Luminous Landscape Forum
Raw & Post Processing, Printing => Colour Management => Topic started by: colrook on February 10, 2012, 05:54:13 pm
-
Hey there,
just playing arround with i1Profiler, but there seems to be something strange, would be very nice if someone could confirm this behaviour or bug, to check if there is a problem on my end..
I try to calibrate with standard settings (D65, flare off, ambient light off, ADC off, Bradford, ICC v2, Matrix based ) but with "sRGB" tonal response curve instead of gamma 2.2!
With gamma 2.2 everything works fine, but as soon as I use sRGB at the end of the measurement the screen keeps white and nothing happens anymore :'(
Tried with i1Profiler 1.2.0 on a Win 7 x64 machine with i1Display pro.
Do you experience the same problem?
Thanks in advance!
-
I have no idea why they have an sRGB setting (or why you’d use it) when 2.2 is available. Just use 2.2.
-
I have no idea why they have an sRGB setting (or why you’d use it) when 2.2 is available.
According to Wikipedia
sRGB also defines a nonlinear transformation between the intensity of these primaries and the actual number stored. The curve is similar to the gamma response of a CRT display. It is more important to replicate this curve than the primaries to get correct display of an sRGB image. This nonlinear conversion means that sRGB is a reasonably efficient use of the values in an integer-based image file to display human-discernible light levels.
and
It is often casually stated that the decoding gamma for sRGB data is 2.2, yet the above transform shows an exponent of 2.4. This is because the net effect of the piecewise decomposition is necessarily a changing instantaneous gamma at each point in the range: it goes from gamma=1 at zero to a gamma of 2.4 at maximum intensity with a median value being close to 2.2. The transformation was designed to approximate a gamma of about 2.2, but with a linear portion near zero to avoid having an infinite slope at K = 0, which can cause numerical problems. (This is surrounded by lots of complicated formulas.)
http://en.wikipedia.org/wiki/SRGB (http://en.wikipedia.org/wiki/SRGB)
Doesn't this mean that gamma 2.2 is wrong for sRGB?
When I dumped an sRGB profile with an sRGB curve that was generated by SpectraView II with Argyll iccdump, the TRC values are lookup tables and not gamma curve specifications. A sRGB profile with gamma 2.2 had "Curve is gamma of 2.199219" for the TRCs.
And FWIW, I have the same problem with iProfiler: when I select the sRGB curve, it hangs at the final white screen. Also using i1 Display Pro on 64 bit Win 7. Gamma 2.2 works. I've been shrugging my shoulders at this, because my main monitor is a PA241W. I only use iProfiler on my secondary monitors so they sort of match the PA monitor when I slide Explorer windows around.
Wayne
-
Thanks for your fast replies guys :D
I thought sRGB would be a tiny little bit more exact for viewing sRGB content (I do mainly 3D rendering) but I maybe wrong here?!
Also thanks to Wayne for confirming that this is a plain software bug. Hope x-rite fixes this with next release.
-
Tried with i1Profiler 1.2.0 on a Win 7 x64 machine with i1Display pro.
Do you experience the same problem?
Yes, I've also had this. I prefer to use Basiccolor and ArgyleCMS (but stick with 2.2 gamma anyway).
-
According to Wikipedia
This does not change my two questions.
There is no reason to select sRGB, especially if you hit what is a bug in the software.
The sRGB color space is theoretical, based on a CRT with specific phosphors circa 1993. I kind of doubt you are either using such a display and question why anyone would need to mimic that. In an ICC aware application, the TRC gamma isn’t at all critical anyway.
-
The argument for calibrating your display gamma to be the same as your working space is that it makes for one less time when output levels need to be modified and collapsed. As Andrew pointed out, sRGB emulates 20 year old CRT display technology. Any reduction in level compression from matching calibration and working space gamma to sRGB will be overwhelmed by the difference between the native gamma of your LCD and sRGB.
In the real world, however, you will likely never see any visible difference in gradient smoothness or output banding between 2.2 gamma and sRGB. i1Profiler's problems with sRGB calibration are another matter. That's a bug. Tried it on a screen here and see the same problem.
-
There is no reason to select sRGB, especially if you hit what is a bug in the software.
As I said in my other post, I now use a NEC PA241W, calibrated with SpectraView II and an i1 Display Pro. The PA supposedly can emulate sRGB exactly. I thought that it would be better to get the TRC curve to match sRGB.
I read somewhere that the difference between the sRGB curve and gamma 2.2 shows up in shadow tones. I noted a big discrepancy in shadow tones with my previous monitor (HP LP2065 calibrated with MonacoOPTIX/DTP-94) and prints. Which are usually from a mini-lab printer. (Yes, I know that there is no such thing as an sRGB printer--I looked at 3D plots of a lot of printers on Dry Creek Photo's site. Most mini-lab printers do fit in sRGB.) I was sort of hoping that the PA monitor's shadow tones would better match prints. (I only got the PA a few weeks ago and haven't edited any images that have been printed yet.)
I will be the first to admit that I only have a fuzzy understanding of color theory, even after reading your book. And reading all the tutorials on Dry Creek Photo.
Wayne
-
The PA supposedly can emulate sRGB exactly.
Exactly? Probably not although if so, I’d sure like to see proof. Note, calibrating to even a luminance that doesn’t fall within the sRGB spec would invalidate exactly.
And if you have a PA, use SpectraView or MultiProfiler and forget i1P.
-
Exactly? Probably not although if so, I’d sure like to see proof. Note, calibrating to even a luminance that doesn’t fall within the sRGB spec would invalidate exactly.
Well, the resulting sRGB emulation monitor profile matches up visually with sRGB when I compare them at http://www.iccview.de/content/view/3/7/lang,en/ (or with the Argyll 3D profile VRML generating programs that do much the same thing that the iccview site does.) A lot closer than how other monitor profiles match sRGB.
But no, it isn't an exact numerical match. I think that would be expecting too much from a $250 i1DP.
And if you have a PA, use SpectraView or MultiProfiler and forget i1P.
I do use SpectraView on my PA241W. SV is the only way to make the monitor be (almost) sRGB, at the monitor level. SV makes the monitor match the target and then writes a linear 'vcgt' table in the profile. I only use iProfilier to calibrate my non-NEC secondary monitors. I did use SV to profile my non-NEC monitors and the profiles were close to the profiles that iProfilier made (when examined as 3D gamut plots.) Which is a good cross check, isn't it?
Wayne
-
I have no idea why they have an sRGB setting (or why you’d use it) when 2.2 is available. Just use 2.2.
Why do you say that? If you calibrate it to sRGB TRC then programs such as IE that don't take into account anything will show images with the proper tone curve. If you calibrate to gamma 2.2 then you will see the shadows too dark and the contrast a little bit different than it should be in programs such as IE.
OTOH for viewing TV and movies and sometimes games, gamma 2.2 often looks a bit nicer.
-
Exactly? Probably not although if so, I’d sure like to see proof. Note, calibrating to even a luminance that doesn’t fall within the sRGB spec would invalidate exactly.
And if you have a PA, use SpectraView or MultiProfiler and forget i1P.
It sure gets closer than almost any standard gamut monitor, most of those fall short in one way or another or don't track saturation curves all that well, etc.
NEC PA241W calibrated sRGB mode:
(http://www.smugmug.com/photos/935962421_Zq8o7-X2.jpg)
(http://www.smugmug.com/photos/935962713_ErzHd-X2.jpg)
(http://www.smugmug.com/photos/935962459_T38sW-X2.jpg)
(http://www.smugmug.com/photos/935962479_Ru5zc-X2.jpg)
(http://www.smugmug.com/photos/935962538_kAxtT-X2.jpg)
(http://www.smugmug.com/photos/899780420_H3qri-X2.jpg)
how many regular old monitors get everything matching so well? Most will have at least one primary fall short, most have the saturation tracking for various colors going all up and down along the way, some have the primary luminance tracking off, etc.
here are measurements from a an sRGB monitor with unusually numerous controls (many have the luminance and saturation color curves a lot farther off):
(http://www.smugmug.com/photos/891293788_ynqiZ-X2.jpg)
(http://www.smugmug.com/photos/889793209_AsCbd-X2.jpg)
(http://www.smugmug.com/photos/889793200_V9Z5E-X2.jpg)
and you might get luminance and saturation primary curves like say:
(http://www.smugmug.com/photos/875789294_dkzBa-X2.jpg)
(http://www.smugmug.com/photos/875789284_qGWbE-X2.jpg)
-
Why do you say that? If you calibrate it to sRGB TRC then programs such as IE that don't take into account anything will show images with the proper tone curve. If you calibrate to gamma 2.2 then you will see the shadows too dark and the contrast a little bit different than it should be in programs such as IE.
The differences in the TRC of sRGB and a true 2.2 gamma curve is insignificant. And in one case, you catch a bug in the process of setting that option. Plus any non ICC aware application is, well not color managed. You really think that your LCD can behave as a true (as exactly defined) sRGB device and even so, that would have any meaningful role in IE or other non ICC aware app’s? Your LCD is using P22 phosphors I suppose? Talk about debating how many ICC profiles can dance on the head of a pin.
-
The differences in the TRC of sRGB and a true 2.2 gamma curve is insignificant. And in one case, you catch a bug in the process of setting that option. Plus any non ICC aware application is, well not color managed. You really think that your LCD can behave as a true (as exactly defined) sRGB device and even so, that would have any meaningful role in IE or other non ICC aware app’s? Your LCD is using P22 phosphors I suppose? Talk about debating how many ICC profiles can dance on the head of a pin.
It's not totally insignificant. With plenty of photos it's readily noticeable (between using gamma 2.2 and sRGB TRC).
And OK sure you get a bug that crashes the calinbration program so obviously you simply can't set sRGB TRC with that program for now but that just means you need to send in a bug report not decide to never use sRGB TRC, come on.
Yes, the better wide gamut monitors in their sRGB mode very, very much can behave as sRGB devices when used with non-display managed programs such as IE. Did you not see my charts above? The first set? Taken with samples that had no ICC color management applied to them whatsoever? Those sRGB mode non-ICC managed measurements on my PA241W are much CLOSER to the exact spec for sRGB than any I have ever taken on any of my regular old non-pro level sRGB monitors ever produced even using ICC management on the sample patches using when using profiles made from well regarded programs such as CEDP.
You are wayyy off base saying that is dancing on the head of a pin. Utterly wrong. You seem to forget that the fancy wide gamut monitors with high bit 3D internal LUTs are fully capable of shifting the primary locations around and you can get them right on and you can also see that these sets are so linear that they track the saturation curves and primary luminance curves essentially perfectly as you move their locations around (many less fancy monitors have somewhat all over the place curves and even after fancy CEDP calibration they are still only partially corrected).
-
Yes, the better wide gamut monitors in their sRGB mode very, very much can behave as sRGB devices when used with non-display managed programs such as IE.
Again, the sRGB specifications are very well defined, down to the phosphors and ambient light conditions. Are you actually targeting the lumanice to 80 cd/m? P22 Phosphors? Because if you don’t, you are not producing sRGB as specified.
You might get “close” with some emulation but there is no reason to split hairs here especially when trying such a setting totally hoses the calibration anyway!
You are wayyy off base saying that is dancing on the head of a pin. Utterly wrong. You seem to forget that the fancy wide gamut monitors with high bit 3D internal LUTs are fully capable of shifting the primary locations around and you can get them right on and you can also see that these sets are so linear that they track the saturation curves and primary luminance curves essentially perfectly as you move their locations around (many less fancy monitors have somewhat all over the place curves and even after fancy CEDP calibration they are still only partially corrected).
Show me where you are getting the exact specifications for sRGB as defined by HP and MS (http://www.w3.org/Graphics/Color/sRGB)
-
Again, the sRGB specifications are very well defined, down to the phosphors and ambient light conditions. Are you actually targeting the lumanice to 80 cd/m? P22 Phosphors? Because if you don’t, you are not producing sRGB as specified.
You might get “close” with some emulation but there is no reason to split hairs here especially when trying such a setting totally hoses the calibration anyway!
Show me where you are getting the exact specifications for sRGB as defined by HP and MS (http://www.w3.org/Graphics/Color/sRGB)
You may as well say why bother for even gamma 2.2 just use whatever it does naturally, it's all splitting hairs if you don't have the exact viewing conditions in the spec, etc.
Anyway, at night time I actually do tend to use it around 80 cd/m^2 and somewhat close to spec ambient lighting, but when it is daytime I don't since that would be foolish and you need to adjust the luminance up, etc.
And as I said look at the gamut plot, the primaries and secondaries are all dead on and not just that, but as I also said, the way the saturation levels and lum track for them is actually better without ICC apps running than on my older, regular old sRGB monitors running ICC apps with a high quality monitor profile.
Sure there can be metameric differences between P22 phosphors and various LCD filters+backlights but it's not like using an ICC aware program does anything about that and it's also true that the metameric differences vary from person to person. Nothing less than spectral distribution color management would really work to fix that but such systems are used by regular programs and the sRGB ICC profiles are not defined that way, etc. so that is all beside the point. As measured by probe, you can dial them primaries in as perfectly as you can anything else with the monitor that he has.
Whether I use it during the day or night I can see a difference between gamma 2.2 and sRGB TRC when browsing images using IE. For some monitors that are quite well behaved, the biggest noticeable difference when using a monitor set to gamma 2.2 and a monitor profile and ICC aware apps can actually be the sRGB TRC vs gamma 2.2 difference and for a NEC PA monitor it is the only visible difference at all. If he want to calibrate one of his sRGB modes (the one for non TV/movie) stuff to TRC why should he not? I mean he should. Ideally. Obviously it goes without saying that if the program crashes during sRGB TRC he shouldn't use sRGB TRC with that program, I mean he can't. All he wanted to know is if anyone else had the same bug so he knows what to report to X-Rite in terms of getting it fixed or whether or not anyone found a way around the bug.
-
You may as well say why bother for even gamma 2.2 just use whatever it does naturally, it's all splitting hairs if you don't have the exact viewing conditions in the spec, etc.
Yes, I would recommend a native TRC gamma to many users.
Anyway, at night time I actually do tend to use it around 80 cd/m^2 and somewhat close to spec ambient lighting, but when it is daytime I don't since that would be foolish and you need to adjust the luminance up, etc.
So nighttime use is different from daytime use? That opens a whole can of worms that are far more critical than the tiny differences between a 2.2 gamma and an sRGB 2.2 TRC isn’t it?
Sure there can be metameric differences between P22 phosphors and various LCD filters+backlights but it's not like using an ICC aware program does anything about that and it's also true that the metameric differences vary from person to person.
Can be differences? I think it is a bit more clear cut than that. Your take is, you need to nail sRGB and to do so, it is critical to have a 2.2 TRC instead of a simple gamma curve. I don’t buy that. Now when asked if you can hit something far more critical, the sRGB spec for cd/m2 and chromaticity, that isn’t all that important to the debate.
The bottom line is, there is yet any justification for either the 2.2 sRGB TRC setting nor the bug it introduces in the software.
As measured by probe, you can dial them primaries in as perfectly as you can anything else with the monitor that he has.
So your take is, no matter what values I dial in, every instrument is going to provide that exact value? It might report what you asked for, I’d be hard pressed to believe you’ll always get it.
Whether I use it during the day or night I can see a difference between gamma 2.2 and sRGB TRC when browsing images using IE.
I have no reason to doubt that. What you haven’t proven is that one is closer to sRGB than the other or that being close to sRGB is even useful. There is a difference, great. There should be. There should be a difference calibrating to 80 cd/m2 (if you can really hit that) and 90cd/m2. I’d expect you would see a difference. Is that useful? Is that really producing sRGB (or is producing sRGB as defined even something to attempt?).
If he want to calibrate one of his sRGB modes (the one for non TV/movie) stuff to TRC why should he not?
Because it breaks the software for one. And it might be just a feel good option or something placed into the product solely for a marketing competitive matrix diagram.
-
Yes, I would recommend a native TRC gamma to many users.
Not sure what you are saying here.
I'm not sure many calibration/profiling programs even allow that do they?
Many monitors have TRCs going all over the place along the way are something way different than anything close to gamma 2.2.
And if it is some messy random TRC how will the profile even characterize it?
So nighttime use is different from daytime use? That opens a whole can of worms that are far more critical than the tiny differences between a 2.2 gamma and an sRGB 2.2 TRC isn’t it?
Whatever the differences you can see the difference between sRGB TRC and gamma 2.2 in either case so why avoid it?
Can be differences? I think it is a bit more clear cut than that. Your take is, you need to nail sRGB and to do so, it is critical to have a 2.2 TRC instead of a simple gamma curve. I don’t buy that. Now when asked if you can hit something far more critical, the sRGB spec for cd/m2 and chromaticity, that isn’t all that important to the debate.
2.2 is the simple gamma curve, sRGB TRC is the TRC that is not a pure simple gamma curve.
The sRGB spec for cd/m^2 is not even something you'd want in all viewing environments. If it is a bright day and the windows are open, come on. When it's dark and controlled they I do tend to set it around 80 cd/m^2.
It depends upon the monitor, but for some when you toggle between profile on and off, the biggest visual difference actually could be gamma 2.2 TRC vs. sRGB TRC and not the other aspects (well other than perhaps the white point, of course). I mean I've tried it and seen. It depends on the monitor though. Many profilers, most even, struggle somewhat to fix up saturation tracking all that well for the monitors where that is not so hot.
And for one of the better wide gamut monitors, they can get all specs so dead on that the TRC is the only thing left. If I put it in sRGB emulation mode with monitor calibrated to gamma 2.2 and compare images between IE and Firefox the only differences, whatsoever, that you see is between IE not correcting the sRGB image's display for the monitor being gamma 2.2 instead of sRGB TRC.
The bottom line is, there is yet any justification for either the 2.2 sRGB TRC setting nor the bug it introduces in the software.
Funny that the ICC and so many others have even bothered storing TRC information then.
So your take is, no matter what values I dial in, every instrument is going to provide that exact value? It might report what you asked for, I’d be hard pressed to believe you’ll always get it.
Of course not, but with that attitude why even bother calibrating at all? And when it comes to TRCs, if you plot the measurements of a monitor in gamma 2.2 and one in sRGB TRC with a bunch of probes, most probes actually do show a straight line across with 2.2 and a curved one for sRGB TRC. Even if the plots don't quite overlap the general shape, from what I've seen, is generally pretty close. They tend to get that better than the absolute values of brightness, which don't matter as much, in agreement.
And it might be just a feel good option or something placed into the product solely for a marketing competitive matrix diagram.
Toggle between viewing an sRGB image as gamma 2.2 or as sRGB TRC and tell me it's just marketing BS. With many images the difference is plenty easy to see.
It's not as huge as swapping D50 and D93 or something and on some monitors the rest of calibration sometimes matters even more (but definitely not always) but it's still readily visible so why is the rest important and this part just marketing BS?
-
Not sure what you are saying here.
I'm not sure many calibration/profiling programs even allow that do they?
Yes, quite a number do, including SpectraView and at least the last number of generations of X-Rite products.
Many monitors have TRCs going all over the place along the way are something way different than anything close to gamma 2.2.
Which is why we have the Native Gamma option. Why alter the behavior as you propose to sRGB when you can leave it alone and just characterize it with the profile.
And if it is some messy random TRC how will the profile even characterize it?
Of course it will.
Whatever the differences you can see the difference between sRGB TRC and gamma 2.2 in either case so why avoid it?
Yes, as I said, you can see the difference in all kinds of calibration targets. So what? Calibration should be set to provide a visual match to a print in most cases (that is what most users desire). The gamma isn’t going to play a role here in ICC aware applications. For years, prior to OS X, Mac users calibrated to a Gamma 1.8 because outside ICC aware applications, the OS made that silly assumption. But inside ICC aware applications, it made no difference. Smart users calibrated their displays to 2.2 because that if far closer to the native gamma of the display and, there was no visual difference in ICC aware applications plus closer to native meant less banding (the reason Native has been an option for years in better software products). IF you cared about how non ICC aware app’s previewed their inaccurate color (cause they are not ICC aware), you used 1.8 otherwise everything appeared a tad dark (yes, there was a visual difference, so what?). But IN ICC aware applications, 1.8 or 2.2 produced the same previews, the former however introducing more banding in many display systems.
2.2 is the simple gamma curve, sRGB TRC is the TRC that is not a pure simple gamma curve.
You are preaching to the choir! If you carefully re-read what I wrote, you’ll see I’ve gone out of my way to separate language discussing a simple gamma curve with a TRC curve when discussing the two.
The sRGB spec for cd/m^2 is not even something you'd want in all viewing environments.
I question in this post, if you’d even really want it (or can actually really hit it). You seem to feel otherwise for reasons I as yet don’t understand.
If it is a bright day and the windows are open, come on. When it's dark and controlled they I do tend to set it around 80 cd/m^2.
As I said, if you have such conditions, you have bigger problems than what gamma to calibrate to! You should control you editing conditions so they are always consistent and controllable.
It depends upon the monitor, but for some when you toggle between profile on and off, the biggest visual difference actually could be gamma 2.2 TRC vs. sRGB TRC and not the other aspects (well other than perhaps the white point, of course). I mean I've tried it and seen.
First off, seeing a before and after difference tells us nothing about the quality of the ‘after’ calibration. Whether it meets sRGB specifications or matches a print. I don’t know why you keep going back to seeing a difference in two settings. That is totally to be expected. Other than there are differences, the results thus far say nothing further.
Look, if you want to calibrate to a setting called sRGB and you get the match you want, great. But to propose that using that setting in some why makes your LCD display produce sRGB specified is not something I’m buying.
-
Yes, quite a number do, including SpectraView and at least the last number of generations of X-Rite products.
Oops you are correct.
Which is why we have the Native Gamma option. Why alter the behavior as you propose to sRGB when you can leave it alone and just characterize it with the profile.
Why not leave it at native gamma? Maybe because it's nice to have the 95% of programs that are not color aware look nice? It is sure nicer using the monitor to watch TV/movies/play games/edit in premiere pro/watch videos/for the times you use IE instead of Firefox/desktop/3D programs/games/etc. So why shouldn't someone calibrate to sRGB TRC for some of that stuff and to gamma 2.2 for the rest? Why leave it to just a few color aware programs and have to trust the CMM in all of them?
And in this case here, he has a wide gamut monitor and most likely the only reason he is even in the sRGB emulation mode is to deal with non-managed software. For the most part, when you are using an sRGB emulation mode on a wide gamut monitor it is because you are dealing with non-color aware programs, otherwise you probably are going to want to have it in native gamut mode.
Of course it will.
Hmm yeah I guess maybe it can also be stored as a table form too which would work.
I will have to look into it more for native gamut photo editing modes. I'm not sure it makes any difference on my PA though since I don't see any banding due to the high bit internal LUT anyway. Perhaps it might improve contrast ratio or something.
Yes, as I said, you can see the difference in all kinds of calibration targets. So what? Calibration should be set to provide a visual match to a print in most cases (that is what most users desire).
The gamma isn’t going to play a role here in ICC aware applications.
Yes, but what about for all the non-ICC aware stuff?
(and in the OP's case, if he is in the sRGB emulation mode, he very well may be in that mode specifically because he plans to use non-ICC aware programs)
For years, prior to OS X, Mac users calibrated to a Gamma 1.8 because outside ICC aware applications, the OS made that silly assumption. But inside ICC aware applications, it made no difference. Smart users calibrated their displays to 2.2 because that if far closer to the native gamma of the display and, there was no visual difference in ICC aware applications plus closer to native meant less banding (the reason Native has been an option for years in better software products). IF you cared about how non ICC aware app’s previewed their inaccurate color (cause they are not ICC aware), you used 1.8 otherwise everything appeared a tad dark (yes, there was a visual difference, so what?). But IN ICC aware applications, 1.8 or 2.2 produced the same previews, the former however introducing more banding in many display systems.
Again, what about when someone flips on a tv show or a movie or to a game or does some browsing and looking at other people's images and happens to be using one of the many non-fully aware browsers?
And I don't think it is a so what if IE shows things with the wrong TRC. You can start telling people to lift their shadows a little and end up giving bogus advice.
As for banding and such it depends whether it's an internal high bit LUT monitor like the OP has or someone relying on an 8bit graphics card LUT. And if you are dealing with non-iCC aware it's all moot since you need it to be calibrated and not merely profiled.
You are preaching to the choir! If you carefully re-read what I wrote, you’ll see I’ve gone out of my way to separate language discussing a simple gamma curve with a TRC curve when discussing the two.
That's what I thought but then you said something about how I insisted things need to be set to the 2.2 sRGB gamma instead so....
I question in this post, if you’d even really want it (or can actually really hit it). You seem to feel otherwise for reasons I as yet don’t understand.
I thought you were the one who was insisting it had to be hit since, if not them why even bother with what the TRC is, etc. etc., maybe we misunderstood each other.
And you can hit it with some monitors. I have it in sRGB emulation mode at sRGB TRC and 85cd/m^2 right at this moment and it is just about near min black level, the contrast ratio is a trace, TRACE lower than at 120 cd/m^2 but so what? the black level is much deeper and my eyes are not getting burned out and the calibration actually tested out to be a bit more uniform than the one made at 120 cd/m^2 so, no, it has not gone all whacked out. I'm in a dim room right now, why do I want a high black level, faded looking image and my eyes burned out?
First off, seeing a before and after difference tells us nothing about the quality of the ‘after’ calibration. Whether it meets sRGB specifications or matches a print. I don’t know why you keep going back to seeing a difference in two settings. That is totally to be expected. Other than there are differences, the results thus far say nothing further.
I keep going on about how you can see the difference between an sRGB image displayed with sRGB TRC vs. gamma 2.2 TRC because you said the difference was nothing and like dancing on the head of a pin and that he shouldn't even bother setting his sRGB emulation mode to sRGB TRC for general work. I was just saying that, to me at least, it is a lot more noticeable than that. Now all of a sudden you seem to say that the difference is noticeable.
But to propose that using that setting in some why makes your LCD display produce sRGB specified is not something I’m buying.
I'm saying that if you are switching from photo editing mode to sRGB emulation mode to do say perhaps some browsing on the web using IE that setting it to sRGB TRC does get you proper sRGB image display while setting it to gamma 2.2 makes it a little bit off and that with many, if not all images, the difference definitely can be seen easily enough even if it is not radically extreme or anything.
-
Why not leave it at native gamma?
No reason not to.
Maybe because it's nice to have the 95% of programs that are not color aware look nice?
Non ICC aware applications simply have no way to show you color appearance consistently and correctly so I don’t worry about them. I might prefer they were color managed however.
So why shouldn't someone calibrate to sRGB TRC for some of that stuff and to gamma 2.2 for the rest?
You are discussing calibration and profiling which work hand in hand, then asking why applications that don’t understand ether should affect how we calibrate and profile FOR ICC aware applications. It is pointless.
And in this case here, he has a wide gamut monitor and most likely the only reason he is even in the sRGB emulation mode is to deal with non-managed software. For the most part, when you are using an sRGB emulation mode on a wide gamut monitor it is because you are dealing with non-color aware programs, otherwise you probably are going to want to have it in native gamut mode.
Yes and I would submit that any combo of gamma, or for that matter calibration settings for the emulation could work equally well. You keep suggesting that the TRC 2.2 gamma is necessary, others are less good. You haven’t provided anything to prove that point. I keep saying that working with non ICC aware applications, all bets are off. Futz around with any settings you want so the non ICC aware app ‘looks good’. There is a significant differences in are disagreements in those two cases. Gamma 2.2, sRGB TRC, gamma 1.9.3, whatever.
Yes, but what about for all the non-ICC aware stuff?
I told you, the previews are science fiction. They don’t know squat about the color space of a document. They don’t know squat about the ICC profile of the display. You might as well just futz around with all the goofy OSD controls and make the display look pretty and move on. A gamma setting is simply not worth considering.
As for banding and such it depends whether it's an internal high bit LUT monitor like the OP has or someone relying on an 8bit graphics card LUT.
Yes it does. That doesn’t change anything about viewing color in non ICC aware applications nor that a sRGB TRC is ‘better’ than anything else.
-
I think the real problem is perhaps that you are so used to thinking of a monitor as nothing more and nothing less than a device on which to proof prints using ICC aware software and that you are too stuck in that mindset.
No reason not to.
Non ICC aware applications simply have no way to show you color appearance consistently and correctly so I don’t worry about them. I might prefer they were color managed however.
I suppose that all the people who calibrate their HDTVs are just wasting their time then?
I guess you like to watch movies using whatever horrid default settings most sets are shipped with then?
You seem to forget that most images and colors on the web are in sRGB gamut and assume sRGB TRC and that most video and movies and TV stuff is in sRGB/REC709 gamut and that one can very well worry about them in non-managed apps considering that things like windows media center, powerdvd, externla blu-ray players, Internet Explorer, desktop, etc. etc. are not ICC aware since you know exactly what the source expects as the destination.
If have a regular monitor with no fancy internal LUT, you should still try to do what you can do bring it closest to sRGB+sRGB TRC or to sRGB+gamma 2.2 (or whatever matches video/tv/movie viewing best under the circumstances). Some programs still make use of the graphics card LUT so at least white point and TRC might be correct and for the ones that are not some things. When talking about HDTV instead some have fairly extensive calibration controls.
In this case the OP has a NEC PA, you can dial in a calibration inside the monitor and any program will make use of the calibration, anything you plug into the monitor will. In fact, as I've said, and even shown, a non-calibrated program fed sRGB gamut material will result in your seeing the material closer to spec than if you using ICC aware programs on some standard sRGB monitor that isn't even capable of hitting sRGB primaries and where the profile might not correct things as well as the PA series does things due to its extreme linearity and 3D high bit LUT.
You are discussing calibration and profiling which work hand in hand, then asking why applications that don’t understand ether should affect how we calibrate and profile FOR ICC aware applications. It is pointless.
When did I say that it affects how we calibrate for ICC aware applications?
Yes and I would submit that any combo of gamma, or for that matter calibration settings for the emulation could work equally well. You keep suggesting that the TRC 2.2 gamma is necessary, others are less good.
Wow when have I been saying that a TRC gamma 2.2 is necessary for sRGB image viewing??
I keep saying sRGB TRC is not that gamma 2.2 TRC is. You are the one who keeps saying gamma 2.2 TRC is good enough. And then when I start talking about sRGB TRC vs gamma 2.2 TRC you tell me that of course you know that sRGB isn't defined to use gamma 2.2 TRC but here you are again.
If you are, say are using IE to browse, i.e. a web browser than does not understand monitor profiles at all, then yes calibrating to sRGB TRC would be better since then it would show things with the proper response and it would look a little bit more accurate than if you had calibrated the monitor to gamma 2.2 and perhaps vastly better than if you left it set to native gamut.
(I also have a very sneaking suspicion that a few color aware apps appear to take a short cut if they spot an sRGB source image and seem to skip any TRC compensation steps and assume gamma 2.2 is good enough and assume your display is gamma 2.2, granted that may be on them for not carrying out all steps. But it could be a reason, if you use such programs, to sometimes also calibrate to sRGB TRC instead of gamma 2.2 or native gamut. I'm not sure what they are doing differently, but the tone response doesn't match what say photoshop or firefox or irfanview do and the author of one appear surprised to hear that sRGB images were not encoded as gamma 2.2)
You haven’t provided anything to prove that point. I keep saying that working with non ICC aware applications, all bets are off. Futz around with any settings you want so the non ICC aware app ‘looks good’.
For non-managed programs you to get everything as close as you can. If you can get the white point and TRC correct then at least do that.
With something like the OP has, a fancy monitor with a 14bit 3D LUT that can be programmed into a beautiful sRGB emulation if you then go and view an sRGB image using photoshop and then with IE, guess what they will look the same if you calibrated the mode to sRGB TRC.
The monitor carries out the mode so well that the image will look better in IE than it would in photoshop using a fancy monitor profile on some whatever monitor, even.
If you set it to sRGB gamut, D65, gamma 2.2-2.4/REC 709 TRC or what not you can then view TV/movies in beautiful fashion, it doesn't matter if the player programs knows a single thing about color management. The set it perfectly calibrated internally to accept the video.
I told you, the previews are science fiction.
what previews?
They don’t know squat about the color space of a document. They don’t know squat about the ICC profile of the display. You might as well just futz around with all the goofy OSD controls and make the display look pretty and move on. A gamma setting is simply not worth considering.
Utter nonsense, tell that to an HDTV calibrator.
And in this case, regarding the OP, it's particularly nonsense since his monitor has such a near perfect emulation mode (the set can be 100% internally calibrated to match spec better than any regular monitor would do even using ICC aware software). If you feed it an sRGB image or DVD or bluray using a non-color management aware program it most definitely will look correct and not like some random choice.
That doesn’t change anything about viewing color in non ICC aware applications nor that a sRGB TRC is ‘better’ than anything else.
it does if the goal is to say use IE to surf images on the web, or use some 3D program that assumes sRGB display conditions, especially in his case, where the monitor can have all the rest perfectly calibrated on top of just the white point and trc.
if he is viewing a movie then he probably does not want to use it calibrated to sRGB TRC though and probably want to use the broadcast video mode. He may wish to setup in SV II a Broadcast Video mode with sRGB gamut, something like gamma 2.2, D65 (for TV/movies/video/certain games and programs) and then he may wish to setup an sRGB emulation mode with it set to sRGB gamut, sRGB TRC, D65 (for web browsing/some games and programs) and a setup a Native Gamut mode with native gamut, D65 and maybe gamma 2.2 or native gamma (for photo editing/viewing/print proofing in color managed software) or maybe even a special print proof mode where he programs the set to be closer to his paper/ink/printer combo (depends whether that or softproof in the software works better), etc.
-
I think the real problem is perhaps that you are so used to thinking of a monitor as nothing more and nothing less than a device on which to proof prints using ICC aware software and that you are too stuck in that mindset.
No, I recognize there are two ways to treat the previews your display shows you. One is correct and based on color management, the other isn’t. I don’t look at the one that isn’t and wonder what those RGB values will look like on a piece of paper. If they don’t look correct, I don’t worry about it, I understand the process isn’t color managed. I’m not expecting a match. I
I suppose that all the people who calibrate their HDTVs are just wasting their time then?
IF they prefer the color they see, great. They are not trying to match the HDTV to something else, they have zero control over altering the RGB values as we have to do in imaging applications.
I guess you like to watch movies using whatever horrid default settings most sets are shipped with then?
No, I don’t prefer horrid defaults over better appearing defaults. But the image I see isn’t something I have to edit or reproduce to another media.
You seem to forget that most images and colors on the web are in sRGB gamut and assume sRGB TRC and that most video and movies and TV stuff is in sRGB/REC709 gamut and that one can very well worry about them in non-managed apps considering that things like windows media center, powerdvd, externla blu-ray players, Internet Explorer, desktop, etc. etc. are not ICC aware since you know exactly what the source expects as the destination.
Not necessarily, and even if I agreed every image is in sRGB (which is ridiculous), outside of an ICC aware method of viewing that data, sRGB is meaningless. I can alter the OSD controls and make a pretty picture just like I can with my HDTV. There is zero rationale to suggest I use a specific TRC or a TRC that matches sRGB as you’ve proposed. Just turn the knobs to make everything look nice. The actual TRC is meaningless, it looks good.
In this case the OP has a NEC PA, you can dial in a calibration inside the monitor and any program will make use of the calibration, anything you plug into the monitor will.
Yes you can. You suggest the gamma should be an sRGB TRC, I say it doesn’t matter. The data viewed is uncalibrated, non ICC aware. Whatever setting anyone visually prefers is AOK. There is nothing here that suggests it must be a 2.2 TRC that mimics sRGB. Further, I submit nailing sRGB isn’t necessary.
Wow when have I been saying that a TRC gamma 2.2 is necessary for sRGB image viewing??
I said in the first post, the sRGB calibration target (which causes a bug) is unnecessary, you appear to disagree and off we go.
To sum it up: The Gamma setting isn’t important in ICC aware applications. Outside ICC aware applications, anything is game. The sRGB or 2.2 setting could equally make one group prefer the non color managed preview on a display over the other. It doesn’t matter.
-
No, I recognize there are two ways to treat the previews your display shows you. One is correct and based on color management, the other isn’t. I don’t look at the one that isn’t and wonder what those RGB values will look like on a piece of paper. If they don’t look correct, I don’t worry about it, I understand the process isn’t color managed. I’m not expecting a match.
It is based on something when you know what the source will be. If you are aiming only for sRGB image display or only for ATSC TV or DVD or blu-ray, yes you do know exactly what the source will be.
And yes you can tune some HDTVs pretty well, as with most sRGB-like monitor they fall a touch short hear and there of hitting the primaries or this or that but you can really get some of them quite close and yes it does make one heck of a difference.
Granted some you can't really do much to and they may be stuck with very far off saturation curves way over pumped primary luminance gamma 2.9 or who knows what and then you are stuck.
But to say there is no point in calibrating something has has 10pt white balance and TRC controls and a full CMS is ridiculous.
And again in the case in the thread the OP has a NEC PA! You can dial in the primaries perfectly! You can dial in the TRC as you wish. You can set the white point. I can get it closer, all internally calibrated to spec for say watching DVDs than I could get my old monitor even using ICC aware programs and CEDP!
The entire point of the 14 bit 3D internal LUT and sRGB emulation ability is so you can tune it in, do the whole thing internally and then it will work perfectly with ANY program or source that sends it the sort of material it got calibrated too.
IF they prefer the color they see, great. They are not trying to match the HDTV to something else, they have zero control over altering the RGB values as we have to do in imaging applications.
WHo says they have zero control? Have you see the controls on a modern HDTV? And yes they sure are trying to match it something specific.
And again with the NEC PA that the OP has, he has every control he could possible need. He can get it set for truer display of those sources than someone could with any consumer standard gamut monitor could evenif they used the most advaned profiling software and wrote idealized custom color management aware video software.
But the image I see isn’t something I have to edit or reproduce to another media.
I'm not sure what you mean by that.
Not necessarily, and even if I agreed every image is in sRGB (which is ridiculous), outside of an ICC aware method of viewing that data, sRGB is meaningless. I can alter the OSD controls and make a pretty picture just like I can with my HDTV. There is zero rationale to suggest I use a specific TRC or a TRC that matches sRGB as you’ve proposed. Just turn the knobs to make everything look nice. The actual TRC is meaningless, it looks good.
1. by far most images on the web and most sites are set up as sRGB and sure there are a few AdobeRGB and whatnot around and yeah a totally unmanaged browser will mess those up, however, pretty much all browsers at least translate everything to sRGB even if they do no more so they would still look OK even in IE.
2. There is not zero rationale. The rationale is to MOVE the primaries to sRGB primary locations! And to set everything else as needed. With the NEC PA everything is set internally. If it is set to the standard it darn well does something for you when you use IE to browse or Media Center to watch TV etc.
To sum it up: The Gamma setting isn’t important in ICC aware applications. Outside ICC aware applications, anything is game. The sRGB or 2.2 setting could equally make one group prefer the non color managed preview on a display over the other. It doesn’t matter.
Anything is not game when you know what the source will be and if the display can be made to approach the source. It's called basic class one sRGB color management where the input is assumed to be in sRGB and the output device automatically handles things as sRGB.
With some HDTV you can adjust them to be pretty close to the source and with the NEC PA you set it as exactly as you can.
Do you know what I see when I set my PA to sRGB gamut, D65, sRGB TRC and then view an image in a non-managed viewer and in photoshop? Zero difference!
You know what I see when I view a DVD in a non-managed player and in a managed one when I set it to sRGB gamut, D65 and gamma 2.2 Zero difference! The color management is being done only in this is has been able to have been done 100% through device calibration.
Do you know what I see when I set my PA to sRGB gamut and then shift the red primary way up, D93, gamma 2.8 and then view an sRGB image in a managed and un-managed viewer? a big difference and even with the managed viewer it still won't look quite as it did in the earlier situation if the image contains certain colors since the monitor has been set so that it can't display all sRGB colors now. In fact this means that using a color managed viewer now the image won't even be able to be shown in all it's glory and it will need some form of gamut mapping or clipping and yet in the earlier case, with an unmanaged program, it would display everything as it should with no clipping.
Anyway this is getting tiresome, so I am done with it. You may know a real lot when it comes to certain scenarios but you definitely are off-base when it comes to other scenarios and there are a lot of HDTV calibrators and color science guys who would be a bit shocked by some of things you have said. Although maybe it's just some sort of misunderstanding.
-
Hey there,
just playing arround with i1Profiler, but there seems to be something strange, would be very nice if someone could confirm this behaviour or bug, to check if there is a problem on my end..
I try to calibrate with standard settings (D65, flare off, ambient light off, ADC off, Bradford, ICC v2, Matrix based ) but with "sRGB" tonal response curve instead of gamma 2.2!
With gamma 2.2 everything works fine, but as soon as I use sRGB at the end of the measurement the screen keeps white and nothing happens anymore :'(
Tried with i1Profiler 1.2.0 on a Win 7 x64 machine with i1Display pro.
Do you experience the same problem?
Thanks in advance!
Wow I just realized that in all our arguing we actually came up with the solution for your problem.
I hadn't used the new i1Profiler yet but as pointed out, it offers the option for native gamma! Since you want to use Multiprofiler to set as much internally as possible and it does allow for TRC settings there is no need for the i1Profiler to be messing with the TRC at all! Do the TRC stuff entirely in MP and leave i1profiler set to native gamma so it doesn't meddle with the TRC, why woulkd it need to calibrate to sRGB TRC or gamma 2.2 if the monitor is already there? And you'd want the TRC to be set inside the monitor to take advantage of the high bit internal LUT rather than the graphics card LUT which is low bit and won't end up getting applied to blu-ray or games. Since you are making a v2 profiler there would be no option for it to have recorded gamma 2.2 or sRGB TRC by function and I guess it would have to be LUT sampled TRC profiling anyway.
So all you do is use multiprofiler to instruct the set to set itself to sRGB TRC and then leave i1profiler on native gamma and then since the native gamma was made to be sRGB TRC that is what you'll have.
So you can use i1profiler to measure a white patch while using whitepoint xy adjuster in Multiprofiler until you hit .313,.329 reading in i1profiler for white point
and then you can use multiprofiler to set sRGB gamut (perhaps also use the i1 Display pro to measure pure red 255,0,0 g 0,255,0 b 0,0,255 patches and adjust in MP the primary locations until they hit the sRGB primary xy coords, if you feel like bothering, for a new set it might not really help anything)
and use multiprofiler to set sRGB TRC (although you may want to make a second custom sRGB emulation mode where you chose gamma 2.2 instead since that would work out better for tv/videos/movies/some games and flip between the sRGB emulation versions depending upon what you are doing)
and then you can run iprofiler set for D65, native gamma, matrix-based and it won't really do much other than associate an ICC profile for the monitor's current internal settings for the screen. So basically how you had it set only change i1profiler to native gamma mode, there is no need for it to calibrate the TRC since the monitor can be programmed to that itself with multi-profiler.
For photo editing/viewing you'd want to make a native gamut programmed mode using multi-profiler. Set MP for D65, native gamut, either gamma 2.2 or perhaps native gamma is better. Set i1 profiler for D65,matrix-based, native gamma.