I am sorry if this has been answered before, I tried searching for it, and found a couple of threads dealing with the topic and useful test patterns, yet they didn't mention how the test patterns were supossed to be used.



I have DVE BR version and there is a test pattern for this too, and still I don't think they explain what I am supossed to look for. So if you please be so kind to teach me how to use it, I'd appreciate it.


Thanks.

Edo

Greetings

Please be more concise with what you are asking. I have no idea what you are talking about.

Regards

I mean, how to know if the player is using either 601 or 709 color space.


Thanks for your quick reply.

Greetings

If it is the restaurant scene ... then the clip shows you what the color palette looks like when you map 601 to 709.

I'd assume your TV is doing 709 when you hook it up. Unless you explicitly told it to pick 601.

Regards

Actually there is a color bars pattern, the upper half of the screen is marked as 601, the lower part is 709 (or it could be the other way around).

How is this pattern supossed to be used?

Greetings

Well the Blu player outputs in 709 ... to the TV which should be 709 too.

If you use a blue filter and look at the 601 patterns, they should not look correct through the filter. Only the 709 patterns should look right.

Regards

Here's a link to the PDF document of program notes from the 'DVE' HD DVD/SD Combo disc: http://www.videoessentials.com/dvehd/images/DVE_HD_Program_final.pdf
That program has a few more patterns than the Blu-ray version of 'DVE: HD Basics,' but there is much that is the same in both titles. These program notes have a lot more detail about the patterns than the documentation associated with 'DVE: HD Basics.' Perhaps you can find better guidance on how to use the various test signals on your version.

Best regards and beautiful pictures,
G. Alan Brown, President
CinemaQuest, Inc.

"Advancing the art and science of electronic imaging"

Thank you guys and thank you George for the very useful document.

Thank you guys and thank you George for the very useful document.

There really are 2 issues... there is a color decoder and there is a color space.

The color decoder comes into play when YCbCr or YPbPr are used. YPbPr is analog component, 3 separate video cables used for the connection. YCbCr is digital component. This is almost always used in consumer video products to send video from one component to another over HDMI or DVI. Displays have to display images in RGB, ultimately. Source data is often stored in RGB on discs and has to be converted to YCbCr, sent to the display, and converted from YCbCr back to RGB. The ENCODE/DECODE process is where the color decoder (and encoder) come into play. If those are done poorly, the colors will not look right on the display. You have to use a color decoder test pattern to determine if the color decoding is being done correctly. If it is NOT correct, it could be the display or the source component (assuming the source component is NOT a video signal generator, a piece of test equipment that has to do this right). The 601 and 709 specifications define how the color decoder SHOULD work (and the encoders). It is up to manufacturers to do this right.

Color Space is a whole different deal - these define the reference coordinates for primary and secondary colors. There are many different color spaces... CIE xyY is one you probably see a lot, but there's also uvL, CIEL*A*B*, and a hoat of others. Reference coordinates in one color space can be converted to coordinates in any other color space. HDTV color space is defined in the 709 spec. Standard definition color space is NOT defined in the 601 spec. You have to go to SMPTE-C specifications for standard definition (NTSC) color space references.

In general, I would have to say that the visual difference between using the wrong color space is minimal. But using the wrong color decoder is quite easy to see. I have a video display that allows you to switch the color DECODER (aka color matrix) between 601 and 709 for each resolution and each input. So I can send a DVD like the first Austin Powers movie and see natural/correct colors when the 601 matrix is selected, but the reds on the double-decker London buses and the red on the UNDERGROUND signs are clearly too orange when the 709 matrix is selected. Similarly, 1080 sources look fairly obviously wrong when you apply the 601 matrix.

But using the right matrix (decoder) with the wrong color space (i.e. SD color space selected for an HD source or vice versa) is quite difficult to detect - partially because most displays' color space(s) are not accurate for HD or SD sources! And partially because the actual difference in "size" of HD color space isn't that big of a difference from SD color space. Errors in the display can be larger than the difference between HD and SD color space references for primaries and secondaries.

There aren't a lot of displays that let you manually switch the decoder/matrix. There aren't a lot of displays that let you manually select SMPTE-C or 709 color space. There are even fewer displays with accurate primaries and secondaries in either color space. And there are even fewer displays with controls that allow you to make the primaries and secondaries accurate. In the end, you are left with a very small number of displays (and sources) with accurate encoding/decoding, accurate primaries and secondaries, and accurate + switchable color spaces.

Wonderful explanation! Most folks who read this are now going to want to know which displays "have all their ducks lined up," as you described above. The only consumer displays I know off the top of my head, that do color right, are Joe Kane's Samsung projectors. What others are there that you can identify for us?

A quick question, when I run my Oppo 981 @ 1080p to my XBR5, should I set the color space on my XBR5 to 601 and calibrate that way? What about on the Oppo itself (YPbPr or RGB)?

A quick question, when I run my Oppo 981 @ 1080p to my XBR5, should I set the color space on my XBR5 to 601 and calibrate that way? What about on the Oppo itself (YPbPr or RGB)?

Edit: What I wrote below assumes that the Sony setting is a color decoder/matrix setting and not a color space setting. If it is a color space setting then setting it to 601 I would assume means the SMPTE-C defined color space. You would have to measure to see how close they actually are to the standard. Since DVDs should be mastered using SMTPE-C then 601 would be the right setting but it really depends on the measurements. What is the exact name of the control/setting in the Sony menu?

Assuming it is a color decoder/matrix setting...

I have an Oppo 980 and the Oppo's should be handling things properly. Setting it to RGB will mean that the Sony does not need to do a YCbCr->RGB conversion (although it could be doing an RGB->YCbCr and back again internally depending on how things are implemented and what controls are accessible to you such as tint). Setting the Oppo to one of the YCbCr options will mean that the Sony will need to do the YCbCr->RGB conversion. What I'm not sure of is if the Oppo is "twisting" 601 YCbCr to 709 YCbCr before sending things out or not when it is configured to output a HD signal (720p or 1080i/1080p). If it is doing this twist and you set the Sony to 601 then things should look incorrect and you should get somewhat different color brightness measures between that and the Oppo sending out RGB. If you set the Sony to 709 then measurements then should look pretty much the same as to the Oppo sending RGB.

In my case, I have my Oppo hooked up to a Samsung LCD and I get the same measurement results no matter if I have the Oppo set to RGB or to YCbCr. When I set it to RGB, I still have access to all the user settings so it may be that the Samsung is doing that extra RGB->YCbCr->RGB step internally - I'm not sure.

hope this helps,


--tom

Yeah, the control for 601/709 is called "Color Matrix". There is also a "Color Space" setting that I leave on "Standard" (the other option there is "Wide").

From what I understand the Oppo players do the "twist" correctly, and if that's the case, can I leave the Oppo to the YCbCr and leave the Sony on "Auto", which applies the 601 color matrix for SD sources and the 709 for HD? What color output (RGB or YCbCr) should I set my PS3 to output with Blu-rays?

I apologize for the "newbie" questions, and I know my knowledge of such stuff is extremely limited. I'm just trying to figure out where my menu settings across my components should be set before I try my first amateur calibration with an EyeOne early next week.

Source data is often stored in RGB on discs

DVD and Blu-ray do not store video as RGB.

what I understand the Oppo players do the "twist" correctly, and if that's the case, can I leave the Oppo to the YCbCr and leave the Sony on "Auto"

It sounds like that's correct. I believe in other posts that you might have some way to take measurements. If that's the case then you can look at Y measurements for colors to check that things are handled correctly. If the decode is off then the Y measurements on colors will be clearly off. It's a simple concept if you understand the basics and http://www.avsforum.com/avs-vb/showthread.php?p=13934963#post13934963 or some of my later posts in that thread are mostly about looking at Y measurements for colors.

From what I understand the Oppo players do the "twist" correctly, and if that's the case, can I leave the Oppo to the YCbCr and leave the Sony on "Auto", which applies the 601 color matrix for SD sources and the 709 for HD?

If the Sony is doing things correctly then that sounds like the right settings. You could also just tell the Oppo to send RGB - but you'll have to see if the Sony then locks out any user controls such as Tint, etc.

What color output (RGB or YCbCr) should I set my PS3 to output with Blu-rays?

I think it is pretty much the same here. The PS3 I would assume is also doing the twist correctly when it is reading a SD DVD and outputting a HD signal to the display. If so, then you could set it to YCbCr and Auto on the Sony and it should work just like with the Oppo. You could also set it to RGB and you should get the same results. I'm just not sure what the Sony will lock out (if anything) if you send it RGB over HDMI.

As Alluring mentioned, measuring may be the best way to determine exactly what is happening. If things aren't working correctly, then your Y readings will be different as you toggle the settings - especially green. So, once you get your i1 and are ready to take some measurements you can do some experiments

So, what I would do is:

(1) Set the Oppo to YCbCr and set the Sony to Auto

Do a primary/secondary measurement and save it.

(2) Switch the Sony to 709

Measure colors and compare to the first result.

(3) Switch the Sony to 601

Measure colors and compare to the first result.

(4) Switch the Oppo to RGB

Measure colors and compare to the first result.

If things are working correctly then I believe results 1, 2, and 4 should all be pretty much the same within variance. #3 should give you different Y results.

You could then repeat the experiment with the PS3.

hope this helps,


--tom

Wonderful explanation! Most folks who read this are now going to want to know which displays "have all their ducks lined up," as you described above. The only consumer displays I know off the top of my head, that do color right, are Joe Kane's Samsung projectors. What others are there that you can identify for us?

For the record, I don't believe the Samsung's let you choose the color decoder, only the colorspace. Correctly calibrated, it does both accurately though, but if your source messes up the color decoding, there is no way of correcting this inside the PJ. It's an accurate display, designed to work correctly with an accurate source, not designed to let you compensate for inaccurate sources. The good thing about this is, that if something doesn't look right, you'll know it's the source that is to blame, so fixing the issue is fairly straight-forward (buy a better source!).

If the Sony is doing things correctly then that sounds like the right settings. You could also just tell the Oppo to send RGB - but you'll have to see if the Sony then locks out any user controls such as Tint, etc.

Also, you'll have to check if the Sony actually displays the RGB correctly, without messing up the decoding by converting to YCbCr and back, using the wrong matrixes. I have seen Sony LCD's that do this (along with a lot of other displays outthere...)



I think it is pretty much the same here. The PS3 I would assume is also doing the twist correctly when it is reading a SD DVD and outputting a HD signal to the display. If so, then you could set it to YCbCr and Auto on the Sony and it should work just like with the Oppo. You could also set it to RGB and you should get the same results. I'm just not sure what the Sony will lock out (if anything) if you send it RGB over HDMI.

I believe the PS3 is clipping BTB and WTW when using RGB, regardless, of the full/Limited setting in the player. Haven't testet thoroughly, though...

As Alluring mentioned, measuring may be the best way to determine exactly what is happening. If things aren't working correctly, then your Y readings will be different as you toggle the settings - especially green.

If you're used to seeing what an accurate SMPTE pattern looks like, it should be farily easy to see by eye. The 601 will look quite dark, when the pattern is displayed correctly. If the display handles the input as 601, the 601 part will look right, and the 709 part will look almost neon. However, some displays have "odd" decoders, that are neither 601 or 709 exactly, and to check this you will have to measure (I don't trust filters much, as I have seen quite varying results from certain panels, especially Pioneer's).

Here's a link to the PDF document of program notes from the 'DVE' HD DVD/SD Combo disc: http://www.videoessentials.com/dvehd/images/DVE_HD_Program_final.pdf
That program has a few more patterns than the Blu-ray version of 'DVE: HD Basics,' but there is much that is the same in both titles. These program notes have a lot more detail about the patterns than the documentation associated with 'DVE: HD Basics.' Perhaps you can find better guidance on how to use the various test signals on your version.

Best regards and beautiful pictures,
G. Alan Brown, President
CinemaQuest, Inc.

"Advancing the art and science of electronic imaging"

Is the flip-side of that DVD(the SD program) intended for non-HD consumers like myself, or is that digital SD intended to be read only by HD-DVD players?

Wonderful explanation! Most folks who read this are now going to want to know which displays "have all their ducks lined up," as you described above. The only consumer displays I know off the top of my head, that do color right, are Joe Kane's Samsung projectors. What others are there that you can identify for us?

I can't think of any others I know from first-hand experience that do everything right. I'm still waiting to get my hands on one of the THX Certified video displays to see if they do everything right or if there are still "holes" in how sources are handled.

Some have come tantalizingly close... Pioneer Kuro plasmas and the new Samsung plasmas (PN-xxAxxx) which aren't as black as the Kuros, but have VERY accurate color and can be calibrated to get Gamma up to 2.4 (plasmas as a group struggle with getting gammas that high). But the Samsung has no manual settings for matrix or color space - Pioneer does offer 2 color spaces, but you need calibration to get the 2 color spaces accurate. Samsung offers 3 color space settings, but only 1 can be calibrated so you have to pick whether you want HD "perfect" or "SD" perfect... though the default Samsung color space is pretty accurate for HD, it's off a fair bit in Magenta (displaced towards Red). So there are displays that dance around getting everything right, but they just aren't quite there yet.

Of course there's always the $4000 Lumagen Radiance! LOL! If you want perfection... that's going to get everything as right as any display could ever get it. But $4000 for a video processor is a LOT of money - especially when you don't know what the future will bring re. new video standards and display technologies.

For the record, I don't believe the Samsung's let you choose the color decoder, only the colorspace.
I'll drop Joe a note to confirm, but he has been so vocal in recent years about the discrepency in various devices on this issue, I'd be surprised that both decoding and color space are not included in his projectors' feature to force SMPTE C/HD/EBU at will in the picture setup menu.

Is the flip-side of that DVD(the SD program) intended for non-HD consumers like myself, or is that digital SD intended to be read only by HD-DVD players?
The flip side is the SD version of 'DVE.' The same version it's always been. DVD is "digital" SD and always has been. If it could only be read by HD DVD players, it would have to be the HD DVD format, and not SD. It could not be called a "combo" disc in that case.

Best regards and beautiful pictures,
G. Alan Brown, President
CinemaQuest, Inc.

"Advancing the art and science of electronic imaging"

This is what the setup should look like for all displays:

http://www.cine4home.de/Specials/Planar8150/PD8150-8.JPG
(from Planar PD8150 projector)

Nice!:D

I keep forgetting to point out that there is, with some combinations of components, a way to completely bypass the whole color decoder/color matrix issue.

If the source component has an RGB mode and you are using HDMI for the connection, the whole issue of needing a matix to encode and decode YCbCr is gone. The player will send RGB and the display will display RGB. Every display will accept RGB over HDMI.

Now... there is a potential for error in the display still... some displays will convert the incoming RGB to YCbCr and back to RGB.... putting the encode/decode sequence back into play. Not a problem if there are no errors in the encode/decode, but if there are errors, you're not going to bypass them by sending RGB. You have to be pretty "deep" into video displays and have instrumentation and a signal generator to figure out if the display is behaving itself or not. Even then, you may not be able to figure it out without help from the manufacturer.

Most people know there are 2 RGB modes... 0-255 is what is used in the PC world... this means there are 256 possible values for red, 256 for green and 256 for blue. This mode may be referred to as RGB-PC or RGB 0-255 or "Low" or "Dark" or some other unhelpful word. Consumer video in RGB mode uses digital values from 16-235. This is referred to with different names in different products... RGB-v, RGB-Video, RGB 16-235, Normal, Bright, or some other unhelpful word. Either mode is fine as long as the display and source are using the same type of RGB. If there's a mismatch, it will be most noticeable in the black level being very high (no black at all, just dark gray) or you could have crushed shadow detail..

Now... there is a potential for error in the display still... some displays will convert the incoming RGB to YCbCr and back to RGB.... putting the encode/decode sequence back into play. Not a problem if there are no errors in the encode/decode, but if there are errors, you're not going to bypass them by sending RGB. You have to be pretty "deep" into video displays and have instrumentation and a signal generator to figure out if the display is behaving itself or not. Even then, you may not be able to figure it out without help from the manufacturer.

One would hope that these errors were small :) since the software developers know which matrix it was encoded with - whichever one they chose - thus the decoding matrix should be known. Of course, there could still be "rounding" errors and bugs in the functions used.

I re-did a quick test last night with my Oppo sending 1080p over HDMI to the Samsung LCD. Using GetGray for the 75 color window patterns, there was little difference in color brightness measures between any of the Oppo's modes: RGB, YCbCr 4:4:4 and YCbCr 4:2:2. I think the biggest difference was blue measured -4.6% in RGB mode and -6.0% in the YCbCr modes - both compared to the HD709 standard. The rest were all within +/- 1%.

It seems to me in an ideal world, each display would have the ability to switch color spaces and do the correct decoding based on data supplied to it from the incoming data stream. Each dvd/blu-ray/digital tv show could be required to contain this information and supply it to the next device in the chain as part of a standard handshake. Then all one would have to worry about is whether or not the display is doing things accurately. There would be no need for the end user to have to worry about which color space or color decoder matrix to choose.

cheers,


--tom