for people with commercial players I'd recommend you keep the commercial player set at its factory defaults (except possibly for having to change 0 IRE vs 7.5 IRE for analog video output). The factory default settings for most modern DVD players really are pretty likely to produce the best signal that player is capable of -- particularly if you pick a "picture mode" on the player described as doing the LEAST to the image. This is quite unlike the case with TVs where the default factory settings are likely to be truly awful.

Then, trusting that your player is already doing the best it can, you now make ALL calibration adjustments at the display end.

What you are doing on the other hand is adjusting your CRT to show it's full dynamic range and then separately adjusting your computer-based player to calibrate the signal from the DVD against that CRT range.

Just to give you an idea of what might go wrong here, suppose the 100% = "white" level on your CRT happens to be driving the CRT a bit too hard so that you are getting beam focus "blooming" and perhaps even modest levels of geometry distortion as the CRT's power supply tries to keep up. Or perhaps the CRT will start to show color shift if you drive the phosphors that hard for very long.

Now when you play Avia test charts via your computer, you will see such problems and will turn down the top end of the output range of the computer player so as to eliminate these problems -- effectively driving the CRT at less than the 100% you previously set up. But that means you have just compressed the dynamic range of the signal coming out of your player. That may in fact be the best way to image perfection, but another possibility is to adjust the top end of your CRT to a lower light output level and leave the player's end at a wider dynamic range so that "white" from the player is a higher voltage but only drives the display at, say 95% of it's maximum light output.

Since neither your player nor display is dependably calibrated to begin with, you might need to play around to find which combo works best.

But on most TVs, for example, you would be very badly served if you set up the TV to show "white" from the DVD as the maximum possible light output level.

A "safer" starting point would be to set the CRT such that the 100% signal level generates an image which is just beyond the point where you perceive it as "gray". I.e., "white" but only just barely so -- as opposed to maximal white. This would be the equivalent on a home TV of lowering Contrast with the intention of watching the TV in a dimmed room.

Ideally you would need a video signal scope and a light sensor to do both ends of this calibration. First making sure your computer-based player is sending the best signal it can, with proper linearity of the gray steps, and then adjusting your CRT to accurately reflect that signal as steps of light output.

I don't know enough about the nature of your computer based setup to know if it has dependable defaults. But if it does, then you might also try leaving the player end at the default settings and doing all calibration adjustments at the display end, just as if you were using a commercial player.

------------------------------------------------------------------------------

Also note that if both your computer based player and your display are equally happy with analog video signals at 0 IRE and 7.5 IRE voltages, then after you calibrate you will be unable to distinguish on the basis of any test pattern or image exactly which voltage you ended up using. The images will look identical once properly calibrated each way.

In particular, any given test pattern off a DVD can't know what voltage standard you are using on the analog video cable. So a test chart that identifies "Black" as "0" for example, doesn't necessarily mean that the signal for black on the analog video cable is 0 IRE (voltage). It could just as easily by 7.5 IRE (voltage).

So when you say you have set up your component connections to use 0 IRE, you can only know this for sure if your test image generator is driven by the 0 vs 7.5 IRE selection, and thus changes according to which voltage you select for output -- which won't be the case for any DVD calibration disc.

The only problem to watch out for is that not all players and not all displays work equally well with both voltages. The player or display might clip BTB data at one voltage and not the other for example, or the display might not have enough calibration range to be adjusted properly at one voltage when it works fine with the other.

The most important thing is to pick an output voltage standard that allows you to get the display into proper calibration at both ends of the gray scale -- both blacks and whites. If only one voltage level allows for that, then that's the one you've got to use unless you can find a setting in your display that lets you use the other. Even if that means you lose BTB data.

Having found a voltage that lets you calibrate blacks and whites, then you should also check that BTB data and Peak White data is getting through properly. If it isn't there may be some setting on either the player or display which enables it -- without forcing you to use a voltage level that won't calibrate.

----------------------------------------------------------------------------------------------

The black levels test chart in the THX Optimizer on some commericial DVDs is OK to use to double check that BTB data is being passed, but it shouldn't be used for actually setting calibration levels because it is not as dependably accurate as what you'll get off of Avia or DVE. On some commercial DVDs the THX images have been massaged as part of the final editing process that created the DVD, and thus inaccurate levels are introduced.
--Bob

I realize I'm sort of thread jacking here since this isn't the HTPC forum, but hey, I got here from there so perhaps others have, so perhaps this will help them.

I think the boys at ATI have set up the component output pretty well. My calibration procedure was as follows:

1. Video card set to defaults
2. Set contrast and brightness using PC test patterns.
3. Check contrast and brightness of AVIA.

As it turns out, I didn't need to change anything in AVIA. After setting up using the PC tests, I was bang on.

I see what you're saying with respect to not knowing whether or not I'm actually sending 0IRE or 7.5IRE from the computer. I haven't actually connected a scope to the HTPC output so I could in fact be sending either. Looks like I'm going to have to bring a scope home one of these days and check it out.


I've heard that the THX optimizers can be "off" so I haven't ever used them before. I only just checked this one out for the BTB test.

One more questioin, I believe the IRE values are for RGB. What are the corresponding voltage levels for YPbPr?

jvincent,
Yes, that sounds OK. If you are going to spend money on test equipment, you'd probably be better off getting a light sensor first so that you can check gray scales precisely and verify things like gamma correction. [I believe Ovation has a "such a deal" on a combo of Avia Pro and a good light sensor if you need to lighten your wallet.]

Also be aware that all CRT based displays drift in settings and geometry both as they warm up and as they age. Give the CRT a good 20-30 minutes warm up time before you adjust it. You might also want to get in the habit of double checking your calibration periodically (at least once a month) until you get a feel for how rapidly your particular display drifts. Re-check image geometry and color convergence at the same time. Critical studio monitors are always re-adjusted on a regular schedule for just this reason.

I don't think there's a variation in voltage standards between the two analog formats, but I'm not sure on that one. Chris?
--Bob

I wasn't planning on buying a scope. We've got plenty around the office that I can borrow for the weekend. I was eyeballing some light sensors just the other week. Maybe next Christmas after the reno / new TV.

Only once a month? My wife/kids would be happy if I was in service mode that infrequently. :)

I'm one of those guys who cringes when my set is out of convergence so it's checked pretty regularly. Up here in the great white north (-40C today with the windchill) my basement gets pretty cold overnight so I get a fair amount of drift in the winter due to thermal cycling of the set so I'm pretty good at reconverging the set.

btiltman
Sorry to be so blunt, but unless you follow the directions and have a player that acts as I have described, the demonstration will not work: you're adding variables to the comparison and therefore will learn nothing from the experiment.


Phat Phreddy
>>>"If I am still stuck miss understanding then just say so but if a display is perfectly calibrated would all those values not be outside of the dynamic range of the display.. Eg all that below black would be 'clipped' at the display ??

Thats why I dont understand the advantage of below black in the outputted signal.. If I am supposed to tune black so that below black is not visible this information is then clipped at display ??"<<<


Not "clipped" at the display, just not "visible" to your eyes. The signal is still there and is part of the entire image processing chain in order to "support" what you are supposed to see.


The side affects of clipping signals in digital video is more or less the same in digital audio. When the audio signal clips in digital, it's usually a rather obvious and detrimental sound. Besides the frequencies clipping, those around them feel the shockwave that is digital clipping. Same with digital video: when you clip a signal, the signals around it (above and below) are negatively assaulted as well -- think of the "pebble dropped in water" scenario. This is why clipping black doesn't mean that digital signals from 1 - 15 are lost. It also means that the signal above the clip (the darker and, to some extent, the lighter grays) feel the affect of that loss. This is why that scene I described looks different on a capable display between the unclipped vs the clipped image.


The range for human hearing is said to be 20 Hz - 20 kHz. Music goes below 20 Hz and the range of instruments/harmonics may extend as high as 100 kHz. Do you think you could hear the difference between these two presentations:

A) a full-range SACD or DVD-A recording of a classical symphony
B) the same recording, but with hard truncation of the frequency extremes at 40 Hz and 15 kHz?


I think if anyone actually takes the time to read through the topic post in this thread and the links that were provided, all questions should be answered. This FAQ was an effort to assemble the information in one place, rather than having it scattered around the forum. If that doesn't do it, then I have to say that Bob said it best -- "TRUST US ON THIS!".

Joe, I thought I had it pinned down more or less, but the audio analogy you presented actually confused me. So, please correct any mistakes in the following:

My SACD setup reproduces frequencies that are outside my hearing range but do exist in reality. Fortunately, the recording and playback systems were able to capture them and so the overall feeling I get from the music is closer to reality.

In the visual domain, however, our recording (and reproduction) media are not able to cover our range of vision. As a result, there is real-world information that does not make it to the film and any BTB information has nothing to do with it. BTB, introduced further down the path, is just engineering tricks that serve to counter display technology and image processing artifacts. Moreover, the PC does not need such tricks, because it (usually) creates information rather than reproducing it.

If we ever make a film with a dynamic range outside our seeing range, and if the processing and display media are able to reproduce it, we will have no need for BTB information, because any artifacts will go undected by our eyes.

jvincent:

Computers don't really follow the 0IRE/7.5 IRE output standards quite the same. Guy covered this in a thread that I linked, here is an excerpt that may help overall:

Computer RGB puts black at RGB 0,0,0 and absolute max white at 255,255,255. No blacker than black footroom nor whiter than white headroom are provided. The output is set up such that black RGB 0,0,0 corresponds to 0 mV and white RGB 255, 255, 255 is at 700 mV. Unfortunately, actual video material also requires footroom and headroom data in the signal to be represented or else image information is lost. This need was provided for by allocating the bottom and top of the digital signal range to permit "blacker than black" and "whiter and white." Digital video encodes a wider dynamic range that goes from blacker than black to whiter than white. Black is at digital 16 and white is at digital 235. Blacker than black data is allocated digital 1 to 15 and whiter than white is 236 to 254.

This creates a problem when it comes time to implement how digital video should be represented on a computer video card. Once could clip both ends of the digital signal range and then expand the remaining range to map black at RGB 0,0,0 and white at 255,255,255. This irretrievably clips video information and can create banding artifacts because a smaller range [16..235] is expanded to [0..255]. The value mapping isn't completely monotonic so banding can be induced. Another way to handle this would be to simply shift the digital data downward by 16 so no range rescaling is done, but this still cuts off blacker than black info and makes white on video much dimmer than white in computer video. The preferred solution for Media Center Edition PC's - computers specifically designed for multimedia use, is one that preserves video signal integrity over computer graphics. Black and white are kept at the values 16 and 235 and the MCE qualified displays are adjusted to properly display digital 16 as black and digital 235 as white. This avoids banding issues and displays video at full range. Computer RGB is less accurately displayed (unless it is in studio RGB with black = 16,16,16 and white = 235,235,235), but since the MCE's primary function is to provide high fidelity multimedia playback, the tradeoff is a reasonable one. Some older (esp. LCD panels) displays may lack the controls needed to adjust black level to make digital 16 true black. Displays with such limitations are not considered to be MCE compatible.

In HTPC's, the end user is free to make their own choices about how video signal levels are to be ranged and offset to fit within their video card output range. The degree of signal clipping and banding will vary depending on the user choices. The tradeoff between computer graphics and video graphics fidelity is certainly grounds for debate in HTPC's and no one right answer will satisfy all owners. For that reason it is not possible to state for HTPC's what mV is "correct" for black or white. It depends on the user goals and display capabilities. For MCE's the choice has already been set and both MCE and display manufacturers will very likely come to follow or allow for Microsoft's standard for the operating system. Black is at digital 16 and white is at digital 235.

If you really need to think of this in mV on an MCE....

Video Black = digital 16 = 16/255 * 700 mV = 43.9 mV
Video White = digital 235 = 235/255 * 700 mV = 645 mV

On such a system the Avia and Avia PRO 7.5 IRE labeled patterns would be at Video Black or 43.9 mV
and 100 IRE labeled patterns would be at Video White or 645 mV.

On a HTPC, your guess is as good as mine because the video signal scaling and offsets are not standarized. At any rate, I hope people adjust their HTPC and display COMBINATION to make black be displayed as black and white displayed as white.


__________________
Guy Kuo
Director - Imaging Science Foundation Research Lab
Video Test Design - Ovation Multimedia / Home of OpticONE Colorimeter, AVIA and Avia PRO

LeptonGR
The fault is that you are, in essence, saying that the only signals that need to be recorded and reproduced are the ones that we can actually see. The fact is recording and reproducing signals below reference black and above reference white support and serve as a safety net, if you will, for the entire film-to-display chain. The decision where to put/reference black and white was not up to the DVD, the display or you. That decision was made by someone further back in the chain. All you can do is try to get your display to reproduce those reference levels based on agreed upon guidelines (eg; a calibration disc). In the film-to-disc chain, if someone got sloppy -- it does happen -- and encoded black a bit low or white a bit high (hence the headroom at both ends), you want your display to get all of the signal, not a clipped/compromised version.

I thought the audio analogy would actually help: I apologize if it made things more complicated. What it was meant to show was that extra bandwidth is needed to properly convey the defined bandwidth. Nasty things happen in digital very fast, unlike analog where things happen slower (distortion, clipping, etc), and those nasty things affect things around them. How's that for a technical explanation? :D You want to keep those nasties as far away from the meaningful data/signal as possible.

If it helps, don't think of digital 16 and digital 235 as black and white (ie; colors): think of these as reference points for signals (ie; intensity levels).

I am usually careful to call the two terms "Reference Black" with data below this point as "Blacker-than-Black." First, this point is fairly concrete at digital 16 (except from some floats in the 'real-world' number value of black to compensate for CRT float as explained). BTB data will not be visible when correctly calibrated (again, except for this float), but the presence of the BTB data through the video chain will affect the final visible image.

I call 235 "Nominal Reference White" and values above this as "Peak Whites." I think this is more clear than calling them "whiter than white." Peak Whites will be normally visible on any CRT display that is properly calibrated. They don't just affect values below reference white, but but peak whites are *directly* visible as detail in bright objects.

Note the difference between Reference Black, and 'Nominal' Reference White. You won't directly see BTB data(usually). You *will* directly see peak white data. This ends up providing video a *larger* dynamic range than for graphics, since graphics does not allow for any data outside the bounds of black and "white."

Chris, how is a digital display device supposed to handle peak whites? Should they be visible there, too?

Displays with hard limits on on/off contrast like current digital displays are a little bit more difficult to deal with. In my opinion, you should calibrate so that peak whites are included in the final image without colorshifting or clipping them. Others do this as well, but I also understand that some may want to increase the white levels slightly, to increase perceived contrast a little bit. I would still hope that at least a portion of peak whites are being maintained, and this choice is a very subjective one. I won't be as stringent to advocate that the full range of whites *must* be seen in the final image, but I would certainly feel that it is preferred. It also does depend on the display. If a display hard-clips, I would find this slightly more acceptable than colorshifting. You don't want your whites to be off-white. If your display is colorshifting with peak whites, I would probably calibrate so that all the peak whites are maintained correctly as white (below the max white that starts to colorshift) to avoid this problem.

Btw, the Radeon's overlay correctly maps the 16-235 range to 0-255. VMR9 does the same except for YV12 or NV12 input modes (I guess this is just a bug).
And as quoted above the 7,5 IRE setup is irrelevant for digital and VGA outputs.

This means that you usually don't have to calibrate a HTPC.

Nils,
That's going to need some more explanation.

There IS NO correct mapping from 16-235 to 0-255. If you shift "black" from it's proper location at 16 for home theater video to the PC-style encoding where "black" is represented as 0 then (1) you have eliminated any possibility of properly passing Blacker than Black data (there's no space below digital 0) and (2) you have inserted additional steps in the gray scale which may very well result in banding. Now there are better and worse ways to conceal that last problem, but the loss of BTB data should not be tolerated.

Furthermore, saying that typical HDTV-ready displays don't need to be calibrated simply because the source is an HTPC setup is also false. Most such displays ship with manufacturer's default settings -- even for digital video input -- that are just ghastly: The justifiably disparaged "torch" mode settings. At the very least you need to calibrate to extinguish the torch.
--Bob

Bob,
There IS NO correct mapping from 16-235 to 0-255. If you shift "black" from it's proper location at 16 for home theater video to the PC-style encoding where "black" is represented as 0 then (1) you have eliminated any possibility of properly passing Blacker than Black data (there's no space below digital 0) and (2) you have inserted additional steps in the gray scale which may very well result in banding. Now there are better and worse ways to conceal that last problem, but the loss of BTB data should not be tolerated.

BTB should not be displayed. This is a part of the DVD/HDTV standard. If you calibrate your chain that it will pass BTB then you not only don't view the movie like it is intended to be also the majority of all movies will have a too high black level causing a "washed out" picture.

Banding may have been an issue with old graphic cards, but with the introduction of internal 10 bit precision this problem is nearly obsolete nowadays.

Furthermore, saying that typical HDTV-ready displays don't need to be calibrated simply because the source is an HTPC setup is also false.

I didn't say that! ;)

I stated there is no need to calibrate the HTPC. The display is a completely different story. In fact every display should be properly calibrated with Colorfacts or similar measuring equipment, because "out of the box" most displays are not calibrated or calibrated wrong.

Nils,

again, we've just been over this and you are not correct. Expanding to PC levels is not desired. Have you read my explanations in the guide, the linked threads, and the subsequent posts I've made explaining why?

I will not continue to argue this. You have a pm.

Thanks,
--Chris

***I will NOT argue this again on this thread. Do not post about expanding to PC levels as being ok. It's been covered for 20 pages before, I will not repeat that here. If this keeps getting dragged down to newbie debates about this, I will close the thread, something that I really don't want to do.***

Nils,
OK, we agree on the calibration. I wanted to make sure we didn't end up confusing people.

If you trust that your HTPC video output stage (or DVD player for that matter) is made properly by people who care about quality, then it is usually best to ASSUME it's default settings will produce the best image it is capable of. For commercial DVD players (unlike TV sets), the factory default settings really are likely to be the correct ones to use. At least as a starting point. If you discover persistent problems after doing the best you can achieve with display calibration, then you need to revisit what your HTPC (or DVD player) is doing.

-----------------------------------------------------------------------------------

As for the BTB data, I'll have to suggest you re-read the earlier posts in this thread.

BTB data should not be DISPLAYED but, nevertheless, it should be PRESENT so as to keep any signal processing in the TV, panel, or projector from generating artifacts in the Black and Above data that IS supposed to be seen. This is all explained in detail in this thread.

If your display exhibits "floating black levels" as is typical with CRT-based systems, it is even MORE important that the BTB data get to the display because it will float up into visibility depending upon (usually) the average light level of what's currently being displayed.

However, if you generate a PC-style digital video signal where "Black" is represented as digital 0, there is no possibility of passing BTB data from the DVD to the display so that the display can take proper advantage of it to protect the quality of the Black and Above data.

So you should generate a video signal that INCLUDES BTB data (and check that it is really there by *TEMPORARILY* raising Brightness on your display until you can see it), but then calibrate your display so that the BTB data is *NOT* visible for normal viewing.

Even though it is not visible, the portions of the image that ARE visible will be improved because the signal processing in the display can take advantage of that "guard" data recorded below the light level which was arbitrarily selected by the producer to be represented as "black".
--Bob

If you calibrate your chain that it will pass BTB then you not only don't view the movie like it is intended to be also the majority of all movies will have a too high black level causing a "washed out" picture.

In theory this is true.

Films are processed using CRT displays. CRTs can't hold black at black, they float. As the APL in the image changes, the some BTB info will be displayed.

You do lower your CR when you calibrate to have all values above 235 visible. I went from 1500:1 down to 1000:1 when I did that. Too me it is worth the loss in CR to not have blow out whites.

The gamma of the display is critical. A lot of new digital displays have a gamma of 2.2, but this will reproduce an image that is not correct. That CRT I mentioned above has a gamma of ~2.5. In order to see what the film makers intended, your display must have a gamma that matches.

Big apology, guys. I didn't read carefully enough. Now I understand why you suggest to use the whole 0-255 range.

I will try this on my 1292 tomorrow.

There's another issue besides the "CRT floating black level" that applies to all technologies, not just CRT: the floating black level of your visual system.

The SMPTE standard for black level calibration is to put up a 75% color bars+pluge pattern, increase the black level (brightness) control until the below-reference-black bar is visible, then back it off until the bar just disappears. In most cases, this will not be exactly when 16 is at the minimum light level. Because of the light washout effect of the 75% color bars, there will still be a little room under 16 that is visible when the average picture level of the whole screen is lower. This effect is by design.

Your visual system calibrates its sensitivity range based on the brightest object in view. Thus when you have a bright pattern (the color bars) on screen, your black sensitivity goes down, partially because your iris closes and partially because of other biochemical processes in your retina. So your perceived extinction point on the monitor will generally not match the physical extinction point on the monitor. This partially explains why your calibration will be different if you use a pure pluge pattern instead of a pluge pattern combined with color bars or a white field. Given the SMPTE recommendations, it's clear you should be setting black level with a pattern that includes either color bars or a white field.

In practice, code 16 from a digital video source like DVD will be calibrated at close to the physical extinction level (the physical minimum black level) of the display. But not exactly - when the SMPTE procedure is used, the perceived extinction level will not match the physical extinction level. Again, remember that the telecine operators are using a CRT monitor, and using the SMPTE calibration procedure. If you want to see what they see, you'll want to use a similar procedure.

Thus I am uncomfortable with statements that suggest that the below-16 values will never be seen. The telecine operators absolutely do see some below-16 values, as well as seeing the indirect effect of below-16 values (the aforementioned image processing and analog issues). Again, if you want to see what they see, you should calibrate your system accordingly. Are you going to see, say, a pixel at 8 surrounded by pixels at 9? No. But you'll almost certainly be able to see the difference between a dark suit with details that vary from 13 to 16 versus a suit that has all of the pixel values clamped to 16.

The upshot is this: toeroom is not just for the pluge pattern. There are good technical reasons for it, and it's not supposed to be clipped off at any stage of the video pipeline.

Don

Apolgies in advance Chris..

There are good technical reasons for it, and it's not supposed to be clipped off at any stage of the video pipeline.

Now we are back to where I began.. If the video chain results in an image that resolves blacker than black at the screen then exact black (16) will be an elevated black level.. If you use that image as a reference then all your programmings 'black' will be higher than the best black of the device and you will have continually compromised CR. If you include that toe room (and the head room) you DRASTICALLY reduce CR when measuring the 16 - 235 range.. You cant help it..

While it may be a desirable compromise in a CRT based system I can see that it would be a good idea in the digital realm where CR is such an issue..

Also could someone explain why above white data is different from below black data ?? Why is above white desirable to be seen and below black calibrated out ?? Why is the reduced CR of above white a good idea (surely a correctly mastered title should have all intended data in the 16 - 235 range)..

Again sorry Chris...

Phreddy, the BTB data is confusing because as we've seen in my explanations and those by Don and Stacey, the "real" level for black that is encoded and seen can float around a bit. If, as Don mentions, you calibrate to make the BTB bars *not* visible in a high APL pattern, in low APL material you probably will see data that is ostensibly "below black."

whiter than white material should be preserved in the final image, and will easily be seen. The BTB detail is more complicated, because you are *not* calibrating for 0 (really 1) to be black. You are calibrating 16 to be black, but data below 16 can and will become visible in a CRT setting, or a display that properly mimics the behavior of a CRT. Meanwhile, you would ideally want to place 254 just at the range of the maximum white of your display, so that peak whites are not clipped off or colorshifted. I can forsee and understand that some digital users may clip off *some* peak whites for a touch more CR, but I think going so far as to put nominal reference white (235) at the absolute max white of your display and clipping or shifting all values above that is too damaging to the picture.

Phat Phreddy,
I'll just add one point to what Chris said, without trying to get too technical:

What you are trying to achieve is not some technical standard of data reproduction perfection -- some law that says this pixel will be visible and that one won't. Rather what you are trying to achieve is to see WHAT THE DIRECTOR WANTED YOU TO SEE.

This is much tougher than it might appear, particularly in the face of the numerous human factors that arise as choices are made during the reproduction chain, as well as the expected mangling that occurs during signal processing.

But correct transmission of BTB data throughout the reproduction chain makes it easier.

-------------------------------------------------------------------------------

For Whites, think of how you calibrate for loud sound. You want to calibrate the loudest sounds coming out of your system to a reference sound level -- again matching what the director wanted you to hear in a theater environment as the sustained level of loudest sound. But there are PEAK sounds that go above that to a certain degree. Good sound systems have a reserve capacity to go beyond the reference sound level, even though that higher sound level should not be sustained -- for comfort reasons if no other.

By calibrating "Reference White" properly, you give your display the best chance of showing typical brightly lit scenes properly, and without blurring, color bias, or geometry distortion. However highlights will go beyond that, and the BEST displays will be able to give that extra 110% performance for small sections of the image for short periods of time, even though calibrating the display to show White as at that level on a sustained basis might be beyond what the display technology can accomplish without introducing other problems.

Of course the display can only do this if the Peak White data MAKES IT too the display in the first place.

--------------------------------------------------------------

So Blacks and Whites really are different, just as inaudible sounds are distinct from peak sounds above the reference maximum sound level.
--Bob

But you are approaching this from a CRT background showing typical brightly lit scenes properly, and without blurring, color bias, or geometry distortion.On my DLP whites do none of the previous..

All exposure of BTB data wrecks my black level and causes dithering for the majority of 'correctly' leveled black data at 16..

Peak white and nominal studio white is less of an issue for me as I have a bright DLP so that I already have a pretty bright image even at 235.. Also rasing whites by 16 points at the upper end of the scale causes far less CR decrease than having black higher than reference (to me 16)..

Black float etc is not an artifact I see with digital.. Dithing and CR is more of a problem and the only 'float' I can imagine in black levels would be due to room reflections.

As an audio analogy I would consider studio white to be simply running a tiny bit short of peak db but leaving data below studio black would create a constant hiss..

What would be worse in an audio system ?? Little bit under absoulte peak sometimes or a constant hiss ??

Phreddy, you're missing the whole point. You calibrate 16 to be black on your DLP. A DLP that does not emulate a CRT in this regard and holds a very constant and precise level for black may be a weakness, but you would still calibrate black (no dithering) to be 16. Some will suggest that you bump your brightness up a couple clicks to maintain some detail a little bit below 16, this is a subjective preference. But we are *not* saying that you calibrate 0 (or 1) to be black on your DLP. Indeed, black (16) would be extremely elevated, and this is *not* what we're after.

I think the audio analogy may serve to obscure things, since we can't really talk about the white or black levels in an audio system quite the same way.

But you're misunderstanding the calibration, maintaining BTB data through to the display does NOT mean that you calibrate your black levels up so that you *see* the BTB data. As Don pointed out, with a CRT you will end up seeing some data that is ostensibly below 16. however, a DLP that does not emulate a CRT in this regard, and maintains an fixed black level should be calibrated to be black at 16, or ever so slightly below it. Either setting is a compromise. But you *don't* calibrate so that you see the full range of values. You shouldn't see the PLUGE bar. If you decide that you want a little bit of the BTB data, you may just barely see the BTB pluge.

Got it?

I am not miss understanding what you say at all..

What I am pointing out is that while you advocate a 'hard' cutoff at 16 (I agree.. Its the only way to keep a solid low black level for most content) yet Don is saying that 'some' below black info should be seen.. This is why I pointed out that some are argueing this point with a purely CRT mindset (blooming, color shifts, distortion) and not considering the digital side.

If you show the below black info you raise black and destroy 'average' black level and hence destroy CR on correctly mastered titles..

I personally would rather the vast majority of titles looked good and had high CR and low black level, than sacrifice this to see some below black that argueably should not be seen if correctly mastered.

Phred, the point is that the standards and mastering chains and etc are all based on the behaviors of CRTs. A display that does not mimic a CRT will not perform in the same way, thus not quite match what the engineers intended. Some may call this a weakness for now, some may call it an improvement. But in any case, you have to compromise this setting since your display doesn't behave like a CRT does. You can fix blak (16) at the lowest black level of your display, and possibly lose some visible shadow detail, OR you can bump up your brightness a couple clicks and maintain a little bit of BTB data. *BOTH* settings are a compromise. You have to decide for yourself how to deal with this weakness, as I've explained. I don't have a recommendation here, since I'm not using a DLP. I know some users that keep some data below black, and I know some that set 16 as the blackest they can go on their DLP (no dithering). I have my own predictions about what setting I'd settle for on a DLP, but I don't live with one, so I haven't spent a lot of time fiddling around with this compromise. I know more about setting my own CRT and dealing with a different set of problems in coming up with a black level setting.

however, a DLP that does not emulate a CRT in this regard

Depends on the DLP. The Samsung HP700AE was designed to emulate a CRT broadcast monitor.

Most DLPs don't do this, however.

I have an NEC 50XM4 Plasma and a Denon 2910 DVD player. DVD's are connected via DVI, HD satellite through component. In general the default display settings look quite nice. I changed the NEC's DVI setting to STB/DVD from PC and black level to high, per the manual black level should be set to low for PC and high for STB/DVD. The Denon's DVI black level is set to normal. Here is my conundrum, when I calibrate with DVE (DVD pluge with gray scale) these default settings are clearly passing BTB, all 3 bars are quite visible. When I throttle back brightness so that the 3rd bar just disappears the performance of DVD's with dark scenes looks terrible, a lot of detail in these dark areas just disappears. Any thought as to why this occurs?

sunra,
If you are correct that near black detail actually disappears then I can only assume your player or display is rounding near black levels incorrectly.

Are you sure that you are not seeing noise instead? I ask this because the below black bar in DVE is reportedly encoded as digital 7 where black (the background on that chart) is digital 16. If you adjust black levels (Brightness control) so that the third black bar just barely vanishes when compared to "Black", then in fact you have adjusted to make a portion of the Blacker than Black data (8 to 15 in this case) VISIBLE. Since Blacker than Black data tends to be of lower quality, this can show up as noisy low black regions of the image. Try lowering black levels (Brightness control) another notch or two and see if that improves the apparent image quality. Be sure to re-check white levels (Contrast control) since these two settings often interact. Remember that if you have black levels adjusted properly apparent "details" recorded in the Blacker than Black data WON'T be visible, and this is CORRECT -- i.e, it's what the film-maker intended.
--Bob

sunra
What gamma setting are you using? The NEC should be set to a 2.2 gamma. If that doesn't fix the loss-of-detail problem (when using the PLUGE pattern in Title 12 Chapter 2 to set black level), try using the Reverse Gray Ramps in Title 12 Chapter 14. The blackest bar is below black -- not sure about the digital level, probably around 12 -- so raise the Brightness control until the black bar next to it (which is at digital level 16) can just be distinguished from the below black bar, then back off 1 click. Black (digital level 16) has 3 white dots above it on the left side of the screen and below it on the right side of the screen.

I was actually considering a 50" NEC last February. Very impressive features and a shitload of adjustment capabilities. Unfortunately for NEC, their black is dark gray. In Plasmaland, Matsushita owns black -- now and for the foreseeable future. That's why I went with the 50" Panasonic 6UY instead.

Chris, et.al:

Hopefully this question isn't to elementary and close enough to on-topic.

In studying the thread, I believe I have a handle on the issue. I'm a little fuzzy on the white level setting. I know this isn't a calibration how-to thread but to help me understand, please allow this tangent. WHen adjusting my (DLP FP) contrast I have used the Avia pattern with the moving white bars. Since this is the consumer Avia, I believe I read that it's white 100IRE pattern levels are at reference white. I've adjusted my DLP so as to make only one of the 2 moving bars on that pattern visible. I have DVE, let me know if there is a more appropriate/better pattern and method there to adjust with.

Back to the real question... On my PJ there is a setting called "White Peaking" Since this is a term that's been used (sort of) in preceding posts, I thought it might be related to this discussion. In the case of a DLP display, can you or anyone enlighten me on what this setting actually does, as it relates to the input? What would be the proper way to adjust said "White Peaking" or account for it in the input?

Or is this parameter not relavant to the input and just a PJ behavior enhancement feature?

Thank you very much for the help and the great info in this thread.
Scott

As to the White Peaking "feature", what are the available settings: on/off, variable, etc? It does sound like a limiter of sorts and, as with most "features" (check out SVM in the asinine Mitsubishi "feature" lists in the Best Buy ads), it probably does more harm than good.

For white level on DVE, you can use Title 12 Chapter 14 (Reverse Gray Ramps). There are 3 black dots marking white and the outer white bars are just above white. Set the Contrast so the the white bar is just distinguishable from the outer above white bar. You can also use the pattern in Title 13 Chapter 2, the SMTE RP 133 pattern. There's a white box (100% white) that has an interior white box which is lower in intensity (95% white). [Note: the actual white box to use is actually to the left of the solid 100% white box] Set the Contrast until both are equal, then back off until the lower level box is distinguishable from the exterior white box. Joe Kane recommends either of these two patterns for digital displays, but, depending on your display, one of these patterns will probably be easier to use.

Originally posted by Joe Murphy Jr
As to the White Peaking "feature", what are the available settings: on/off, variable, etc? Thanks Joe. It's a 1 through 10 setting. The manual is kind of ambigious as to what it does. It says" Use the white peaking control to set the white peaking level (DUH) of the DMD chip. "0" stands for minimal white peaking, 10 is for maximum." But then it goes on to say: "If you prefer a stonger image, adjust toward max, if you prefer a smother more natural image, adjust toward min."

And thanks for the contrast adjusting methods. I think I lost you on the boxes but with the pattern up I'm sure it will become clear(er). :)

Thanks,
Scott

Joe Murphy Jr:

Gamma settings range from 1-4 on my NEC, I am using the default setting of 2.

Bob P.:

With the default NEC brightness setting I am actually very happy with the black level and resolution of detail with almost black images , very little black crush. Of course just because the default looks good didn't stop me from trying to make it better, that's when I used DVE to adjust brightness to where the 3rd black bar just disappears ( this requires substantial reduction from the default setting). Calibrated in this way, when I watch "Matrix" or "Planet of the Apes" there is significant crushing of black or near-black detail. When I go back to the default setting it looks fine. More than anything I trying to understand why what should be properly calibrated black level looks so bad. I will have ISF calibration done in the next few weeks, hopefully that will shed some light on my darkness question.

>>>Gamma settings range from 1-4 on my NEC, I am using the default setting of 2.<<<

sunra
I believe the 1 - 4 could be replaced by A - D. Setting 2 (aka "B") should correspond to a gamma of [insert a real number such as 2.0, 2.2, 2.5 (2.5 is probably the max)]. The manual may give you the actual numbers.

Make sure you have at least 100 hours (preferably 150 - 200) on the NEC before spending the money on an ISF calibration. I would hope that the calibrator made this clear when you set up the appointment.


scott
Yes, it will become more clear.

Very interesting read Chris,
But this leads me to an obvious question. If we know that NTSC OTA is all over the map and that there needs to be a correct calibration done for DVD playback, why would MCE2005 have you do your brightness and contrast calibration ONLY on the display device?? I'm sure there are ways abound this but MCE would lead you to believe that these are the absolute reference calibrations using there very well mastered reference material as the source.
As we all know this just doesn't exist and there needs to me adjustment done within the DVD player and the tuner to actually get these levels right. Otherwise one or all will suffer.
All of the Software DVD players out there have this ability and in some like TT you can setup individual DVD's to remember their individual setting.
But in MCE2005 I don't see the brightness, contrast, saturation and hue setting in the player (unless I missed them??)
I didn't read all of this thread so if something has been written on this, my apologies.

Terry

Maybe this will help anyone trying to set up DVD software on a HTPC...

My HTPC has an nVidia 5200 Ultra (66.93 driver), outputting over DVI to a Samsung 43" DLP. I use Zoom Player Pro + ffdshow (10-12 version), with DSclaer 5 audio/video codecs.

I had been using Overlay for my video renderer for a long time, because with the way my DLP is calibrated, the VMR settings looked out of whack (and used a lot more CPU power). I used the Phillips Pattern Generator and www.displaycalibration.com to match my DLP's settings to my HTPC. Doing so meant that my Overlay settings were perfect, and I didn't need to use any adjustments in my playback software or ffdshow filters. (I also changed my HNL DLP's gamma setting to 0, which is the best one and greatly improved PQ, especially SD cable, no more "clay faces")

However, after finally understanding a lot of other things in the HTPC/DVD-software realm, last week I finally got to this thread and discovered the "truth" behind levels. I was finally able to understand it, after having spent time using Levels in ffdshow to compensate for darks/lights, then actually using my display's Service Menu to calibrate it properly.

So, using Overlay, I tried to see if I could see the BTB section of the THX dropshadow screen (Monsters. Inc.). It wasn't there.

So I tried switching to VMR9, and it suddenly appeared. So I adjusted my Brightness (black level) to where it was before. But this was still too bright (blacks were grey). So I dropped Brightness from 50 (out of 100) to 40~41, and I got the blacks/levels perfect.

By switching to VMR9, I was able to get my BTB back. I checked a number of scenes in Finding Nemo, Matrix Reloaded, and more, and noticed that the picture seemed noticeably (yet subtly) better. What stood out the most was an almost complete reduction in banding (water in Finding Nemo, skin tones in Matrix (rave and Neo+Trinity chapter)), and also better shadow detail in certain scenes (ie., detail/buttons in Neo's jacket in Chapter 3 of Matrix Reloaded, when he is handed Smith's earpiece by the guards).

So, although VMR9 uses a lot more CPU power and I have to resize at 1280x720 rather than 2000x1600, my picture overall is much better. All I have is some Luma Sharpening within the Lanscoz 4 resize anyway. Supposedly there is a renderless VMR9 mode in the works for Zoom Player, so that should speed up VMR9 a whole lot.


I know Overlay vs VMR9 issues may depend on video cards and drivers, but if you're trying to tweak your HTPC, definitely check if you have BTB or not. If not, try switching renderers and get that correct. You'll definitely improve the PQ.

the problem is that with an htpc we gan adjust contrast,brightness and gamme in the player and thus compress the image dynamic, so how to calibrate the computer and the display at the same time ?
is a scope the only way to calibrate the analog outputs of a htpc ?
What about the digital outputs ?

Terry,

sorry I didn't get back to your question sooner, partially because I glossed past the MCE part, because I've not used MCE. However, in general, PC video cards just don't follow the same voltage standards as consumer equipment. This is why you use the brightness and contrast controls (when using analog output on a PC) to match the display to the source. Doing the reverse accomplishes the same, however, too many times source devices can implement what should be analog alignments digitally, and screw up your image. This is why I always suggest using the default settings from sources, including PCs, and using the brightness and contrast controls on your dislplay to match the source. As a CRTer, most any CRT that you'll come across will have plenty of quality adjustment in both black level (brightness) and white level (contrast) to accomodate whole ranges of sources. Especially the pro CRTs that you deal with, are designed to be able to handle graphics sources, and non-standard voltage ranges just fine. There is tons of adjustment range for this reason.

As for other issues, like color balance, etc, this is more software specific, etc, so I don't know that I can help particularly in this regard. It should be correct in the PC, and when it's turned into analog RGB there won't be any issues between the PC and the projector.

is a scope the only way to calibrate the analog outputs of a htpc ?

Some users sope their cards, then adjust the settings to try to match consumer voltages, but again, it's not usually that necessary since most displays will have plent of adjustment range to handle the outputs of a video card, especially pro-CRTs that Terry would be using.

Chris:

I think I understand the BTB thing, and hitherto have calibrated my HTPC/DLP PJ using DVE. Feed is DVI from Nvidia 6800GT card. I use ffdshow as the final post-processor. In ffdshow you can map the input digital levels to the output and there is the option to map the range 16 -235 to 1-254. i have tried doing this because at some point in the past there was a post recommending it. At the moment I am not doing it. Am I right to pass thru the full range including BTB and 'above white'?

The PJ (Sim2 HT300+) has Video and PC settings. I use the video setting so I tell myself I am seeing Studio levels.


Andrew

seyfang
Use the 16-235 output on the HTPC and the Video settings on the projector for Studio RGB.

Originally posted by Joe Murphy Jr
seyfang
Use the 16-235 output on the HTPC and the Video settings on the projector for Studio RGB.

I thought that if you did this in ffdshow you were compressing BTB and WTW into the 16-235 range?

Correct me if I am wrong, but you want to output full range, i.e. 0-254, but make sure that everything is calibrated for black = 16 and reference white = 235.

I'm not familiar with the program, so I wasn't commenting on how to utilize the settings.

The point that I was making was that the output from the HTPC should correspond to 16-235 for the video signal, thus placing black at 16 and white at 235. Below black and above white would be output as 1-15 and 236-254, which are the correct levels for video.

Agreed.

I just wanted to highlight that depending on how you use the levels filter in ffdshow you can either clip or compress ranges.

If you've been pacing the halls regarding purchase of Avia Pro, onecall.com wants you to stop. They're running a special right now ($100 off retail) on this multi-disc calibration set, but it won't last long.

Been there, bought that! Thanks for the pointer.
--Bob

Bob
I thought Guy had already given you a complimentary copy of Avia Pro? :confused:



;) :D

Nope, no tchotskas here. :(

By the by, I just spoke to Ovation, and they expect the audio test discs for Avia Pro to go into Beta testing any day now and hope to be able to start shipping them out to Avia Pro owners around the beginning of March. That would add the final two discs to the set as originally conceived -- Audio and DVD-Audio tests.
--Bob

Quote from first post:
"To reiterate, all the previous statements were nonsensical and meaningless. They represent confusion on the facts surrounding analog voltage outputs(expressed in mV or IRE units) as opposed to what is digitally encoded on the DVD."

If I'm designing a video A/D converter (which are used everywhere today) to sample a standard 0.714v (or 1v p-p with sync included) analog signal with a standard 75 ohm termination, what are the correct digital values to assign using limited digital black and white?

Is 0.000v assigned a digital value of 0?
Is 0.714 volts assigned a digital value of 235?

To deny that there is no relationship or simple linear conversion algorithm only serves to confuse the matter.

The relationship changes depending on the standard in use, and how accurately a piece of equipment conforms to the voltage standard. You cannot universally state that "X" voltage represents "y" digital level, or represents well-defined image content. The voltage values change dependin on the standards. This is why there need to be qualifiers about what standards are in use before you can state a voltage value that has any concrete meaning to the image. This should make itself clear through the explanations.

Further, my quote has to do with 4 very specific statements, all of which, are without *any* reservation, completely meaningless and absurd. Your question may be a more legitimate, but the statements that are "nonsensical and meaningless" were these:

‘Movie studio A encodes their DVDs at 7.5 IRE, while studio B uses 0 IRE.’
‘Test disc X is at 7.5 IRE while test disc Y goes down to 0 IRE.’
‘You must always switch any DVD player to 0 IRE for: darker blacks/better -contrast/blacker-than-black.’
‘HD can provide better blacks because it goes down to 0 IRE.’

Those are all mind-boggling for the reasons explained. They are not sensical in any logical universe.

Originally posted by Bob Pariseau
Phat Phreddy,
I'll By calibrating "Reference White" properly, you give your display the best chance of showing typical brightly lit scenes properly, and without blurring, color bias, or geometry distortion. However highlights will go beyond that, and the BEST displays will be able to give that extra 110% performance for small sections of the image for short periods of time, even though calibrating the display to show White as at that level on a sustained basis might be beyond what the display technology can accomplish without introducing other problems.

Of course the display can only do this if the Peak White data MAKES IT too the display in the first place.

--Bob
For non-computer based displays:
How does one calibrate common fixed pixel display "properly?
Does one calibrate for maximum output for an input value of 235?
Or does one calibrate for maximum output for a value of 255?
What do the service manuals state?

Does a fixed panel source light bulb generates the same amount of light?
Or are the "BEST" display manufactures modulating it with the input signal?

Most fixed pixels have a fixed CR range, that is also non-moving. Some use variably irises that maintain(roughly) the fixed instant CR range, but move it up and down.

Because of the fixed CR limitation, (this was discussed earlier in the thread), there is the problem of on/off CR range. You should place black, digital 16, at the black limit of the display (with some slight exception explained earlier). The white point is somewhat user subjective, but you would want to include some peak white detail within your calibration range in my opinion.

Originally posted by ChrisWiggles
Most fixed pixels have a fixed CR range, that is also non-moving. Some use variably irises that maintain(roughly) the fixed instant CR range, but move it up and down.

Because of the fixed CR limitation, (this was discussed earlier in the thread), there is the problem of on/off CR range. You should place black, digital 16, at the black limit of the display (with some slight exception explained earlier). The white point is somewhat user subjective, but you would want to include some peak white detail within your calibration range in my opinion.
Well it is rare indeed on AVS forum when someone earns my respect. And you just did Chris:
"The white point is somewhat user subjective, but you would want to include some peak white detail within your calibration range in my opinion. "

A huge problem with with the Internet and especially AVS forum is that their a many self proclaimed "experts" who state their opinion as fact. Or even more importantly, attempt to add to or change an existing international standard because, in their opinion it is wrong. Such arrogance, especially when their solutions are unable to withstand even elementary scrutiny.

You, in your reply, realize that their is an important compromise involved in SELECTING THE ONE WHITE VALUE where the projector generates its maximum light output to the screen.

The dilemma is if the projector is calibrated to white of 255, then the average light output will decrease substantially.

If you calibrate to 235 then these over spec values are clipped and simply displayed as white.

So the choice is either greater average light output OR reproducing an occasional peak with greater fidelity. Since most projectors are too dim, I would expect most people to choose greater average light output. The fact is there should only be one value defined for white, just as their is for black. That is, in my opinion. Have a nice day!:)

Now, from my perspective, I would probably err towards including peak white in the display's range. I don't live with a digital as I have a CRT(without this limitation), so perhaps if I lived with one my preferences would change. Whatever the case may be, novices should make sure that they aren't clipping the bars in Avia, as this means that they are losing white detail that is below nominal reference white.

From page 13 of the Avia Pro Guide. (http://im2.onecall.com/Image_Products/Ovation%20Software/AviaPROVideoChapter.pdf)

The above white bar can clip without endangering details of properly mastered material.

Dave

Guy is less stringent about this, it seems, than I and some others are. He still demands that video levels be maintained, but is assuming that properly mastered material will not have excursions above nominal reference white. I disagree on that, and that is why Stacey, etc calibrate their displays to include peak white detail.

Note also that his primary display is a CRT, so the CR limit here does not apply, as a CRT will easily be able to include these peak white details.

He still demands that video levels be maintained, but is assuming that properly mastered material will not have excursions above nominal reference white.
I can’t say what his assumptions were, but based on what I’ve observed, measured, captured and documented, I’d agree with the assumption about the assumption. He goes on to say this about the usefulness of BTB/WTW…

Because of the occasional poorly mastered recording, an ability to process signals in the below back and the above white regions can be useful.

Note also that his primary display is a CRT, so the CR limit here does not apply, as a CRT will easily be able to include these peak white details.

Not sure if you’re referring to Guys or Stacys display? but Stacy did state that he had to sacrifice over 33% of his CR for it.
You do lower your CR when you calibrate to have all values above 235 visible. I went from 1500:1 down to 1000:1 when I did that.

Dave

Stacey is using the Samsung 700A DLP projector that was designed with input from Joe Kane.

Excuse the ignorance here guys, I've been going through this thread and still have a hard time understanding everything. Are we saying that we need to change the input output levels in ffdshow to correct pc levels? Input at 16 and output at 235 in VMR9?

VC

No, VMR9, depending on your video card and drivers and such, should be outputting video levels. I believe overlay will expand to PC levels. You can easily check by looking at pluge patterns and ramps on DVE, or deep ramps on Avia PRO. You should not have to use the ffdshow levels slider to achieve video levels.

At one point with old drivers I was not able to achieve video levels in VMR9, so I used ffdshow levels slider, but it still caused banding problems. I was not able to get ffdshow to successfully eliminate the banding. I would avoid ffdshow levels for this reason. You shouldn't have any problems with VMR9, and it is simple to check.

If I have understand things correct with the DVE chapter 14 calibration I should not see any information outside the black and white dots?

So then calibrating blacks I will just see to that I dont have any info outside the white dots and then calibrating Whites I should not have any info outside the black dots??

Am I correct or?

Regards
Daniel

Daniel: you should not really see blacker-than-black, you should generally calibrate black to be the maximum black on your display. However, whites are a different matter. This is subjective, as we've discussed earlier on the thread. With CRT displays, you have plenty of on/off CR range to handle peak whites without any problem. Whether you include peak whites or clip them on a digital display is a user-preference, I personally would lean more towards including them.

Chris: I, too lookd at that particular pattern and wasn't real clear on how to use it properly. Can you describe again, or elaborate on using the patter (the one with the dots) on adjusting the whites. WHether or not you go for clipping?

Thanks,
Scott

Again, it depends on your display type. What you do for a CRT is different, right now I am assuming you are using a device with a hard-limit on on/off CR, a digital.

See posts 41 and 51 in this thread, they discuss the tradeoffs between including peak whites or clipping them on a display with a limit in on/off CR.

You can either clip down to reference white, you can include it all, or do something in between. You also have a similar choice if your particular display color-shifts instead of clips. IMO color shifting is probably more damaging and I would lean towards not seeing colorshifting, hence including everything.

Hi Chris: Thanks, but I wasn't clear. It's is a DLP incidentally. Whether or not I choose to clip whites, my question is more elementary...

What I'm asking is literally _how_ to use that particular test pattern to adjust the whites (to either include above white). I pulled it up once and saw the dots, but didn't understand what to look for and where on that test pattern. I ended up using a different pattern (I think I this thread that pointed me to it as well, posibly by you/Chris), the one with thewhite box inside another white box.

So, can you/anyone explain how the other test pattern (the one with the little dots on it) is properly used?

Thanks,
Scott

NEVERMIND, I found it:

>>David: the DVE pattern has ramps and steps, as bob notes, that are marked at reference black and white by three dots. The areas outside these dots are below black, and peak white. If there is clipping, the brightest and darkest steps (which are peak white, and below black respectively) will not be seen as a distinct step, and the ramp portion will "flatten".<<

Thanks,
Scott

Originally posted by Joe Murphy Jr
For white level on DVE, you can use Title 12 Chapter 14 (Reverse Gray Ramps). There are 3 black dots marking white and the outer white bars are just above white. Set the Contrast so the the white bar is just distinguishable from the outer above white bar. You can also use the pattern in Title 13 Chapter 2, the SMTE RP 133 pattern. There's a white box (100% white) that has an interior white box which is lower in intensity (95% white). [Note: the actual white box to use is actually to the left of the solid 100% white box] Set the Contrast until both are equal, then back off until the lower level box is distinguishable from the exterior white box. Joe Kane recommends either of these two patterns for digital displays, but, depending on your display, one of these patterns will probably be easier to use.

Thanks, Joe for the info on where to set contrast with DVE. That is not a very intuitive program and though I was able to figure out where to adjust for brightness fairly easily, I couldn't find out till now what pattern to use for contrast setting.
Problem I'm having now is that even maxing out the contrast setting on my tv (Sony lcd rpt) the interior is not equal to the outer white box. And on the Reverse Gray Ramps the white bar can be distinguished from the outer white bar at the highest contrast setting.
I'm assuming I'm doing something wrong with this, because the tv picture sure looks really bright at this high of a contrast level. Can you, or anyone else, give me some more pointers on this? Thanks.

Update: I think I see what I was doing wrong with the Reverse Gray Ramp image - I thought the bar under the black dots was the above white bar. So now using that image I can set Contrast closer to wha I would have expected it to be.
Still don't understand why I can't match up the two boxes on the SMPTE RP 133.

With an LCD, there are two things that you may see as you raise your white level: you may see clipping, or you may see subtle colorshifting as one color "runs out." With LCD/DLP etc, and unlike CRT, there is no danger to maxing out your white level. The negative, of course, is loss in image quality due to colorshifting or clipping of white details by setting it to high. You should strive to raise your white level as high as possible on these types of displays, to *just* below the point of clipping or colorshifting. You are trying to maximize as much as possible the available on/off contrast range in the display.

Note! Do NOT do the same for CRT displays, or other phosphor displays (plasmas) as you reduce the life of the display.

Thanks for the info, Chris. Some of this stuff is beginning to seep into my old brain. But think I am still doing too much 'analog' thinking rather than 'digital.'
After this las adjustment with DVE, my contast setting is actually very close to the default setting for the PRO mode on my tv. Still looks a tad too bright to me but I think I've gotten used to watching it on a lower setting. Going to give my eyes a chance to get used to this before making any other adjustments.

If it's too bright, the correct solution is to use a neutral density filter to lower the output while maintaining the same on/off CR range. A filter will lower by the same amount the points for black and white. You certainly don't want to be watching too bright an image, this is quite fatiguing! If you lower the contrast for a pleasing image, you are lowering the white point, but you can't lower the black point at all, so you are crushing the possible on/off contrast that you could have achieved.

Say, for those reading, this is a pretty good writeup of how calibration and PC versus Video levels differences emerge in the real-world.

Remember that you can't correctly display video levels and graphics levels simultaneously. You may need a separate calibration on your display that you switch between when doing different tasks with an HTPC.

http://www.avsforum.com/avs-vb/showthread.php?s=&threadid=523614

Thanks cyberbri!

Great thread! I didn't read it all yet, so I hope I'm not repeating someone.

First a comment about the first post.

A DVD is a digital SD source, as such the YCbCr on it was created using the SD 601 equations and should be decoded into RGB using the same SD 601 matrix or color errors will result.
This is very wrong. As can be seen by using GSpot, many dvds are encoded using Rec.709 (or BT.709 or whatever you call it). This info is present in the "matrix coefficients" field in the "sequence display extension" of the MPEG2 header. If it is empty Rec.709 coefficients are assumed.

More info:
http://forum.doom9.org/showthread.php?s=&threadid=82217

Also, when talking about hdtv clips. A number of them is encoded using Rec.709 coefficient, but others use Rec.601 coefficients.

Let me conclude by a question about the lumarange [16,235] of dvd material. Using AviSynth it is possible to see which pixels have a certain luma value.
Many times I see the following when considering a dark frame: the majority of the pixels have Y=16, but also a significant (but pretty much randomly distributed) part have luma between 10-16. How is this possible? Did that happen during processing, resizing, compression (before authoring the dvds) or a combination of them? Don't they clamp the lumarange before compression? For some screenshots (this can be done with the limiter in latest AviSynth version):

http://forum.doom9.org/showthread.php?s=&postid=515255#post515255

This can happen with the luma for example when resizing:
http://forum.doom9.org/showthread.php?s=&postid=205324#post205324

This is very wrong. As can be seen by using GSpot, many dvds are encoded using Rec.709

No, DVD is a SD source and should be, if encoded correctly, a 601 encode/decode. 709 applies to high-definition content, which should be encoded using the 709 matrix coefficients. If you are moving between the two, this difference in color encoding and decoding needs to be recognized. whether HD or SD material is encoded correctly, or mastered using the correct primaries (which are slightly different between SMTPE C and 709) is a different story...

but also a significant (but pretty much randomly distributed) part have luma between 10-16. How is this possible? Did that happen during processing, resizing, compression (before authoring the dvds) or a combination of them?

As discussed, all values from 1-254 are legal values, with a nominal reference range that defines black at 16 and white at 235. the headroom and footroom is included the for reasons explained in my guide. Values can fall outside (and as you have discovered often do) the nominal reference range.

i will read your links/threads when I have a bit more free time, but in the meantime, the issue of values falling outside the nominal reference range is summarized in my writeup here, but also see the thread I linked, paying close attention to Don and Stacey's explanations.

No, DVD (...) should be, if encoded correctly, a 601 encode/decode.
Is that stated in the dvd-specs (which i don't have and thus can't check)? But if that's the case it means that >90% (that's just a guess but i have the impression that most of it is encoded using Rec.709) is encoded incorrectly.

i will read your links/threads when I have a bit more free time
Ok, nice. If you do and have comments could you pm me, because i don't visit this forum weekly :)

Great info, I didn't read everyone's replies so I apologize if this question has been asked, which should be calibrated first, the DVD player or the projector?

I was told in another thread that I should leave the DVD player's settings, but that doesn't match what has been said in this thread.

Make sure your DVD player is outputting pluge/btb levels first. There may be some sort of black level setting in the setup menu, and you may have to temporarily turn the brightness (black level) on the display to confirm that information/data is being maintained all the way to the display.

Then set everything on it to default. You don't want to use any picture modes or adjustments on the DVD player if possible. You want to calibrate the display itself.

Thanks cyber, just wanted to make sure that I only calibrate the projector and not both

But if that's the case it means that >90% (that's just a guess but i have the impression that most of it is encoded using Rec.709) is encoded incorrectly.

Why would you think this is the case? I don't have any concrete info one way or the other, and there have been conflicting statements in the past about this, but it still means that if a studio is encoding a DVD in 709, it's making a serious and fundamental mastering error.

Chris,
See if this makes sense to you. On a HTPC using an NVidia 6600 card, I can set up as the PC background the DVE stairstep pattern that I have tested the values on and clearly the last black step below the dots is 0,0,0, the next black step up is 16,16,16 which should be black. 235, 235, 235 is the next to last white step and the last white step is WTW at 255, 255, 255. I have a capture of this pattern that I can put on an all black desktop and then "scope the output so that 0,0,0 is 0mV, and 255, 255, 255 is 0.713 mV or there abouts using the NVidia's "desktop" controls which do not directly effect the "overlay" controls for the software DVD player. I am assuming that what I have done at this point is to preserve the Studio RGB values on a PC RGB system. Agreed??
If so I can then go and using this desktop set it so that the desktops brightness is not showing BTB and it looks the same as Black but I can just barely see the differentiation of the next step, and WTW looks the same as White but I can see the differentiation of the next step down. BTW this is al done after the greyscale has been properly calibrated. Do you agree so far??
If so I can then go into my DVD player (TT in this case w/ VMR9 and renderless) and using the software players brightness and contrast controls adjust the BTB and WTW as I did on the projector above. Would you agree with this, and could we say that I have calibrated to Studio RGB levels first in the PC, then from that reference to the projector, and then once the PJ is set to the reference, the controls on the software DVD player.
Let me know if you agree with all this and I will come back with some additional questions.

Terry

I am assuming that what I have done at this point is to preserve the Studio RGB values on a PC RGB system. Agreed??

Yup. It'd be preserving even if you didn't scope the output, as long as it's being output. I am assuming we are talking VMR9, probably renderless in TT. (For others reading: the VMR9 should be preserving Studio levels by default, regular Overlay usually expands to PC levels by default. This can vary depending on software, drivers, etc etc of course.)


If so I can then go and using this desktop set it so that the desktops brightness is not showing BTB and it looks the same as Black but I can just barely see the differentiation of the next step, and WTW looks the same as White but I can see the differentiation of the next step down. BTW this is al done after the greyscale has been properly calibrated. Do you agree so far??

Set "it"? Do you mean the display? I'm confused as to what you are adjusting here. If you mean the video card settings, there isn't a way to have a reference display response really, (that I see), and also there is the danger of lowering the output and clipping BTB below 0mV. Here I am assuming you mean calibrating the display to this desktop that yo've set up above. This sounds good, but I would tend towards preserving the differentiation of peak whites even on digitals. On CRTs, there shouldn't be any problem with peak whites since a CRT won't clip. Try to avoid colorshifting the peak whites though, and this would indicate a pretty high white level setting on a CRT projector anyway...

If so I can then go into my DVD player (TT in this case w/ VMR9 and renderless) and using the software players brightness and contrast controls adjust the BTB and WTW as I did on the projector above. Would you agree with this, and could we say that I have calibrated to Studio RGB levels first in the PC, then from that reference to the projector, and then once the PJ is set to the reference, the controls on the software DVD player.

I think I see what you're getting at. I'm not sure how precise this would be, but it would be close. I like the idea of taking screenshots of the rendered image, and looking at the RGB values in paint and getting them lined up that way. I have been extremely busy lately, so this is something on my "to experiment with" list still. I think this method of measuring the actual values of captured patterns (making sure they're captured correctly, I have to figure out the print-screen thing, because screen captures in TT for me captures regular overlay or something because it's expanded and not what I'm seeing rendered).

I think your method relies more on visual alignment, and I think it'd be pretty likely to end up a click off or more, especially if you're not looking at 100% identical patterns on your desktop and in video, because of ANSI and things like that.

Chris,

Thanks for getting back so quickly,

I have a couple of tools that I use to capture the fames from the DVD that I believe mat be better than doing screen grabs from say TT. The tool I use for this is DGIndex and it seems very accurate.
When I said set "it" yes I was referring to the display device be it a monitor, CRT projector, Digital, etc.
I can use a meter to insure that what is being displayed from the desktop is identical to the DVD playback and that hey are accurate. Most folks don't have that kind of equipment and for practical purposes it is faster to do it by eye. It is being done in a darkened room with none or little reflections coming back to the screen.
Here is what I was driving at with all of this. I think that this method sound pretty accurate and the correct way to preserve Studio RGB. But in using it what I am finding is that the brightness is a few clicks to high and when there is a sene that I know the entire frame is black (16 and measured), the screen isn't completely black. There is a small amount lighting it up. If I then turn brightness down to insure that the screen is black during thus scene (btw, a good one is the first 10 sec or so of chapter 11 of finding nemo), then I feel that I loose some lack detail. This loss is noticable during higher light levels (a good exacple of this is the tux's that come out of the closet in The 5th Element.
I may be fighting the lower ANSI contrast of CRT projectors (LC's) and trying to get them to perform better than they are capable. One of the things I have wrestled with in using these captures is, is it better to set up brightness with only very low light level patterns (say btb to 25%) or setting it with the entire step pattern maybe more representative of what we will normally view?? The low light level pattern will allow you to adjust for great black detail and set the brightness lower, but as the pattern gets brighter, the black detail will go away.
What I wanted to insure was that what I was doing to calibrate the chain made sense and was preserving BTB and WTW.

Terry

I may be fighting the lower ANSI contrast of CRT projectors (LC's) and trying to get them to perform better than they are capable.

You surely have more experience than I do, but this is a real-world limitation that is subjective. You can't have complete can't-see-the-screen blackout on a full-black frame (at 16) and preserve full shadow details in brighter (higher APL) scenes. LC helps with this a good deal, but you still have to make a choice when setting your black level: do you prefer complete black-out for those rare frames? Or do you prefer to retain all shadow detail in even the brightest scenes? These are extremes, and most will end up somewhere between these when deciding on a final black level *at the display*.

I wouldn't use a pattern that is brighter than half 50%, like the half-gray plus bars on Avia. A full steps/ramp pattern is very bright, and this is too bright a black level for me personally, but it will make sure you are preserving every tiny bit of detail possible, including probably stuff well below black that may be present.

My way is actually to display a full-black frame(or with just bars), and raise brightness a bit, I sort of know where my sweetspot is subjectively, so that I can clearly make out the screen, but it's still very very dark. I feel this works well in preserving a good bit of detail that a *complete* black would obscure, while still giving you a sense of totally losing the screen when moving from a normal scene to a black frame, for at least a half-second or so. Too much brighter than this and you're not gaining too much detail and you're losing that total black moment or two as your eyes adjust/wear off, and too much darker and you start losing a lot of detail in brighter scenes.

What I wanted to insure was that what I was doing to calibrate the chain made sense and was preserving BTB and WTW.

It makes sense to me, but as before as long as you can see BTB and peak whites at the display when you artificially lower white, and raise black, then you're good to go from there to start calibrating. With HTPC, VMR9 should allow this pretty easily by default. I don't want to say that VMR9=correct Studio levels, because you can still screw it up with other software, drivers, etc, ffdshow, but in general if yo're using VMR9 you should be getting Studio levels pretty much right.

It looks like 30-40% on the screen setting 16 black to black and BTB is invisible is about as good as it gets. I worked on this today on a G90. You get about 90% (OK I'm guessing here) of the black detail and almost complete black when 16 black is on the entire screen (or like you said your eyes think it's complete black when the scene goes there).
This method works very well as far as I can tell to preserve Studio RGB, BTB and WTW information. I'm pretty satisfied with it for HTPC setups.
On TT2.11 we aren't finding that we have to use any post processing at all which is just fine with me.
Now back to the drawing boards to see what I can do to impreve the ANSI CR on these machines......

Thanks,

Terry

Originally posted by ChrisWiggles
In a well-designed player, the IRE option will have no effect whatsoever when using digital outputs. If this setting causes any change in the image when using the digital outputs, you should use test patterns to see which option leaves data un-clipped. In this instance, the degree of clipping or image alterations may be severe, so Avia, DVE, or any good test disc will come in handy.


Been reading over your excellent article again, Chris. And am wondering about your above statement.
I recently got the denon 5910 and have it hooked up via hdmi to my sony rp lcd tv. When changing the IRE setting from default of 7.5 to 0 there is a big change in the display. At 0 IRE I can only get the BTB bar to appear on DVE by changing the brightness setting on the denon to +2. Even with brightness maxed out on the tv it doesn't show up.
I'm wondering if this could be due to the digital signal from the player being converted to analog when it gets to the sony. In other words, is your above statement about IRE not affecting the digital outputs only true if the signal remains digital throughout its journey to the screen display?

If there is a change in the digital outputs when adjusting the IRE option, this is sloppy player design. I'm pretty skeptical of this being the case on the 5910.

I'm wondering if this could be due to the digital signal from the player being converted to analog when it gets to the sony. In other words, is your above statement about IRE not affecting the digital outputs only true if the signal remains digital throughout its journey to the screen display?

Not sure what's going on in your setup, but no if it's output digitally, it's digital there and IRE is out o the picture. If it gets turned back into analog at some point, it's beyond the player's ability to have any control of external devices and what they are doing in conversion and voltages etc.

Could the 5910 need a firmware update?

Originally posted by cyberbri
Could the 5910 need a firmware update?

Guess that is possible.:)
I'm not really complaining about it. Just trying to understand what is going on here. The default setting is at 7.5 IRE and it behaves well with that.

By the way. Thanks, cyberbri, for all the excellent links you provide. I've found them very useful.

Originally posted by ChrisWiggles
If there is a change in the digital outputs when adjusting the IRE option, this is sloppy player design. I'm pretty skeptical of this being the case on the 5910.

I would have thought so too. But then I can't disbelieve what my own eyes see. This happens with the DVE pluge and the THX Optimizer patterns. Once I set IRE to 0 the only way I've so far found to get the BTB bar to appear is by adjusting the Brightness setting on the Denon.

Maybe I've got a defective player? Guess I better try emailing denon. I have about 3 weeks yet to return it to Crutchfield if I'm going to have to go that route.



Not sure what's going on in your setup, but no if it's output digitally, it's digital there and IRE is out o the picture. If it gets turned back into analog at some point, it's beyond the player's ability to have any control of external devices and what they are doing in conversion and voltages etc.

Unfortunately, this is the only dvd player I currently have on which I can change the IRE settings. Have the denon hooked up to the Sony WEGA lcd rptv only through hdmi at 720p.

Thanks much for the info you've provided.

Originally posted by ChrisWiggles
If there is a change in the digital outputs when adjusting the IRE option, this is sloppy player design. I'm pretty skeptical of this being the case on the 5910.


Well, Chris, I'm sitting here with egg all over my face. I think I found out why I was having trouble seeing the BTB bar with IRE at 0.

With the HDMI hookup, Denon gives two options for output: "HDMI Y Cb Cr" and "HDMI RGB". I had mine set on the first option. When I switch to "RGB" then the BTB pattern appears normally.
Kris Deering talks about these two settings in his review of the Denon. Since the "HDMI Y Cb Cr" allows for 10-bit video output I mistakenly assumed that would work best with my display. Apparently not. :(

Ahab, what display are you using which needs the HDMI to be set at RGB.

Originally posted by WTS
Ahab, what display are you using which needs the HDMI to be set at RGB.

Sony WEGA KDF-50WE655.

Not really sure if it's not ok to use the "HDMI Y Cb Cr" output with this set. Things seem to work ok with IRE set at 7.5, But strange things happen with IRE at 0 that are resolved by setting to "HDMI RGB". So I'm assuming there's a problem with the Sony handling the 10-bit signal properly.
Please keep in mind that I'm pretty much of an ignoramus when it comes to figuring out how all these setting interact with one another.:o

Well I doubt it has anything to do with being a 10 bit signal, it would have more to do with it being coded for RGB or YCrCb I would think. But sometimes I wonder if these companies know what format(s) they should be using with these different interfaces, including Sony which is part of the HDMI group.

With an LCD, there are two things that you may see as you raise your white level: you may see clipping, or you may see subtle colorshifting as one color "runs out." With LCD/DLP etc, and unlike CRT, there is no danger to maxing out your white level. The negative, of course, is loss in image quality due to colorshifting or clipping of white details by setting it to high. You should strive to raise your white level as high as possible on these types of displays, to *just* below the point of clipping or colorshifting. You are trying to maximize as much as possible the available on/off contrast range in the display.



When you say raising the whites you mean increasing the contrast.
I have calibrated my Sony HS-50 (same as Hs-51 in the US) with DVE NTSC version and the calibration options for white (besides enhacned and other crap) is:
* Brightness = used to set BTB correct - so that the BTB bar disappears TOTALLY in the background = so that it just blends in and all "BTB pixels" are NOT lit up (looking very closely at the screen you see that when closing in on reference black only a few BTB pixels are "lit" (or should i say "not lit maybe" :).)
* Contrast = used to set the white clipping point with the "boxInBox" pattern. Here i set it so that the inner white box is just visible (set by contrast) whlie the "BTB box" is totally black (set by brightness).

Am I right here? This is how i interpret the text so far.
(My levels are studio on DVD and display - no problem there.)

Thanks for a very good layout - i have been wondering about how much to see of the BTB bar but as i understand it i should not see it at all (they say "just blends into the background" and it might look like it does from a distance but up close you see some pixels so you have to lower the brightness a few steps more).

Regards
Boogieman

Boogiem']When you say raising the whites you mean increasing the contrast.
I have calibrated my Sony HS-50 (same as Hs-51 in the US) with DVE NTSC version and the calibration options for white (besides enhacned and other crap) is:
* Brightness = used to set BTB correct - so that the BTB bar disappears TOTALLY in the background = so that it just blends in and all "BTB pixels" are NOT lit up (looking very closely at the screen you see that when closing in on reference black only a few BTB pixels are "lit" (or should i say "not lit maybe" :).)
* Contrast = used to set the white clipping point with the "boxInBox" pattern. Here i set it so that the inner white box is just visible (set by contrast) whlie the "BTB box" is totally black (set by brightness).

Am I right here? This is how i interpret the text so far.
(My levels are studio on DVD and display - no problem there.)

Thanks for a very good layout - i have been wondering about how much to see of the BTB bar but as i understand it i should not see it at all (they say "just blends into the background" and it might look like it does from a distance but up close you see some pixels so you have to lower the brightness a few steps more).

Regards
Boogieman



You are correct. "Brightness" refers to "black level" and "Contrast" refers to "white level."


For example, if you view the attached images in the attached zip folder, the black bottom-end image shows where "Black" should be for video levels. This is at level 16. Viewed on a PC set to PC levels, the blackest black will be at 0, and everything above that shows increasing shades of grey. Levels 15 and below are BTB, so you want to adjust your brightness setting so that btb blends into the black area, but so you can still notice a difference between level 16 and up (17, 18, 19, etc.)

If you were to adjust your display so that BTB was visible, everything encoded to be level 16 "black" in DVDs would appear as dark grey rather than the darkest black your display is capable of.

The same thing goes for the whites, although with digital displays setting the white level can be tricky. Take a look at the HTPC calibration thread in my signature (even if you're not using an HTPC), as there is good, basic information on calibrating, and one of my personal experiences about trying to set the white level correctly.


HTH

You are correct. "Brightness" refers to "black level" and "Contrast" refers to "white level."


For example, if you view the attached images in the attached zip folder, the black bottom-end image shows where "Black" should be for video levels. This is at level 16. Viewed on a PC set to PC levels, the blackest black will be at 0, and everything above that shows increasing shades of grey. Levels 15 and below are BTB, so you want to adjust your brightness setting so that btb blends into the black area, but so you can still notice a difference between level 16 and up (17, 18, 19, etc.)

If you were to adjust your display so that BTB was visible, everything encoded to be level 16 "black" in DVDs would appear as dark grey rather than the darkest black your display is capable of.

The same thing goes for the whites, although with digital displays setting the white level can be tricky. Take a look at the HTPC calibration thread in my signature (even if you're not using an HTPC), as there is good, basic information on calibrating, and one of my personal experiences about trying to set the white level correctly.


HTH

Very nice pictures for calibrating.
A pity i cant put them on a picture CD and put them in the DVD player.
But maybe if i scale them down to 480p it should work as fine.

It looks very strange though - on my pc screen the 0 IRE black is blending with the background and the same thing with 0 white IRE - the other shows increasing or decreasing brightness. So the PC is correct.

But if i put them on pic cd and insert into DVD player then the 16 bar should blend in and 0-15 not visible right (blend in with background)?
Same thing goes for white only 235-255 should blend into bacground instead.

Feels really strange since it really feels as though you "loose" greyscale but on the other hand a DVD can only reproduce 16-235 levels of black so there is no way to enhance the grey levels ramp since there is really no use to "stretch" the pattern.
Then you will only get strange "scaling errors" like
16 = 0
17 = 1
18 = 3
19 = 4
20 = 5
21 = 7
..
..
234 = 253
235 = 255
and so on since the range will be stretched and there are no "half" IRE steps .
Correct right.


It feels odd though - they should have opted for a bit better greyscale on DVDs but i guess it will be the same for HD material.
0-65535 would be a lot better - like a newly cleaned window with no visible grescale transitions :D

Regards
Boogieman

If your display is set to black=16, and your DVD player passes BTB signals, then turning up the brightness should bring those areas into view (like the THX dropshadow screen). But some players don't pass btb signals - which is why the btb/pluge tests are used to test for this.


But if i put them on pic cd and insert into DVD player then the 16 bar should blend in and 0-15 not visible right (blend in with background)?
Same thing goes for white only 235-255 should blend into bacground instead.


Yes, that is correct. This is theoretical though, especially for the white level. I actually have my Samsung DLP set to 16=black and 255=white when looking at those images. Trying to achieve the brightest white possible on a digital display (which acts differently than a CRT, where you can see blooming when contrast is up too high) and set that to 235 may not be ideal. Like I said, check out my HTPC calibration thread for my anecdote about contrast/white level and Finding Nemo.

Boogie man:

If you search back through this thread, you will see more advanced discussions about setting the black point. The easiest way for me to recommend is to shoot for setting 16 as the maximum black on a digital display, and this will mean lowering the black level *just* until the BTB bars completely disappear. On a DLP this is easy, because it is that level where the the 16 background just stops dithering. However, as the "real" level of black can actually move around a bit around 16, some users may elevate 16 just a touch to reveal just a little bit below 16. This is a user-subjective choice, and I only mention it for particularly picky viewers that may want to play around with this. As Stacey Spears mentioned, some like the JK Samsung actually move the black level around a bit to emulate a CRT, so trying to fudge for floating black level by preserving a little BTB material may be less desireable. In any case, it's a worthy experiment for really anal viewers. Not that being a really anal viewer is the most worthy thing though... ;)

No, 235-255 should not blend into the background, not really. I'm not saying that you should preserve it, but it is *not* synonomous with BTB. It's why I call those levels "peak whites" rather than "whiter than white." They contain white details that SHOULD be seen. However, with a display of fixed and limited contrast range, you have the choice of clipping off some of those details (and there may not be that much there, depending on the mastering) to increase the contrast in most scenes slightly. This is, again, a subjective choice. I do not live with a digital display, so I don't know where I would place my maximum white level, but most likely just it would be close to but not at 255. A little bit more contrast, in trade for a very very minor loss of possible(rare) image data.

Boogie man:
However, with a display of fixed and limited contrast range, you have the choice of clipping off some of those details (and there may not be that much there, depending on the mastering) to increase the contrast in most scenes slightly. A little bit more contrast, in trade for a very very minor loss of possible(rare) image data.


Chris,

You hit the nail on the head (again). Makes me wonder how many calibrationist set TVs to what their test equipment implies without actually going that last step to improve the picture.

Chris:

My apologies if I missed in in the thread and was covered. I am trying to get all my devices set to the same video level so I can properly calibrate them. My video chain is:

Denon 2900 SDI -> Lumagen HDP -> Optoma H79 via DVI

The Optoma DVI is RGB video levels by default we are told. It has a quirk where you can press "resync" and it will change to PC levels. Why?, thatt's another story.

The Lumagen is configurable on the inputs, the output says N/A which based on your description in part 2 of your initial posts I believe I see why.

The Denon is where I'm fuzzy. It has a "Black Level" setting of "Lighter" and "Darker". Lighter is the default, and I'm presuming this has teh digital signals mapped to studio voltage levels. But since this is digital out SDI as I understand your initial descriptions, the setting should have no relavance in a proper design, it woudl just send the digital signals to the Lumagen and be done with it. Alas, if you change this setting, from light to darker, the difference is clear so it is making a change. My concern is if hti sis actually making a studiolevel vs PC level change, or just bumping up the brightness as some consumer "feature".

In both cases I can adjust the PJ to clearly see the BTB 3rd bar on DVE. So it's not clipping it anyway in either case.

Can you advise me where this player should be set? Or how I can test and tell definitavely?

Presuming the PJ really is at video levels and I can set the Lumagen to video levels (not it's default for the SDI input) then I can start calibratiing.

THANKS very much for the help. Scott

Not directly familiar with this player, but the black level option is most likely to adjust the analog outputs between 0 and 7.5 IRE. Yes you are correct that it shouldn't affect the digital output but it is very common for consumer devices to make this adjustment at least in part internally digitally rather than purely at the analog output. You should view a full-range ramps pattern on DVE and check for clipping at both ends of the range as you switch between the two options. You say that you can see the BTB PLUGE bar at your display in both settings, which bodes well. I suspect that one of these settings will show clipping on the ramps pattern, and avoid that one.

Chris: Thanks very much for the quick reply. I went to use the ramps to determine which 2900 (vis ADI) setting clipped if any, and now I can't get it to visibly change. I was sure it did before, maybe I confused myself, I was a little groggy after studying this thread :). I'm going to power cycle the 2900 to be sure it's stable and not a FW glitch that it's not changing now. So in any case I don't see any visible clipping using DVE T12/CH14 ramps.

That's a good thing if it does not change then, and jives with your explanation. Then I will set the Lumagen to Video levels (vs default PC). Presuming the H79 DVI is Video, then all devices should be at video levels and I can calibrate from there.

Then I can move to understanding what to do with a HD STB input to the Lumagen via component.

Thanks again for the quick help. Best, Scott

Are you sure the lumagen defaults to PC levels? I think it defaults to video levels, but it should be quite obvious if you runit through the lumagen. If it doesn't clip at the dots on the DVE ramp which mark 16 and 235 (ref black and white) then it's preserving Video levels correctly. You will need to raise your black level, and lower your white level temporarily to observe clipping clearly.

“Why does data below black even exist? That makes no sense! What can be blacker than black!?”
In video, headroom and footroom is important for a number of reasons. The most basic is that mastering can be less than perfect, so some “fudge-room” has always existed. However, even with ‘perfect’ mastering, data regularly extends outside reference black/reference white. Peak white data allows for highlight details to be maintained, which is common in clouds and other bright objects. BTB data helps prevent image anomalies from hard clipping of the analog waveform at black when converted to analog. BTB can also sometimes become visible as the actual black level on a CRT display floats up and down with image content because black level retention on CRTs is not perfect. The mastering engineer is viewing on a CRT display and actively changing the encoded levels so that they are imaged correctly on the display. This reverse-float compensation in black is allowed with BTB footroom. This reverse float compensation should be the only times data encoded below black is visible in the final image. If you are using a PLUGE pattern with BTB bars to calibrate, you should calibrate so that the BTB data is not visible. BTB data also helps define dithering duty patterns on DLP projectors. Lastly, BTB and peak white data is quite useful for any image processing/scaling etc applied to an image. For all these potential reasons, video engineers advocate preserving full BTB and peak white data whenever possible. Lastly, this data is helpful for any image processing that is applied on an image, such as sharpening, scaling, etc.

While some may say: "Ah I see. now that makes perfectly sense" to me this is not an explanation at all. It appears more to be a rationalization post factum than anything else. Listening to the "great advantages" we might hear next that it cures the common cold as well. :D
Feel free to explain why we would need to lose 12% of contrast resolution for all this.

Let's look at some things you write here:
The most basic is that mastering can be less than perfect, so some “fudge-room” has always existed.

Sorry but I do not buy this argument, imagine we would have the same attitude with sound, so now a 16 bit CD would need some overhead as well because the mastering engineer likes to screw up once in a while.

However, even with ‘perfect’ mastering, data regularly extends outside reference black/reference white.

Really now, if you master with enough bits in the first place you can normalize the data to fit perfectly within the 8 bit resolution space. But even without this, we need a full 16 levels on the bottom and the top to cater for such problems?

Peak white data allows for highlight details to be maintained, which is common in clouds and other bright objects.

So how does this supposed to work then? What if the display device gets say a 253 white?

BTB data helps prevent image anomalies from hard clipping of the analog waveform at black when converted to analog.

And you need a full 16 levels for that?

Sorry Chris, while the rest of your posting is excellent and very informative I find your "explanation" of why we need 16 bits shaved off at the bottom and the top hardly convincing.

Why not instead state the obvious, that the original idea was a bad one, that it would have been far better to have the full 8 bit of resolution but that we simply got to live with it for historical reasons?

Why not instead state the obvious, that the original idea was a bad one, that it would have been far better to have the full 8 bit of resolution but that we simply got to live with it for historical reasons?

this is my understanding why the "incomplete" 16-235 range was decided in CCIR-601 as a standard: The 8 bit range are meant to encode an analogue signal (think of it as a analogue waveform). Ideally the entire wave should be contained within 0-255. However the analogue waveform is not perfect, there are "spikes" "over/undershoots". To allow these overshoots to be recorded, either we find a total "envelope" of all known levels and encode with 0-255 such that all info is within, or we chose from start (or within the process ) this is my "reference black/white level" but I give some allowance over these levels in case there are overshoots. I would imagine the latter is easier, or sometimes, the only option. For example a digital video cam that converts to YCbCr as it records. There is no way it can know for sure the "total envelope" to contain all levels. So it must choose a ref black/while (with some intelligence so not to get the intensity too diff between frames), and give the allowance for "spikes". (This is probably why u find YCbCr values from DV camcorders frequently into the BTB/WTW range, as they are not too smart choosing the ref levels.)

As to why it's 16-235 and not , say 8-247, I have no idea.

c722,

The overshoot data is to accommodate the large (significant edge detail area), that is pre distorted for a (reverse) ringing effect .. on a CRT, & needs this pre distortion to produce "edge" high freq detail ie the edge of a grey or black bar.

I'm conjecturing that the overshoot while useful in a pure analog chain, helps the CRT produce fast transition ON/OFF (CR).<Although>

With digital this info is rounded off and not a factor upstream once encoded.
The rubble (above and below black & white) are 'noise floor' and not codable?
Remember even the waveform is a imperfect representation of the actual sidebands
that are the (real image information) ie some form of PM phase modulation.
Although it's perfect for finding non valid info!

With the nyquist (4 X scf) we have to deal with backward/legacy issues.

thomas

Are you sure the lumagen defaults to PC levels? I think it defaults to video levels, but it should be quite obvious if you runit through the lumagen. If it doesn't clip at the dots on the DVE ramp which mark 16 and 235 (ref black and white) then it's preserving Video levels correctly. You will need to raise your black level, and lower your white level temporarily to observe clipping clearly.Oops, Chris I missed this reply. I'm was getting into the bad habit of reading the email notices from the forum that include the text of replies. It took me a while to nitice the fine print that said "there may be other replies" and so I missed some.

Anyway, yes, with the current firmware, the default "out of the box" HDP setting was level=PC for my particular input (SDI). Yuu have to manually change it to video levels. This made me wonder (again :)) about my levels though, so I'll double (triple) check them. Since I added the SDI I'm always fuzzy about what's coming out of the player as far as levels.

As I understand it, if it's being done the most proper way then:

a) my DVD's SDI output is sending the digital signals freshly decoded from the mpeg stream, right off the mpeg decoder chip.

b) The decoded mpeg stream from the disc should have a data where 16 represents video black and 235 represents video white. The disc creation process could also include data above/below those levels for various reasons.

c) The only way that would be different is if the DVD circuitry or mpeg decoder chip intercepted the data decoded from the disc's mpeg stream, and "adjusted" it to PC levels translating 16-235 scale to 0-255. In that case any data above/below 16/235 is truncated (and BTB/WTW is lost). Then sending it on to my SDI output.

I know you must get tired of folks beating re-stating this over and over. Thanks for your patience.

Anyway, yes, with the current firmware, the default "out of the box" HDP setting was level=PC for my particular input (SDI). Yuu have to manually change it to video levels. This made me wonder (again ) about my levels though, so I'll double (triple) check them. Since I added the SDI I'm always fuzzy about what's coming out of the player as far as levels.

My experience with lumagen was just once, and I believe it was being fed analog signals, no SDI. Anyway, I just vaguely remember that it defaulted to video levels, somewhere, so it's quite possible on SDI this isn't the case. Strange though. With SDI you should be getting the full output from the DVD player without clipping, never seen a player clip when sending out SDI, but it's certainly possible. An easy thing to check with a test disc though!

htpcfan Feel free to explain why we would need to lose 12% of contrast resolution for all this.

I am not going to re-trace this argument, you can do that on your own. I will correct your misunderstandings though. First, these standards are written for the reference display, which at the time, and still currently, is CRT. Professional and high-quality CRT displays have virtually infinite on/off contrast ratio capability. Excursions above their calibrated "white point" are well within their means. Digital displays have very limited on/off CR capabilities, so if you align the display's peak white to the reference white point of the video, then incomind peak whites will be clipped. If you align both the peak white points, the display's and the incoming video, then all that data will be maintained and rendered correctly on the display. The latter situation will lead to lower in-scene contrast ratios as measured between black and reference white, but data above reference white will be included. This tradeoff is a weakness of the display type, as the display has a limitation in on/off contrast ratio. It is not a weakness when viewed from a CRT perspective, which is what the standards are designed for.

Last, you have it backwards in a fundamental respect. Video levels provides *more* dynamic range (but slightly less bit-depth within the reference range) than PC levels. This may seem counterintuitive, but keep in mind that the reference points in PC levels are at the extreme end of the available coding range. These would be aligned to your calibrated white and black point of your display. No data will excurse beyond these points, obviously. With video levels, you align the reference points 16 and 235 to your calibrated black and white points of your CRT, and then there are possible excursions beyond these points. This is actually *greater* contrast provided, because PC levels does not allow for excursions beyond these nominal points.

Sorry but I do not buy this argument, imagine we would have the same attitude with sound, so now a 16 bit CD would need some overhead as well because the mastering engineer likes to screw up once in a while.

Audio is a completely different animal. Headroom is certainly included by any competent mixing engineer, however, to allow for loud transients, and to prevent compression and clipping. But there's no quieter than silence with audio.

So how does this supposed to work then? What if the display device gets say a 253 white?

It displays it, as brighter than 235. You see it as brighter than 235. Like magic. If you clip it off, then you don't see it.

your "explanation" of why we need 16 bits shaved off at the bottom and the top hardly convincing.

Why not instead state the obvious, that the original idea was a bad one, that it would have been far better to have the full 8 bit of resolution but that we simply got to live with it for historical reasons?

It's funny that the computer world is beginning to adjust to this as computers are now becoming appropriate sources for media, video etc. Funny that computer engineers are adapting to the video standards for media, and not the other way around. This room is needed for the floating black and white points of CRTs, as well as the other uses I mentioned. Is that much coding space necessary for below black and peak white? Perhaps not as much any more, however it would be much more difficult to create a consumer standard that provides less toeroom and footroom codes than currently, because Studio standards for mastering are 16-235, and there the ample space is needed as things are being mastered and aligned. Convenience and simplicity carried this over to consumer video. Remember that if you think this is somehow degrading your video, it is the video engineering community who decided that this was the best way to encode and transmit and playback digital video. Every good professional piece of video equipment has no problem doing it this way, and it is simply my goal to help PC users (myself included) achieve this level of performance from their HTPCs.

Audio is a completely different animal. Headroom is certainly included by any competent mixing engineer, however, to allow for loud transients, and to prevent compression and clipping. But there's no quieter than silence with audio.




I for one think this analogy could help here, there is a "noise floor' for audio and video. The spec for audio is (-20dBfs=0VU=+4dBu), now if one encodes audio with levels higher than they should, this compress(s) the data, & it no longer is representing what should be a nominal dBspl. Also the dynamic range is always a struggle with the noise floor.

Also Chris, the noise floor for digital is -90 dB down, so it's not infinite>

For video there is also a 'noise floor' and if you emphasize a nominal level for black or white, then this to is compression, & distortion, and for all I know of the SM , this is not recoverable IMO :D

thomas

thanks to tbrunet and chris for the explanation !

"It's funny that the computer world is beginning to adjust to this as computers are now becoming appropriate sources for media, video etc. Funny that computer engineers are adapting to the video standards for media, and not the other way around."

A fact that's lost on the non-industry computer people.


"thanks to tbrunet and chris for the explanation !"

The two people who you are thanking pretty much disagree with each other on this subject.

The two people who you are thanking pretty much disagree with each other on this subject.

both provide their opinions on this topic, (and on different areas also). This is very informative for me to read and digest.

First, these standards are written for the reference display, which at the time, and still currently, is CRT. Professional and high-quality CRT displays have virtually infinite on/off contrast ratio capability. Excursions above their calibrated "white point" are well within their means. Digital displays have very limited on/off CR capabilities, so if you align the display's peak white to the reference white point of the video, then incomind peak whites will be clipped.

Well that may all be true, but we are now living with modern display devices that range their outputs from black to white. I understand what you are saying but to defend it as a better solution is ludicrous IMHO.
It would be the same as saying that all CDs really ought to be masterd for some old tube amp with 14 bit resolution and that two more bits are for special information that will compensate the dynamic behavour in the extremes of this particular amp.

If you align both the peak white points, the display's and the incoming video, then all that data will be maintained and rendered correctly on the display. The latter situation will lead to lower in-scene contrast ratios as measured between black and reference white, but data above reference white will be included. This tradeoff is a weakness of the display type, as the display has a limitation in on/off contrast ratio. It is not a weakness when viewed from a CRT perspective, which is what the standards are designed for.

You write this completely from a CRT perspective. :)
The behaviour for a CRT with information BTB and the response curve for information above white can hardly be called a great feature. From a modern day display device perspective it is rather abnormal.

Last, you have it backwards in a fundamental respect. Video levels provides *more* dynamic range (but slightly less bit-depth within the reference range) than PC levels.

No I don't.
It seems that you simply did not read what I wrote, I was clearly refering to the contrast resolution and not the dynamic range.

Audio is a completely different animal. Headroom is certainly included by any competent mixing engineer, however, to allow for loud transients, and to prevent compression and clipping.

Feel free how compression and normalization as used in digital audio would be completely different in video. I disagree 100%.

But there's no quieter than silence with audio.

Yes, and it can't be darker than black. What is your point? :confused:

Funny that computer engineers are adapting to the video standards for media, and not the other way around.

I hardly find that funny but that is besides the point.
We have no alternative, as you should understand, practically all source material is encoded this way.
We are stuck with a definition that is a fuzzy as hell, "black is not really black and white is not really white and don't worry on a CRT is will all be fine, just trust us because we are the experts".
Yeah, that is a great help for modern day display devices. :rolleyes:

This room is needed for the floating black and white points of CRTs, as well as the other uses I mentioned. Is that much coding space necessary for below black and peak white? Perhaps not as much any more, however it would be much more difficult to create a consumer standard that provides less toeroom and footroom codes than currently, because Studio standards for mastering are 16-235, and there the ample space is needed as things are being mastered and aligned.

Feel free to explain why ample space is needed.
Mastering can easily be done at a higher bit rate, with compression and normalization one can make a good master that uses the full 8 bit resolution.
You are making it sound as if the video world is still living in the medeval period compared to the audio world.

Remember that if you think this is somehow degrading your video, it is the video engineering community who decided that this was the best way to encode and transmit and playback digital video.

That is my whole point Chris.
But what bothers me is that I see a number of "experts" defending and rationalizing this crap.

Clearly you should understand that in the age of modern display devices this is not the best way to encode video sources.

It is far better to use a higher bit rate to capture the original source and then apply proper compression and normalization procedures to provide a well mastered 8-bit resolution. Just like they do in audio.

Now I realize that that is not going to happen since we aleady have too much stuff encoded the other way, but that does not mean we have to defend and rationalize it.

I mean what is next, a topic of why it NTSC croma encoding is really the right way of doing things?

Every good professional piece of video equipment has no problem doing it this way, and it is simply my goal to help PC users (myself included) achieve this level of performance from their HTPCs.
You obvisously should know that that is simply not true.
The dynamic behaviour of CRTs in the BTB region and the above white level is completely different from modern day display devices.

Bottom line: correcting display device anomalies ought not to be done at the video encoding level but at the display device level.

Main problem: we got everything encoded this way so we got to live with it. :)

I said before in this thread, that this would not be a platform to argue over and over about the merit of maintaining video levels. I have tried, many times, to explain this matter to those who are interested and value proper video reproduction as per standards, and further why the standards are written this way, and even further still why they are not yet at all outdated. I will not continue to argue this, this thread should be to assist those who are trying to get the best out of their system, not another drawn-out argument. I ask that you take the argument elsewhere.

The dynamic behaviour of CRTs in the BTB region and the above white level is completely different from modern day display devices.

I will say that CRT is still the reference, and that some digital displays are designed to emulate CRT behavior, such as Joe Kane's samsung. Not all digital displays have non-moving black and white targets, such as plasmas.

Bottom line: correcting display device anomalies ought not to be done at the video encoding level but at the display device level.

That is your opinion, and you are welcome to it. The video engineering community is overwhelmingly universal in using Studio levels in their Studios (surprise there!) and at the end of the day you can think that they are wrong, but that's what we have to preserve. Is it possible that in the future new digital video standards will arise without toeroom and headroom? Yes, though it is unlikely, not only because of the merits of providing this extra space, but also because of the difficulty of proposing a new standard for video.

I will not continue arguing this here, please take it to another thread or to private messaging. This helps no one.

Man you won't give it up. Here, go argue with Joe Kane. He say's it's there. His disc miraclulously produces it. He made buckets of money selling it to the world. tbrunet is the first I've run across to dispute it.

http://www.videoessentials.com/ve_d_alb_discnotes.php

Man you won't give it up. Here, go argue with Joe Kane. He say's it's there. His disc miraclulously produces it. He made buckets of money selling it to the world. tbrunet is the first I've run across to dispute it.

http://www.videoessentials.com/ve_d_alb_discnotes.php

FWIW: All information below reference black "has" serious gamma distortion, and how one calibrates to distortion is beyond my ability to comprehend :) Joe Kane and Chris need classes in "Video Level 101" and "% IRE"

Also: show me a commercial DVD that has BTB information on it, and I'll show you
a DVD that is worthless, ie "incompetent" mastering!

thomas
:cool: gamma@2.35

I admit to being sympathetic to htpcfan's desire for crisper reference standards in video. There are some advantages to the "computer" way of doing things, especially with newer fixed-pixel technologies that don't have the fuzzy margins that an analog CRT display chain has.

However, it's worth noting that video engineers are not idiots. They didn't dedicate that many codes to below-reference and above-reference without thinking things through very carefully. They did have to deal with the fact that analog video has room for slop in the levels and they didn't have a crystal ball to predict today's modern display technologies. I'm sure if they were designing the system today they would have made adjustments.

There is a disconnect in the video standards between the reference levels and the "real" practical range. Very few people, in my experience, really understand this disconnect, though a careful reading of all the documents helps.

First off, black level. Even though 16 is specified as the "black level" in, for example, BT.601 and BT.709, practically speaking it's not the darkest level a video professional can see on a calibrated Broadcast Video Monitor (BVM). Telecine operators and video editors calibrate a BVM's black level by putting up a SMPTE color bars pattern and using the PLUGE section in the lower right to adjust the brightness. The procedure is to turn the brightness up until the below-reference bar becomes visible, then slowly turn it down just until that bar disappears visually into the background.

Note that this process does not use a light meter. It doesn't use an oscilloscope. It's a perceptual process. Furthermore, that below-reference bar is encoded at 4% below reference, which is digital level 7.

Consider that for a moment. Even disregarding the perceptual nature of the calibration process, the established process for calibrating for video actually sets the blackest perceivable level at somewhere between 7 and 16, depending on how good the calibrator's eye is at distinguishing between levels. Not 16.

This means in general that some of the levels below 16 are still visible if the calibration is done correctly. Stacey and I have done some informal tests that suggest that the actual number of visible levels below 16 on a calibrated BVM is around 4, but we need to do more work to get a more accurate number. I suspect (purely an educated guess) that it may be as much as 6 or 7 levels under the right circumstances.

And just as obviously, the levels above 235 are completely visible, because BVMs are CRT displays and have lots of inherent headroom.

A good video technician is not going to let large amounts of picture information go above 235, but all they care about is that the bulk of the histogram stays below that level - they are perfectly OK with small spikes above the reference level, which happens with specular highlights, details in bright or highly saturated objects, etc. Those are exactly the kind of details that headroom is designed to handle.

There are people who master DVDs who run the video through a "legalizer" before finalizing the master. The worst of these "legalizers" clamp the video levels (in YCbCr) to 16-235. Most don't - most just ensure that there aren't huge swaths of above-reference or below-reference values, and to do so apply a non-linear compression to the values with a smooth knee and shoulder (essentially an S-shaped gamma), and don't compress everything into the 16-235 range. These are completely unnecessary for DVD mastering but people have gotten used to using them because television transmitters have stringent range requirements.

Moreover, even with the crappy legalizers that clamp everything into the reference range, the RGB range can (and does) still stray outside the reference ranges, and it's ultimately RGB that matters. I have yet to see a DVD that didn't have above and below reference RGB values. And those values are visible on a properly calibrated CRT (such as a BVM).

Some companies (the good ones) don't apply legalizers at all for DVD mastering. There's really no point. There's nothing inherently wrong with picture information extending above or below the reference levels. That's the way the whole system was designed. In practice, very little of the image will typically extend above or below, but that's to be expected. Most real images have very little data at the very high or low end of the range. Unfortunately, even though it's not a lot of data, it can be fairly important data.

There really shouldn't be any controversy about this - we all are (I hope) trying to get the picture on our home theater screens to look as close to the picture the director and/or cinematographer approved. The director or cinematographer supervised the transfer on a calibrated BVM connected to pro equipment. That equipment doesn't clip off the above or below-reference range, and neither should consumer equipment.

Don

Don,

Y= 16-235 with a offset of "16" and an excursion of 219

CbCr= 16-240 with an offset of 128 and an excursion +/- 112

There are 256 discrete levels of "Chroma" , and NOT 256 levels of luminance!

FWIW I must be one of the idiot engineers you are refereing to! Also headroom was built into the digital payload, it's called "Peak White" and it's bound at 235.
According to you theres NO offset? I have read Poynton and I'm going to quote him when he states!

"The headroom and toeroom are NOT PICTURE INFORMATION"

From what I've read so far, 75% of the original R'G'B' information is lost by the time
it get to us in DVD form! It's called compression :)

thomas

Don,

Y= 16-235 with a offset of "16" and an excursion of 219

CbCr= 16-240 with an offset of 128 and an excursion +/- 112

I'm with you so far. Those are the reference ranges. The encoded levels are allowed to stray outside those ranges. That's why there are more codes.

There are 256 discrete levels of "Chroma" , and NOT 256 levels of luminance!

I agree. There are 254 discrete levels of luma because codes 0 and 255 are reserved. How this is relevant is beyond me.

Also headroom was built into the digital payload, it's called "Peak White" and it's bound at 235.

601 calls it "peak white" but Poynton does not. I'm going to go with Poynton:

Digital Video and HDTV, Charles Poynton, pg. 22:

"Headroom allows code values that exceed reference white; therefore you should distinguish between reference white and peak white." (italics in original)

According to you theres NO offset?

Huh? When did I say this?

I have read Poynton and I'm going to quote him when he states!

"The headroom and toeroom are NOT PICTURE INFORMATION"

I've searched through Poynton and can't find anything remotely resembling this quote. Could you provide a cite, please? Book? Page number? Web site?

All this doesn't really matter - I've looked at hundreds of DVDs at this point. I'm confident in saying that ALL DVDs contain information coded above 235. And ALL DVDs contain information coded below 16. You can scream and shout about how it's wrong (and I disagree, as I've made clear), but that's the way it is.

As for the codes below 16, many of them are invisible and were invisible on the BVM, and aren't supposed to be seen. I would still like to keep that range, partially to allow proper calibration and partially because some of them were visible on the BVM.

I have had it demonstrated to me that clipping below 16 is visible on real-world DVDs. It's subtle, but it's there. So all the theory about "why" is kind of immaterial. It actually does change the picture to clip the below-reference values.

The stuff above 235 is much simpler - ALL of it is visible. ALL of it was visible on the BVM. The director approved it. It shouldn't be clipped off.

Don

This started out as a Reference thread providing information. Somehow it is not side tracked. Please do not post questions to this thread.
I will be cleaning this up when I get a chance and removing some of the off topic posts.
Thank you.

This started out as a Reference thread providing information. Somehow it is not side tracked. Please do not post questions to this thread.
I will be cleaning this up when I get a chance and removing some of the off topic posts.
Thank you.


Awesome thread, helped me a lot! :D

I literally studied this thread a while back. And learned a great deal from it thanks mostly to Chris's help. I don't remember this issue being discussed. IRE vs % yes, but not the fact that Avia Pro and DVE pro have levels that are different based on IRE and % respectively.

A fly in the ointment when I when to use DVE Pro with software expecting IRE levels:

Probably not appropriate for this thread, but certaintly related I think. So here's the link to a new thread to see aht I'm talking about:

http://www.avsforum.com/avs-vb/showthread.php?p=7044635&&#post7044635

grrrr. Another lack of standards.

Here is a simple question. If I have an lcd projector and an iScan Ultra, is a dvd players inability to pass blacker than black a deal breaker? How much of an issue is it in practical terms?

EDIT - ok, for the sake of not carrying on in this thread, thanx for the feedback. I still don't understand exactly how the blacker than black signal works in practice.

Is the btb signal really mostly only "useful" when you set your calibration and then after that it has no effect? Or is the btb signal actually involved in preventing the clipping etc during actual playback? Is it easy to tell if a player has btb and what do you notice if the player does put out have btb?

It sounds like a good thing which can really help in calibrating, but something which is not the end of the world picture wise. I suppose I will try to check the dvd players I have right now to see which units pass btb. I will be pissed if the ones I have now pass btb if the one I am getting does not....but then again, the one I am getting is really only for region free and dvd/audio sacd and other format use.

Any more simple bottom line info about the implications of btb would be greatly appreciated.

thanx chriswiggles [ PS enjoying your Marantz receiver? I still enjoy the ole SR 8000 :) ]

It's not preferred if you can use a setting that doesn't clip BTB, or if you can get a better source, but you should still have a fine picture, especially with an LCD I wouldn't worry about it too much at all.

Say, what the heck is an IRE?
As often as the term IRE is thrown around when discussing video, relatively few actually know what it means!
An IRE simply a representation of volts:

Yourself included :p

The IRE scale is not absolute, it's relative...and also it's a measure of the
gray scale or luminance. "IRE are NOT simply a representation of volts"!

Yourself included :p

The IRE scale is not absolute, it's relative...and also it's a measure of the
gray scale or luminance. "IRE are NOT simply a representation of volts"!

Sorry Thomas but you are wrong - do we have to get into quoting Poynton? IRE is referenced back to Video signal voltage range.

White sits at ~714 mV, Black (for NTSC systems) sits at ~53 mV. If that isn't a direct relationship I don't no what is.

Amplitude units for video signals, ranging from 0 for the blanking level to 100 for pure white. For NTSC signals, one IRE unit corresponds to 7.14 mV.
http://zone.ni.com/devzone/nidzgloss.nsf/webmain/B7A03F495D5B30B5862568A100781B2A

ted

From Snell & Wilcox glossary:
“Acronym for the North American 'Institute of Radio Engineers'. Also refers to the measurement units introduced by this organization. As defined by ANSI/IEEE standard 205, 'IRE units are a linear scale for measuring the relative amplitudes of the component.”

From Tektronix Standards overview:
“Eventually the IRE (later to be the IEEE) established a unit of measure for video signals. This "IRE unit'' was defined as 1% of the video range from blanking to peak white, without reference to the actual signal voltage. Although defined as a ratio,…”

Oh yea, one other term that I have been accused of misusing…
From Tektronix glossary of useful terms:
"GAMUT-
The range of voltages allowed for a video signal, or a component of a video signal. Signal voltages outside of the range (that is, exceeding the gamut) may lead to clipping, crosstalk, or other distortions."


Dave

Thomas,

I've no problem with the use of the term "relative" after all an IRE is a "unit". However within the context of this thread what do you think that unit should be?

Offer another unit outside of its applicability to NTSC that might be useful here.

Since as you claim "it's a measure of the gray scale or luminance" then state the unit measurement of that graphing, and while you are at it tell me how many Ft Lamberts, Ft Candles, Lux or Lumens are represented by 100 IRE?

As a master of tangential thinking you should be able to find something, and when you do please explain it within the context of this thread.

btw did you notice the '~'? that means "nominally" in that context.
ted

Say, what the heck is an IRE?
As often as the term IRE is thrown around when discussing video, relatively few actually know what it means!
An IRE simply a representation of volts: there are with 140 IRE units in one volt. This simplified expressing the ~700mV excursion of an analog video waveform, with white represented using 100 IRE instead of an odd 714mV. As I will describe, IRE units are ambiguous at describing the intended image content unless you know whether or not the 7.5 IRE setup pedestal is present. Many do not grasp that IRE is merely another way to represent volts, and as such confuse IRE into some mysterious value that magically describes the actual image information, which cannot do without caveats.


Sorry Chris, this entire section is WRONG! I already corrected you, yet you continue to discuss things you don't comprehend!
"IRE is a relative linear scale. It doesn't refer to any particular voltage or digital level until you specify the signal standard being used."

thomas

I've searched through Poynton and can't find anything remotely resembling this quote. Could you provide a cite, please? Book? Page number? Web site?


http://www.poynton.com/PDFs/Merging_RGB_and_422.pdf

“Test signals characterize the electrical performance of a video system. Standard video test signals include elements that are synthesized elec-tronically as sine waves, and injected onto the signal. Many of these elements have no legitimate R’G’B’ representation. Since these signals can be conveyed through Y’C B C R without incident, some people claim
Y’C B C R to have an advantage. However, in my opinion, it is more important to allocate bits to picture information than to signals that cannot possibly represent picture information.”

I have edited the original post to update the links, they had been swept into the archives.

I have also added a link to:
http://www.sigmadesigns.com/public/Support/chromaticity.html

In case anyone missed that link, it has an excellent series of images at the bottom of the page which illustrate correct color decoding, and then incorrect color decoding with a 601 to 709 mismatch and a 709 to 601 mismatch. I had originally refrained from including a link to the site because the image captions were confusing and ambiguous, but Keith has graciously changed the captions to make them crystal clear. It illustrates well the kind of green push and green depression people may run into, especially on sources where you can't get colorbars on (like HD boxes etc) which would indicate unambiguously a color decoding error.

This is a great resource.

2) When using digital outputs, the major adjustment option you have is the one for digital levels. As is common in consumer labeling, the labels for this can be confusing. The most common labeling will at least hopefully show that you are making an adjustment to the DVI/HDMI digital outputs, and usually the options will read ‘Normal/Enhanced’ or ‘Normal/Expanded’ or ‘Video/PC,’ or some such label. The latter is clearer, as this adjustment is choosing whether or not the digital image data is correctly output using Studio (also called Video) levels, or is incorrectly re-mapped to PC levels. You should choose to maintain Studio levels by checking to make sure this option is properly set. Usually the default setting will correctly choose the option for Studio levels. Check to make sure.

in the above paragraph (found from part 2 of Chris' awesome write-up), there appears to the a typo. It appears that the intent is to maintain Studio levels, so am I correct in assuming that you should always select "Normal" or "Video" on your DVD player? If so, don't you mean to say "the former is clearer" rather than as printed "the latter is clearer" (4th sentence in paragraph)?

I think he is referring to the phrase ""Video/PC" as opposed to the phrase"Normal/Enhanced(or expanded)"..

IOW he was referring to the description, not the result, as being clearer.

Hopefully, the author will reply and clear things up.

I think he is referring to the phrase ""Video/PC" as opposed to the phrase"Normal/Enhanced(or expanded)"..

IOW he was referring to the description, not the result, as being clearer.

Hopefully, the author will reply and clear things up.

That is correct. I can see how that could be confusing. You want to be choosing the equivalent of "video" or "normal" in most all cases, whatever the label may be. I was referring to the labeling of "video/PC" as just being clearer labeling terminology. I'll edit to make that more clear though beecause it is unclear. Thanks for the heads up!

No reason for a bump, thread is a sticky. :)

larry

Nice guide!

Im 32 years old. How did I survive all this long without reading this?

I would love to say "I DONT CARE", but I do care and this thread was a great help!

This thread was referred to on a current thread as a source of information that would answer another posters question. So I have read it and find it does not appear to agree with what I have read elsewhere, so thought I would ask why?

From the original post:

"I will refrain from calling them ‘whiter than white’ because this implies that they shouldn’t normally be present or visible in the final picture. They should be, unlike BTB"

"In a system that outputs black at 0 IRE, BTB data will be output at voltages slightly below 0mV (simply negative volts). If you’ve digested that correctly, you realize that BTB data can be maintained in BOTH situations"

Below is my limited understanding, is it correct?

In the Pal analogue signal there is a blanking interval to give the display time to move to the next line or the next frame of the image. In this blanking time there are sync signals. Also during this blanking time the analogue signal is clamped to 0volts by using a capacitor. Because if it is not clamped the signal will wander and the black level in the image will not remain constant (since white is created by higher voltage, a bright image could drag the bottom of the signal up from 0v). If this clamping extends into the part of the signal that syncs black it is called a black level clamp, PAL black is 0volts so it is being clamped to 0volts.

I do not understand how if the above is correct, why blacker than black should or could be passed through the system. My understanding is it was neccessary in analogue systems because they might not have very good clamping of the signal or it might wander up as it displayed a image line with a high average brightness, so some tolerance to black level wandering was needed. Whiter than white information is also I think present in analogue systems because signal level could wander up, at the begining of the line whiter than white would equal white, but at the end of the line white could equal white, if whiter than white was not present in the signal white would be grey at the begining of the line. So my understanding was blacker than black and whiter than white are holdovers from the analogue age when image brightness might not hold steady across the image, you might need to drag an analogue wave signal down to black or up to white when the voltage of that signal could be wandering about.

In a perfect CRT I think the idea is that it is not needed as the signal level does not wander, unfortunately people did not have perfect CRTs so it was present in the analogue signal, but on a perfect CRT the information would not be displayed, in effect it would be clipping below black information and whiter than white information would be white anyway. Calibrating to incorporate blacker than black and whiter than white apears to me as calibrating to be equivelent to the worst crt display it was designed to cope with. I thought they only existed in digital as a legacy to analogue and served no usefull purpose with digital displays.

If blacker than black is desirable up to the digital non-crt display, and according to some displaying some blacker than black information is desirable. What visible negative impact does it have if the source is clipping it and are sources that clip it failing to meet the consumer spec and of a faulty design for doing this.

If whiter than white if it is desirable to actually display. What negative impact on image quality is there from clipping it and again are sources that clip it failing to meet the consumer spec and of a faulty design.

If I do not display blacker than black or whiter than white I gain contrast, which improves perceived image quality. What do I gain if I display them.

I am confused...
My player Pioneer DV-600 has an option to show BTB and WTW.
How should I configure it? The readme of the firmware that allows this option says: Enable BTB, put the players Brightness to +1 and Contrast to -1. So it gives BTB and WTW info.
If I do this I can see BTB and WTW bars from GetGray DVD. Should I calibrate my tv to see or not to see BTB?
Thank you!

This thread was referred to on a current thread as a source of information that would answer another posters question. So I have read it and find it does not appear to agree with what I have read elsewhere, so thought I would ask why?

From the original post:

"I will refrain from calling them ‘whiter than white’ because this implies that they shouldn’t normally be present or visible in the final picture. They should be, unlike BTB"

"In a system that outputs black at 0 IRE, BTB data will be output at voltages slightly below 0mV (simply negative volts). If you’ve digested that correctly, you realize that BTB data can be maintained in BOTH situations"

I have a PAL dvd player connected via analogue component to a dlp projector.

Below is my limited understanding, is it correct?

In the Pal analogue signal there is a blanking interval to give the display time to move to the next line or the next frame of the image. In this blanking time there are sync signals. Also during this blanking time the analogue signal is clamped to 0volts by using a capacitor. Because if it is not clamped the signal will wander and the black level in the image will not remain constant (since white is created by higher voltage, a bright image could drag the bottom of the signal up from 0v). If this clamping extends into the part of the signal that syncs black it is called a black level clamp, PAL black is 0volts so it is being clamped to 0volts.

I do not understand how if the above is correct, blacker than black should or could be passed through the system. My understanding is it was neccessary in analogue systems because they might not have very good clamping of the signal or it might wander as it displayed the image line, so some tolerance to black level wandering was needed.

My display has controls for setting the clamp width and position. Also brightness and rgb bias settings. I think the correct settings are to not display any blacker than black information. If the clamp is acting as a black level clamp and the signal does not wander (which displaying a 0% black screen with +4% +2% -4% black bars it probably will not) I think there should be no blacker than black information left after it.

Whiter than white information is also I think present in analogue systems because white level could also wander and display white level may not be consistant with varying image brightness. In a perfect CRT I think the idea is that it is not needed, unfortunately people did not have perfect CRTs.

In a display that excepts an analogue signal but then converts it to digital to display it, like my DLP projector, I do not understand why you would want to display whiter than white information.

The blacker than black and whiter than white information I thought existed to cope with analogue signal wandering on a analogue display - crt. Calibrating to incorporate it strikes me as calibrating to be equivelent to the worst crt display it was designed to cope with.

On a digital display by enabling any blacker than blakck or whiter than white information to be displayed you are reducing image quality by lowering the contrast ratio of the desired black to white information.

Since this view seems to contradict what was in the original post. Wanting to have blacker than black information pass through the system, and then be lost by setting the display brightness and some posts saying you want to display a little blacker than black infomation. Wanting to have whiter than white information displayed. Can someone explain why?

Let's see how much I remember from all this stuff when it was the hot topic. For analog video, you usually have the choice to have "black" be 0 or 7.5 IRE. IRE voltages are not absolute, they're sorta like "voltage differences". I'm not sure if the PAL standard is for 0 IRE to be black or not, but then again that is not necessarily 0 volts (or mV). 0 IRE can be +2mV as long as both sides "agree". So when the player outputs a "level" for black, the display recognizes it as black and displays it. That's my best shot remembering off the top of my head. You may want to PM Chris Wiggles and point him at your post. He should be happy to explain.

larry

I am confused...
My player Pioneer DV-600 has an option to show BTB and WTW.
How should I configure it? The readme of the firmware that allows this option says: Enable BTB, put the players Brightness to +1 and Contrast to -1. So it gives BTB and WTW info.
If I do this I can see BTB and WTW bars from GetGray DVD. Should I calibrate my tv to see or not to see BTB?
Thank you!
Sounds like a load of BS to me. In digital video, "black" has value 16 (for luma in component video). So, on a DVD where there is "black" the digital value 16 appears. If the DVD player works correctly, the value 16 will be output on the HDMI cable to the display. If you increase brightness +1 in the player, it will take the value 16 and add 1 (or 2, or 3, or more depending on how the brightness control is "tuned"). So, now the 16 (black) is no longer output as 16, but as 17, or 18 or some other number higher. Your (properly calibrated) display sees it and it does not output black, but something lighter.

Ok, so now if you have BTB on a disc, eg. 15 or 14 or lower, the "player brightness at +1" will cause the value on the disc to be output at 16 or 17 or something. BTB is not being output, but some brighter shade of black/gray. Changing brightness (or contrast, or any other color/tint) on the player modifies the digital video values coming from the disc. That's not why you buy a DVD player, you buy it to output what is on the disc. :)

The way to tell is BTB is being output by the player is to properly calibrate your display so that "black" (16) is "black" (no light) on your display. Once you do that, anything at 16 or below will show as "black". If you want to see if BTB is passed by the player, view a test pattern with BTB information. Then raise the brightness on the *display* and the BTB information should be seen after raising brightness a tick or two or three. If you don't see the BTB bar or whatever, the the player is probably not sending BTB data. Raising the brightness on the display takes the BTB (eg 15) and adds adds 2 or more to it so that it is 17 or 18 allows them to be seen. In fact, you should be able to lower brightness on the *player* and still be able to see BTB video on the display by raising brightness. It's all just math. Subract some on the player end and if you add enough on the display end it will be visible. Most players don't modify the video when controls are in the "neutral" position, but lowering the brightness on the player will guarantee it is passing BTB video if you can see it on the display after raising brightness. If the player is not configured properly at the factory, it could possibly add (or subract) something to the video data value. That's why we have the nifty calibration discs.

BTB video should not be seen on a digital display when properly calibrated. "black" should be set to "panel off/no light" so it's kinda hard to display something darker. :) BTB is necessary for scaling and to help maintain proper black levels for CRT displays if the BTB data is on the disc.

Most digital displays can output WTW when calibrated correctly. And if the player can output BTB then it is most likely outputting WTW too. If no BTB is output, then WTW is probably clipped also.

larry

Thank you PooperScooper!
I did the changes I wrote (players brightness +1, contrast -1) (that is the only way to pass BTB and WTW) and reduced my tvs brightness and increased my tvs contrast. I calibrated so, I can't see BTB, but now I can see WTW and the picture seems a lot better than before.

I have a Pioneer Elite DV-58AV with three HDMI Color settings:

Full Range RGB = 0 Black to 255 White
RGB = 16 Black to 235 White
Component = ? Black to ? White

I wish to know the missing values for the Component mode. If the Black to White range is the same as the RGB mode, then what's the difference between the Component and the RGB modes? Thank you.

I have a Pioneer Elite DV-58AV with three HDMI Color settings:

Full Range RGB = 0 Black to 255 White
RGB = 16 Black to 235 White
Component = ? Black to ? White

I wish to know the missing values for the Component mode. If the Black to White range is the same as the RGB mode, then what's the difference between the Component and the RGB modes? Thank you.

Isn't component going to be YPbPr (http://en.wikipedia.org/wiki/YPbPr) instead of RGB? I'm not sure if the luma channel has an extended vs studio range corresponding to RGB.

-Bill

Isn't component going to be YPbPr (http://en.wikipedia.org/wiki/YPbPr) instead of RGB? I'm not sure if the luma channel has an extended vs studio range corresponding to RGB.

-Bill

I believe YPbPr is analog, and YCbCr is digital, but they are both Component Video levels. I am curious about the digital numbers from Black to White.