- This topic has 161 replies, 25 voices, and was last updated 1 month, 1 week ago by
.
-
Posts
-
How do I proceed in doing a profile only run? Will the embedded gamma overwrite the lut from novideosrgb?
One more question regarding the profiling only run on top of calibration after applying the initial profile in novideo: In this second profiling run, everything should be As Measured, but what display profile should DisplayCAL be using for the profiling run? I gather it should be the simulated target, correct? pure sRGB, for example?
So, the final workflow would look like:
- Calibrate and profile using DisplayCAL’s sRGB preset without applying the profile
- Clamp to sRGB and gamma 2.2 using the generated profile in novideo.
- Run the second profiling setting DisplayCAL to use pure simulated sRGB profile as display profile (not the profile created at point 1) and everything as measured.
Is this correct? And, what’s the state of black point compensation in the 3 steps? I think it should be ON in steps 1 and 2, and OFF in step 3.
I hope an answer to this can put together all the pieces in the puzzle, because even the README doesn’t cover the workflow top to bottom. Thanks!
Curious if this tool has been invalidated by the advent of Windows Advanced Color Profile, has anyone tested both. Limitations? As first glance it seems like ACM is achieving the same result, and I misunderstanding?
I tried both methods a few months ago. My conclusion at the time was that the Windows tool gave inconsistent results depending on the application and seemed to still need to solve many compatibility issues. In the end, I went back to novideo_srgb, as it has never failed me and is even compatible with full-screen 3D applications.
Hello. I am photographer. I have Dell U2724D and X-Rite ColorMunki Photo.
I made two profiles:
Instrument mode – LCD (generic) Adaptive HiRes) | Profile type – XYZ LUT + Matrix | White point – 6500K | Tone Curve – 2.2 | Profile quality – High | Amount of Patches – 596
and the profile that the author recommends
Profile type – Curves + Matrix | White point – 6500K | Tone Curve – sRGB | + add TestChart with 256 Neutral patches.
As a result, I got slightly better measurements with the profile XYZ LUT + Matrix. After I select the created profile in novideo_srgb (gamma calibration to srgb) and selecting the profile sRGB IEC61966-2.1 in the OS. Then I made a Extended verification testchart with setting – current and get the next result. Can you check if everything is okay, because I’m a little concerned about the maximum ΔE of 2.57 for one color and the gamut of up to 30%. I did the same thing on two different monitors and always got an error of about ΔE 2.6 on the same color. If check the profile without the novideosrgb program, everything is fine, the maximum error is ΔE 0.63.
ThanksCalibrite ColorChecker Studio on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.novideosrgb uses GPU lut-matrix-lut hardware, hence the volumetric info in the 3d mesh of XYZLUT profile gets ignored, only using the matrix info.
Both calibrations will store the same or almost the same calibration. The only diference is the after calibration “taylot made suit” measureing how it behaves: idealized (matrix) vs volumetric 3d mesh (xyzlut)
Everything is as you say. The difference between the profiles is minimal. I understand that everyone has their own acceptable level of deviations, but I am interested is my result is not bad for an average photographer. Maybe I made a mistake at some point.
Why are you even using workarounds like novideo_sRGB instead of proper color management as a photographer?
I like that novideo_srgb works on the GPU side and I am sure that everything I see on the monitor will be displayed correctly. I have a monitor with increased sRGB coverage. I decided for myself that it is better to limit it using novideo_srgb, and not to enable sRGB emulation, when emulating, I have a white point of 7000 and there is no way to control the RGB gain. But maybe I’m wrong. I read here that Windows Advanced Color Profile has been released, but I don’t know anything about it yet. Is it better to use Windows Color Profile and sRGB simulation?
Why are you even using workarounds like novideo_sRGB instead of proper color management as a photographer?
+1,
@Valchuk0
almost every image editor is color managed hence there is no need to simulate sRGB… and if you need to do that and aim for accuracy, nobody will use novideosRGB.If display is bad behaved and needs XYZLUT description of its behavior maybe it would be easier to use DWMLUT to simulate sRGB or an idealized widegamut description of your display (and use a simplified synth profile as display profile).
sRGB simulation like you did may be interesing for non color managed apps like some CAD/CAM/game design or to avoid color management rounding errors in browsers or incompatibilities (it’s unlikely they will support XYZLUT profiles)
Why are you even using workarounds like novideo_sRGB instead of proper color management as a photographer?
@Valchuk0
almost every image editor is color managed hence there is no need to simulate sRGB… and if you need to do that and aim for accuracy, nobody will use novideosRGB.“Unless gaming” or other GPU performance related tasks.
Thank you very much for the detailed answer. And how do you fight with oversaturated colors in DaVinci, YouTube, in video players? To be honest, even the Windows interface does not look good for me)
Downloaded the last DWM_LUT from Lauralex, but it doesn’t work on my latest version of Windows 11.
The funny thing is that when you think you have already understood something, you get even more confused)Resolve is color managed and you should have no trouble looking up the specifics of how to do it (you’ll need a LUT file, not just an ICC profile). Browsers are generally color managed as well these days and by default use your Windows color management settings AFAIK. Windows interface didn’t used to be color managed, but there’s the new automatic color management feature in 24H2.
But yes, if you want to clamp Windows UI or a browser to sRGB, then I guess novideo_sRGB is a reasonable enough choice. I’d just not use it for color critical work in color managed applications, at least not without using a color profile on top of it.
Would love a confirmation on this too, for my second profiling should I profile towards the original run I did for making the icc for novideo or do a generic 2.2 profile only run?
And followup, is using a synthetic gamma 2.2 profile from displaycal good enough, or could that cause an additional gamma shift.
And if I choose to do a second profiling, should I disable black offset, it seems to add an additional level of black crush in ICC managed apps with it enabled.
When I try to check accuracy with the synthetic profile, most of the color values are crazy off.
@noiremk2 Set your monitor profile in Windows to sRGB. Profile your monitor. Set that profile in novideo_sRGB. Enable clamp. Profile the monitor again. Set this new profile as your monitor profile in Windows.
