or Connect
AVS › AVS Forum › Display Devices › Display Calibration › Guide: GREYSCALE CALIBRATION FOR DUMMIES
New Posts  All Forums:Forum Nav:

Guide: GREYSCALE CALIBRATION FOR DUMMIES - Page 2

post #31 of 258
Wow Kal, that's a fantastic guide you've written up! I've only just now skimmed through it, I want to take a much more thorough gander this weekend. I did catch a few minor, somewhat obscure technical quibbles in my speed-read, I'll go through the whole thing though and see if I catch any more and let you know. Really really cool.
post #32 of 258
Thread Starter 
Thanks guys! Any quibbles or issues any of you can bring up are more than welcome!

I don't consider myself a colour calibration expert by any stretch of the imagination: Just an enthusiast who found himself frustrated when reading other guides as I had to constantly search to figure out what certain terms meant... so I figured I'd throw something together that covered it all from start to finish.

Kal
post #33 of 258
Thanks, kal, great job.
post #34 of 258
Thank you Kal for validating the ideal gamma target is indeed 2.20 and not 2.5.
post #35 of 258
Is using the SD AVIA disk going to REALLY skew my resuts for grayscale? Or will it just bit a "little" off?

It's the only calibration DVD I currently own

Also, when I try to calibrate my PC can I use CalMan test pattern generator as an accurate way to calibrate the grayscale for that input?

I am receiving my i1Display LT today, so I haven't tried anything yet, but I'm VERY anxious to get into this!

On a side note: Like a lot of people on here, your guide has given me the confidence to purchase and try out the i1Display colorimeter. Thank You!

-Brian
post #36 of 258
Quote:
Originally Posted by tbrunet View Post

Thank you Kal for validating the ideal gamma target is indeed 2.20 and not 2.5.

The second bold area from http://www.avsforum.com/avs-vb/showt...0#post13486170 would seem to indicate that at 2.2 gamma is not necessarily ideal, depending on such things as room lighting.
post #37 of 258
Quote:
Originally Posted by kamui View Post

Is using the SD AVIA disk going to REALLY skew my resuts for grayscale? Or will it just bit a "little" off?

The Avia IRE test patterns are contminated by chroma, but you can make them usable by either turning your color control all the way down or by using component and only connecting the Y cable. Only connecting the Y (green) cable will cause your TV to only display the luma information and not the chroma information. I've sucessfully used Avia to calibrate my grayscale in the past using the component method. The good part about the component method is that it negates any problems if you have a TV that for some reason the color controls interfere with the grayscale.
post #38 of 258
That is wonderful news, thank you so much.

I didn't know exactly to look to see what the problem with AVIA was, I just remember hearing a lot that it's grayscale was off.

I can't wait to try this out. Hopefully the difference will be so great that I can calibrate it when the girlfriend isn't around and then when we watch a movie together, she will comment on her own that for some odd reason she thinks the TV looks amazing all of a sudden

That would be a true test for me :-p

-Brian
post #39 of 258
Thread Starter 
Quote:
Originally Posted by alluringreality View Post

The second bold area from http://www.avsforum.com/avs-vb/showt...0#post13486170 would seem to indicate that at 2.2 gamma is not necessarily ideal, depending on such things as room lighting.

That's true. I've actually flip-flopped a bit on this. I've updated the guide to mention that the target should be somewhere between 2.2 to 2.4 depdning on room lighting. Ie: A pitch black room with zero reflections may be fine at 2.4, a more typical room may want to target 2.2. Everyone's welcome to try both and see what they like.

I'm about in the middle (2.27 average) with closer to 2.2 at the lower (0-40 IRE) range which is perfect as it brings out shadow details nicely



The average is 2.27 (cyan line).
The solid yellow line is my "after" greyscale calibration & gamma boost reading.
The dotted yellow line is the "before" greyscale calibration and without a gamma boost reading.

Not all displays require a gamma boost but pretty much CRT displays (like mine) do. I use a Crescendo Systems RTC2200.

Kal
post #40 of 258
Quote:
Originally Posted by alluringreality View Post

The second bold area from http://www.avsforum.com/avs-vb/showt...0#post13486170 would seem to indicate that at 2.2 gamma is not necessarily ideal, depending on such things as room lighting.

Yes by definition gamma is subjective and will correlate with a given viewing ambient, but the ideal target is indeed i.e. to match a properly calibrated (telecine) mastering reference monitor (Sony or Ikegami). These monitors are designed for a ruler flat 2.20 transfer characteristic response.
post #41 of 258
Thread Starter 
Quote:
Originally Posted by kamui View Post

Is using the SD AVIA disk going to REALLY skew my resuts for grayscale? Or will it just bit a "little" off?

It's the only calibration DVD I currently own

DVE:HD basics is only $17 or you can also go for AVS HD 709 which is free.

As long as you follow the instuctions above to correct Avia, it'll work too.

Kal
post #42 of 258
Thread Starter 
Quote:
Originally Posted by tbrunet View Post

Yes by definition gamma is subjective and will correlate with a given viewing ambient, but the ideal target is indeed i.e. to match a properly calibrated (telecine) mastering reference monitor (Sony or Ikegami). These monitors are designed for a ruler flat 2.20 transfer characteristic response.

Good enough for me. I'll add that point if you don't mind!

Kal
post #43 of 258
Here is a link and reference for the information

Quote:
The gamma for the Sony CRT agrees perfectly with the 2.20 standard value. (CRT monitors from Ikegami, another major brand of professional studio monitors, also have a gamma of 2.20 according to their Director of Engineering.)

http://www.extremetech.com/article2/...1736944,00.asp
Dr. Raymond Soneira, President DisplayMate Technologies Corp

Dr. Raymond Soneira is President of DisplayMate Technologies Corporation of Amherst, New Hampshire, which produces video calibration, evaluation, and diagnostic products for consumers, technicians, and manufacturers. See www.displaymate.com. He is a research scientist with a career that spans physics, computer science, and television system design. Dr. Soneira obtained his Ph.D. in Physics from Princeton University, spent 5 years as a Long-Term Member of the world famous Institute for Advanced Study in Princeton, another 5 years as a Principal Investigator in the Computer Systems Research Laboratory at AT&T Bell Laboratories, and has also designed, tested, and installed color television broadcast equipment for the CBS Television Network Engineering and Development Department. He has authored over 35 research articles in scientific journals in physics and computer science, including Scientific American. If you have any comments or questions about the article, you can contact him at dtso@displaymate.com.

http://www.extremetech.com/article2/...1736943,00.asp
post #44 of 258
Thread Starter 
Thanks! Interesting reading.

I've flopped back the other way a bit in my guide and am now suggesting again that 2.2 be the gamma target that everyone should be trying to achieve, with up to 2.4 possibly being adequate in some completely light controlled rooms where light reflections are minimized. 2.2 is still the target that everyone should be trying to achieve however.

Thanks for the extra info guys!

Kal
post #45 of 258
Uh oh. Here we go again. I have read and read on this and everytime I think I understand the whole gamma thing something else pops up like this.
post #46 of 258
There is no need to make this excessively complicated.

Shoot for 2.2 gamma measured by the standard power function. If you have a display with a very high native contrast ratio that you view in a completely dark environment, you can experiment with higher values up to 2.5. But don't go below 2.2 or above 2.5 in any case.
post #47 of 258
This is good and reasonable advice. The end result of how the system performs with actual program material is what matters. 2.2 works pretty well in many systems, and is a reasonable starting point.
post #48 of 258
Thread Starter 
Thanks Tom. That makes a lot of sense. Explains why a lot of CRT projector owners don't mind higher gamma values (2.4/2.5).

A question for all of you: Now that I've added colour decoder and primary/secondary information, should I rename this whole thing to "COLOUR calibration for dummies" instead of "GREYSCALE calibration for dummies"? Or is that going to confuse people?

Kal
post #49 of 258
Quote:
Originally Posted by kal View Post

calibration for dummies" instead of "GREYSCALE calibration for dummies"? Or is that going to confuse people?
Kal

Kal,

how about just Display Calibration for Dummies?

cheers,


--tom
post #50 of 258
Quote:
Originally Posted by thomasl View Post

Kal,

how about just Display Calibration for Dummies?

cheers,


--tom

Not that it really matters, but name used (though incomplete Guide Removed until completion).

-Shawn
post #51 of 258
Thread Starter 
I don't think "Display Calibration for Dummies" would work as that goes beyond just colours and greyscale. That would be everything else about the display too.

Kal
post #52 of 258
Quote:
Originally Posted by kamui View Post

Is using the SD AVIA disk going to REALLY skew my resuts for grayscale? Or will it just bit a "little" off?

Also, I thought the labeling for gray windows in Avia correspond only to correct output IRE levels if you use 7.5 IRE black setup. If setup is 0, then the labeling doesn't reflect output IRE.

I don't think this matters in trying to achieve a flat-tracking grayscale. But it might matter when trying to measure gamma?
post #53 of 258
Did a quick read through, a couple things I would change:

Quote:


The target value for average gamma we are trying to achieve for our display is 2.2. The ColorHCFR software tries to target 2.2. The ideal target is to match a properly calibrated (telecine) mastering reference monitor used by the movie studios. These monitors are designed for a ruler flat 2.20 response.

Target gamma is actually inherently assumed 2.5, which is inherent to a CRT and thus the reference. The encode gamma from NTSC days is 1/2.2, which is where the 2.2 comes from(which is an incorrect interpretation since that defines the encode, NOT the assumed de-gamma), which is a very commonly assumed target which is not actually correct. The flow is actually not linear, and ends up about 1.25 for better subjective viewing results. It is unfortunate that this software and others uses 2.2 as the target. Even accupel does the same, despite the fact that in their white papers they are quite clear that 2.5 is actually the desired display standard gamma. Lower gammas can be preferred for subjective reasons, but are not standard.

Quote:


You'll need a test disc with 0 to 100 IRE window patterns. We recommend the Digital Video Essentials: HD Basics disc (in either Blu-ray or HD DVD) as it's the most up to date and the least expensive. We do not recommend using the Avia DVD as the greyscale patterns are incorrect! (The authors have admitted to this).

I would though, include Avia II which is correct, and is a handy disc to use. I recommend it. Additionally, I strongly dislike using the IRE units at all, and would prefer % be used instead as IRE units are very confusing, and not really correct to use anyway. Although, Avia II still uses the asinine IRE labels and is based around 7.5IRE setup, which makes it very confusing. So actually, maybe I take that back, using DVE which is labeled in % is more clear for greyscale. I prefer Avia II for many other settings, and usability though.

Quote:


Note: The 4% below black (blacker-than-black) bar may not be visible on your setup. That's fine and normal. Don't worry about it. Not all setups will display blacker-than-black (or whiter-than-white) content. You aren't losing out on anything if you cannot see this content.

That's not normal, nor "fine." It may not be a big deal, but such a system is clipping the video range. It deviates from reference performance and is not preferred, and can have impacts on the image. That may not be significant for some viewers, but that doesn't make it "fine" or "normal."

Quote:


If the blue and green primaries on your digital projector are too far out causing the colours to be too vibrant, you can possibly tame them a bit by adding a neutral density (ND) filter in front of your lens. The CC20R, CC30R, and CC40R filters (for example) are popular choices. These are light filters that can be added into the light path in front of your digital projector. It limits the light output which can help with the black level and also has the effect of pulling in the blue and green primary points (red remains unaffected). The number represents the percentage of attenuation that is performed, so for example, a CC20R filter attenuates light output by 20%.

Slight confusion, those are not neutral density filters. A neutral density filter is by definition neutral, and will only reduce the light output. It will have no effect on color, or any other projector settings whatsoever. All it will do will dim the light output. The filters listed are not neutral, thus they will affect greyscale balance across the entire spectrum evenly. Obviously you'll want to use the right filter for the right purpose.

Quote:


Since the filter reduces the bulb light output at low light levels, the color correcting filter gives you the added bonus of deeper blacks and improved contrast ratio.

The filter itself is passive, and will hace no inherent impact on contrast ratios, regardless of whether it's a colored filter or a neutral density filter(and the ND filter will not have any at all). The reason to use a color filter on a digital display is that often a digital display will have too much blue or green, but using the gains in the digital projector to pull those in can have the effect of reducing the contrast ratio since you have a limited on/off CR range for each primary, essentially. As you essentially lower the gain of each primary control, you also will have to lower the global white level setting to stay below the display's white level clip point as the primary gains are lowered. This sacrifices contrast ratio. Using a colored filter achieves this without needing to sacrifice CR at all, or as much since you may not need to adjust down the primary gains at all or as much in order to achieve good greyscale. In this manner, a colored filter can have the end effect of improving your calibrated contrast ratio, but it in and of itself will not affect CR at all, except insofar as it helps during the calibration process.

Additionally, it may be worth including a note as to the distinction between D65 and 6500K. They are not really the same thing. You describe 6500K as a neutral color early on in the guide, then switch to D65 terminology later on. D65 is the standard, though 6500K is very close to it. Not that big a deal for the novice, but that should be clearer.

Hope that helps tweak the guide a bit, and thank you so much for writing up such a useful guide Kal!
post #54 of 258
Thread Starter 
Quote:
Originally Posted by ChrisWiggles View Post

Did a quick read through, a couple things I would change:

Thanks for taking the time Chris. It's really appreciated!

Quote:


Target gamma is actually inherently assumed 2.5, which is inherent to a CRT and thus the reference. The encode gamma from NTSC days is 1/2.2, which is where the 2.2 comes from(which is an incorrect interpretation since that defines the encode, NOT the assumed de-gamma), which is a very commonly assumed target which is not actually correct. The flow is actually not linear, and ends up about 1.25 for better subjective viewing results. It is unfortunate that this software and others uses 2.2 as the target. Even accupel does the same, despite the fact that in their white papers they are quite clear that 2.5 is actually the desired display standard gamma. Lower gammas can be preferred for subjective reasons, but are not standard.

Interesting. So here's another arguement for 2.5 instead of 2.2. I've gotten enough arguments for from both sides and frankly they both make sense. I see why this is such as confusing topic.

Quote:


I would though, include Avia II which is correct, and is a handy disc to use. I recommend it.

I've added a note about Avia II being correct as well now. Thanks!

Quote:


Additionally, I strongly dislike using the IRE units at all, and would prefer % be used instead as IRE units are very confusing, and not really correct to use anyway. Although, Avia II still uses the asinine IRE labels and is based around 7.5IRE setup, which makes it very confusing. So actually, maybe I take that back, using DVE which is labeled in % is more clear for greyscale. I prefer Avia II for many other settings, and usability though.

Agreed. I don't have Avia II, but do have DVE, but still prefer to use AVSHD709 due to the simplicity. Unfortunately not everyone has HD yet or the means to burn their own DVD or Blu-ray discs (nor is AVSHD709 even done yet). So for this guide I wanted to pick an inexpensive easy disc that everyone had access to.

Quote:


That's not normal, nor "fine." It may not be a big deal, but such a system is clipping the video range. It deviates from reference performance and is not preferred, and can have impacts on the image. That may not be significant for some viewers, but that doesn't make it "fine" or "normal."

Thanks - I've changed the text. I purposely didn't want to confuse the issue too much by going into the differences between 0-255 and 16-235.

Quote:


Slight confusion, those are not neutral density filters. A neutral density filter is by definition neutral, and will only reduce the light output. It will have no effect on color, or any other projector settings whatsoever. All it will do will dim the light output. The filters listed are not neutral, thus they will affect greyscale balance across the entire spectrum evenly. Obviously you'll want to use the right filter for the right purpose.

The filter itself is passive, and will hace no inherent impact on contrast ratios, regardless of whether it's a colored filter or a neutral density filter(and the ND filter will not have any at all). The reason to use a color filter on a digital display is that often a digital display will have too much blue or green, but using the gains in the digital projector to pull those in can have the effect of reducing the contrast ratio since you have a limited on/off CR range for each primary, essentially. As you essentially lower the gain of each primary control, you also will have to lower the global white level setting to stay below the display's white level clip point as the primary gains are lowered. This sacrifices contrast ratio. Using a colored filter achieves this without needing to sacrifice CR at all, or as much since you may not need to adjust down the primary gains at all or as much in order to achieve good greyscale. In this manner, a colored filter can have the end effect of improving your calibrated contrast ratio, but it in and of itself will not affect CR at all, except insofar as it helps during the calibration process.

This is the stuff that frankly I know very little about as I don't own a digital projector - so thank you! I've added some extra information and quoted you on the second part (hope you don't mind).

Quote:


Additionally, it may be worth including a note as to the distinction between D65 and 6500K. They are not really the same thing. You describe 6500K as a neutral color early on in the guide, then switch to D65 terminology later on. D65 is the standard, though 6500K is very close to it. Not that big a deal for the novice, but that should be clearer.

Yes, I really pulled my hair on this one: How much detail do I go into? In the end I took the approach to mention 6500K up front as meny people are used to the idea and see it in their display menus, but then when it comes to the actual calibration I mention calibrating greyscale to D65 (not 6500K) to make it (hopefully) clearer that D65 is a point. I kind of bridge the gap when I talk about the Colour Temperature graph after I've already explained the concept of D65 and the CIE diagram. I wrote:

Quote:


Some magazines love to show this graph as it's easy for people to follow. They tell us that if the graph is perfectly flat with values of 6500K from black to white that the greyscale is therefore perfect. Not true. What you need to understand is that many points on the CIE diagram actually add up to a colour temperature of 6500K, but only one is correct: The D65 point where x=0.313 and y=0.329. You can have a display that measures 6500K across the board but still look incorrect. The previous RGB Levels graph and the upcoming CIE graph give us a far more accurate picture of what's actually going on. The less we say about the colour temperature graph the better. Just ignore it, it's truly useless. Don't make the beginner mistake of only posting your colour temperature graph in online forums. Without the other graphs or data for context, it's useless.

Maybe not perfect but I found too often I was sticking too much detail up front to make everying 100% "correct" but then when I re-read it, I found I was diving into the gory technical details too fast for the average person to understand. For that reason I didn't want to talk about D65 at all in the intro as I'd have to talk about CIE first, and that's not really an intro topic. So I talk about 6500K (I don't even mention it's 6504K actually) as many people have seen or heard of 6500K.

Quote:


Hope that helps tweak the guide a bit, and thank you so much for writing up such a useful guide Kal!

Thank you for commenting on it!

Kal
post #55 of 258
Quote:


Interesting. So here's another arguement for 2.5 instead of 2.2. I've gotten enough arguments for from both sides and frankly they both make sense. I see why this is such as confusing topic.

It is very confused, sadly. Personally I very much side with Poynton.

I made a thread here, where I argue that point. Also included is a very old thread by Guy Kuo making the same point:

http://www.avsforum.com/avs-vb/showthread.php?t=1008297

The problem though, is that if you dig through some of the Insider's threads, it seems some mastering houses are also wrongly assuming that 2.2 is the standard (when it isn't, nor has it ever been) for viewing/playback, and aligning the mastering monitors to that. This is unfortunate, and yet another place where confusion reigns, and sadly appears to have the upper hand at this point.

Though, as a CRTer myself, I run a little bit lower than 2.5, about 2.3, for subjective preference reasons because of the low ANSI-CR of CRTs.. So while I think that 2.5 should be forwarded as the reference, it is perfectly fine to deviate from that to yield an improved image so long as you know what you're doing and understand the tradeoffs. But I do not think that 2.2 should be given as the target or reference gamma, because it is not. Calibrating to 2.2 for various reasons anyway is fine, but it is unfortunate that people wrongly assume that it is the reference.
post #56 of 258
I have a question that I'm hoping I can convey without too much confusion:

If all displays and monitors are supposed to be scientifically calibrated to standards then why is there so much allowance for deviation in luminance?

For instance, say I had two completely identical displays that I calibrated. On one display I calibrated it to acheive 30 fL and on the other I calibrated it to acheive 50 fL, both of which could be considered correct.

How in the world can the two images be considered the same?

If I displayed a bright white on the one monitor, it would appear brighter than the other, so how are they the same? (substitue white for red, blue, etc.)

If for instance I had a picture of a bright red apple on the one monitor, on the other wouldn't the same picture of an apple appear as a darker red? How is that scientific?

-------

I hope it's not something simple I'm overlooking and I hesitated to ask, as I didn't want to seem like an ignorant fool. But I've been unable to determine the answer by myself and would appreciate any insight.

I ask because I'm trying to calibrate my LCD, and I've seen suggested ranges from about 30 fL to about 60 fL. I've tried both and liked both, but I ended up choosing based on opinion, not science.

Thank you in advance for any input,

-Brian
post #57 of 258
Quote:
Originally Posted by kamui View Post

I have a question that I'm hoping I can convey without too much confusion:

If all displays and monitors are supposed to be scientifically calibrated to standards then why is there so much allowance for deviation in luminance?

For instance, say I had two completely identical displays that I calibrated. On one display I calibrated it to acheive 30 fL and on the other I calibrated it to acheive 50 fL, both of which could be considered correct.

How in the world can the two images be considered the same?

If I displayed a bright white on the one monitor, it would appear brighter than the other, so how are they the same? (substitue white for red, blue, etc.)

If for instance I had a picture of a bright red apple on the one monitor, on the other wouldn't the same picture of an apple appear as a darker red? How is that scientific?

-------

I hope it's not something simple I'm overlooking and I hesitated to ask, as I didn't want to seem like an ignorant fool. But I've been unable to determine the answer by myself and would appreciate any insight.

I ask because I'm trying to calibrate my LCD, and I've seen suggested ranges from about 30 fL to about 60 fL. I've tried both and liked both, but I ended up choosing based on opinion, not science.

Thank you in advance for any input,

-Brian


Allow me to take an initial stab at this...as it relates to luminance the issue is NOT whether the display should be 12ftL, 30 ftL or any other amount...the issue is the relative luminance of the colors; that is R, G and B should be approximately 21%, 72% and 7% of white's luminance respectively.

The issue that you are referring to is brightness and in this regard (i) there is no general consensus but rather guidelines and the different brightness levels that you read about are for different conditions; foe example, in a pitch black room something in the order of 12 to 20ftL is the norm as 60ftL would be like a flashlight whereas in a bright room with lots of sun 12ftL would not produce a watchable image whereas 40ftL would...

HTH
post #58 of 258
"there is no general consensus but rather guidelines"

I think that's what I needed to hear.

I'm mostly curious as to why there is so MUCH variance between levels. I understand that room lighting and enviroments dictate this, but how do I know what to shoot for in a given enviroment?

In my case I have a completely light controlled room with an Idealume bias light and since I'm using an LCD panel, should I just aim for the lowest possible foot lambert measurment suggested?

In other words, is there a mathmatical ratio I could assume for calculating the difference between viewing enviroment and brightness of the display?

Thanks for the reply,

-Brian
post #59 of 258
Thread Starter 
The smaller the set the brighter people tend to run them. Hence the reason my guide mentions 12-16 ftL for projector setups (larger screens) and 30-40 ftL for smaller screens.

It's all about eye strain. You don't want a 10 foot screen in a light controlled room to be showing 40 ftL. That would be unbearable.

Bulbs and tubes get dimmer over time, so there's no way for manufacturers to calibrate them ahead of time even if they wanted to. They also don't know what sort of environment you're going to be watching in (dark or daylight). A TV that's set lower in terms of ftL output will not sell. They're all set to "torch" mode by default to appear brighter than the competition on the showroom floor.

Kal
post #60 of 258
Quote:


The smaller the set the brighter people tend to run them. Hence the reason my guide mentions 12-16 ftL for projector setups (larger screens) and 30-40 ftL for smaller screens.

It's all about eye strain. You don't want a 10 foot screen in a light controlled room to be showing 40 ftL. That would be unbearable.

Not sure I agree with this? It is about field of view. The smaller the FOV the brighter the image needs to be.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Display Calibration
AVS › AVS Forum › Display Devices › Display Calibration › Guide: GREYSCALE CALIBRATION FOR DUMMIES