confusing color-management mess all around, each program has it's own mind

Started 8 months ago | Discussions thread
ForumParentFirstPreviousNextNext unread
Flat view
bronxbombers4
Senior MemberPosts: 2,841
Like?
confusing color-management mess all around, each program has it's own mind
8 months ago

A recent post by Earthlight where his findings were NOTHING like mine at all and what I'd been saying for a few years got me thinking and testing after I tried some of his tests and suddenly got some results more like his (I had switched from a NEC PA to a Dell and been using xrite/Dell profiles instead of SV II NEC profiles since).

OK, wow I discovered an utter mess!

It not just depends upon what program's CMS is used but upon how each program's CMS seems to react to different profiles as well! So although a program might act one way with a NEC SV II profile it might act totally the opposite way say an xrite/Dell made profile! And even worse it even varies upon what settings you chose for calibration with some brand profiles!

So basically all of my talk all this while about gamma 2.2 vs. sRGB TRC and how some programs handle it correctly and others not and how even if you have a regular gamut monitor IE will show things wrong while Firefox will show them right is apparently ONLY valid if you were monitor profiles made by certain programs since some of the results I find using Dell profiles instead of NEC seem to 100% reverse how some of the programs act!

What a mess!!!!!!!!!!!!!!!!!!!

Here are some crazy findings, make of them what you will, this is incomplete and preliminary:

NOTE: For Dell profiles I refer only to vs matrix choice type and internal Dell monitor calibrations done through the Dell/xrite software combo. I didn't try table or v4 types yet.

1. FastPicViewer's native full speed built-in CMS seems to do VERY weird things to tone response. It tends to add a LOT of contrast and really crush blacks and just generally shows every image no matter what else is going on incorrectly. I would NOT use it this way at all! Beware it is the default setting!

2. Switching to the Windows WCS CMS option in FastPicViewer tends to make it behave but not always, sometimes worse.

When used with SV II profiles it still, in almost all cases, tends to make dark tones a bit darker and adds a touch of overall contrast compared to PS (other than for the trivial scenario where you calibrated the monitor to sRGB TRC and are viewing sRGB images). For sRGB images it appears to fail to translate from sRGB TRC to gamma 2.2 when presented with a gamma 2.2 SV II profile for a monitor calibated with SV II to gamma 2.2.

For some reason it might also make things a bit too punchy when using a ProphotoRGB image as well even though those are stored as gamma 1.8.

When used with Dell profiles, however, it starts to become closer to PS, I think, being full screen only it's a bit harder to compare FPV to other programs. Anyway this seems to be the case at least so long as the Dell is calibrated to a native sRGB TRC (trivial case) or to gamma 2.2 and sRGB.

Very oddly if I calibrate my Dell to native gamut and settings it seems to make everything faded out in this mode and doesn't even use 0,0,0 for 0,0,0 but some lighter shade to start out from whne viewing sRGB images, no other program seems to behave this weird way at all. It seems totally hosed in this case and uses a totaly wrong tone curving for some reason.

Not quite sure of all of the details yet. There appear to be very few scenarios where it will show images the same way PS does though.

***************************

3. Irfanview seems to be relatively well behaved using NEC SV II profiles and seems to account for sRGB TRC vs gamma 2.2, if required. Although it almost seems to account for it trace TOO much becoming even a trace brighter than PS now for such a case of viewing an sRGB image on a monitor calibrated to gamma 2.2. (while FstPicViewer is a bit too dark and punchy and I'd say farther off for sure than Irfanview). I guess it's still mostly not quite like PS but at least usually pretty close I think.

********************************

4. If I calibrate NEC to sRGB + sRGB TRC or set Dell to sRGB precal mode and set the monitor profile as sRGB's definition (or use the NEC SV II profile) and view sRGB imgaes:

then basically all sRGB images look correct and look the same in any program since basically they all decide to turn off CMS (or don't have one to begin with) since the profiles match monitor to image. My special sRGB gray scale ramp (created so that the ramp steps match sRGB 8bit steps which is different than just st

********************************

5. If I calibrate Dell to sRGB + gamma 2.2 and set the xrite/dell profile and view sRGB images:

it seems like Irfanview makes more steps black than PS and overall it makes things a touch darker and more contrasty, IE even a touch more so and oddly Firefox even just a hint even more still.

Chrome also makes things a touch too dark.

Oddly FPV might actually be a bit closer to PS this time.

It seems like just about all programs other than PS (and maybe, maybe FPV) seem to for some reason decided not to bother doing a translate from sRGB TRC to gamma 2.2 compensation step when using the xrite/Dell sRGB+gamma 2.2 profile.

I guess in this case PS gets a pass and FPV maybe gets a pseudo pass and everything gets a fail.

I didn't check what happens yet with ProphotoRGB images under this scnerio.

**************

6. If I apply the NEC SV II sRGB+gamma 2.2 profile and view sRGB images:

irfanview, firefox and chrome do appear to carry out this translation to account for sRGB TRC vs gamma 2.2 differences along with PS. It seems that irfanview, firefox, chrome actually slightly slightly OVERdo the compensation and come out a trace brighter in very darkest shadows and actually all along most of the gradient than PS though (with Irfanview seeming to overdo it the least of the three).

In this case though FPV appears to fail to do this translation and comes out darker and punchier looking than PS and seems farther off from PS than Firefox/Chrome/Irfanview.

IE of course does not do this since it never does anything other than convert any non-sRGB image to sRGB and definitely comes out darker and punchier looking compared to PS.

I guess in this case PS gets a pass and irfanview, chrome, firefox get pseudo passes and FPV and IE get fails.

I didn't see what happens yet if a ProPhotoRGB images is viewed in this scenario.

*******************************

7. If I calibrate the Dell to native gamut and native settings (which I thought were gamma 2.2, but I'm not so sure anymore about that gamma) and set the xrite/Dell profile and view sRGB images:

the gradients looks the same in PS as it did when I had the monitor calibrated to sRGB+sRGB TRC (in the gamma 2.2 cases above PS shows it instead with the bottom two tones 0,0,0 instead of just the very bottom one) with only the darkest bar as 0,0,0

irfanview/firefox/chrome seem to show with a touch more contrast than PS and all seem to end up with about the lowest four bars as 0,0,0 instead of the lowest two for PS

IE very oddly seems similar to PS for gradients now somehow (not sure how this can be!) but goes way saturated of course for images that are not gray scale gradient tests

FPV goes crazy and sets 0,0,0 to something lighter than pure black
so shadows a bit too light and punch is a bit less for sRGB images (I didn't test it but I seem to recall that prophotorgb images though end up a bit too dark in shadows and punchy with it in this case though!)

I guess PS passes and most of the rest fail with FPV and IE mega-failing.

ForumParentFirstPreviousNextNext unread
Flat view
Post (hide subjects)Posted by
alsoNew
ForumParentFirstPreviousNextNext unread
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow