- This topic has 51 replies, 4 voices, and was last updated 3 years, 1 month ago by
.
-
Posts
-
Photoshop is not limited to sRGB images, regarding colospaces arbitrary means anyone. Application works in the same way, with the same code, colorspace agnostic.
It transforms from one colorspace (image), whatever it is, to another colorspace (display, or printer, or convertirng to other output file with another colorspace ).Your display colorspace IS NOT “EXACTLY” sRGB, so no identity/null transformation is possible => rounding errors, which can be mitigated or eliminated with dithering or high bitdepth end to end… or making things easy to app ( for example “idealized” matrix profiles for devices that can be modeled that way)
Photoshop is working as expected. 1xTRC matrix profile + basic GPU aceleration should reduce rounding error issues to the best it can with your current HW, although matrix 1TRC profile may reduce accuracy since it may not be an extremely close fit for your display. IMHO good enough given the HW but YMMV.
so not profiling my monitor yields better gradients transitions but poorer colour accuracy, and profiling my monitor yields poorer gradients transitions but better colour accuracy. Do you mean that for this monitor, this is my tradeoff?
I don’t only use photoshop, I also use irfanview which I’d like to not show banding.
Also, if I use dithering in photoshop to improve greyscale transitioning, then this means I only treat an image based on my monitor, no? If you have a better monitor, calibrated and profiled without greyscale banding in photoshop, you don’t need to apply dithering, but you might see if I applied dithering.
So this is a limitation of my monitor and not the profiling? because as I said earlier, when it wasn’t profiled, it didn’t show banding. This leads me to believe that this monitor can be profiled and display grayscale without banding.
so not profiling my monitor yields better gradients transitions but poorer colour accuracy, and profiling my monitor yields poorer gradients transitions but better colour accuracy. Do you mean that for this monitor, this is my tradeoff?
I don’t only use photoshop, I also use irfanview which I’d like to not show banding.
Use it innacuratelly, an U2412M is close to sRGB after all.
Also, if I use dithering in photoshop to improve greyscale transitioning,
You can’t. Dithering stage must be after color transformations and before truncating to desktop composition bpc (8bit). Photoshop or Indesign or Illustration have not that feature. LR, C1 or ACR do. PS has 10bit output to an OpenGL driver. OpenGL driver is on vendor side, some take this 10bpc input and dither to 8 (macbooks + embeded display, some Firepros in the past, you can name it the good approach), some expects 10bit end to end.
then this means I only treat an image based on my monitor, no? If you have a better monitor, calibrated and profiled without greyscale banding in photoshop, you don’t need to apply dithering, but you might see if I applied dithering.
I said before you can’t, maybe with PS 10bit Open GL but all other Adobe design apps won’t. HQ monitor with HW calibration usually profile single curve + matrix profiles that minimizes it, but if calibration (non color managed) is bandless DisplayCAL can do the same. Here banding will be limited to TRC brightness diferences, “colorless”.
So this is a limitation of my monitor and not the profiling? because as I said earlier, when it wasn’t profiled, it didn’t show banding. This leads me to believe that this monitor can be profiled and display grayscale without banding.
It is a PS (+Gimp, + Infranview, +Indesign, + Illustrator, + Firefox, +Edge, ….) limitation.
LR, CaptureOne, ACR, madVR… do not have it.On my U2412M I noticed some color shifts on skin tones using a matrix profile. Since I found that unacceptable I’ve learned to live with some banding using XYZLUT. Maybe I’m anal about colors. PS will introduce banding on it’s own anyway under some conditions unrelated to the monitor/profile. With the exception of the Firepro GPUs Vincent mentioned. They’re more costly than the standard AMD GPU I use.
BTW your gamut looks similar to mine. This correction might be a good match for you: https://filebin.net/ohoxd4dmn3ceci5d Now way to verify 100% without a spectrometer(you’d make your own correction) but if it’s a good match you can evaluate any improvement by viewing images with skin tones. Also you don’t need to profile with hundred of patches, the 175 set already produces the least amount of error in the measurement report with this display.
On my U2412M I noticed some color shifts on skin tones using a matrix profile. Since I found that unacceptable I’ve learned to live with some banding using XYZLUT. Maybe I’m anal about colors. PS will introduce banding on it’s own anyway under some conditions unrelated to the monitor/profile. With the exception of the Firepro GPUs Vincent mentioned. They’re more costly than the standard AMD GPU I use.
BTW your gamut looks similar to mine. This correction might be a good match for you: https://filebin.net/ohoxd4dmn3ceci5d Now way to verify 100% without a spectrometer(you’d make your own correction) but if it’s a good match you can evaluate any improvement by viewing images with skin tones. Also you don’t need to profile with hundred of patches, the 175 set already produces the least amount of error in the measurement report with this display.
Are you using a specific image or set of images to judge skin tones? Is it a stock photo? I wonder if I could run a comparison with you and see if we arrive at the same conclusion.
As for the correction file, do I need to use it and re-run the calibration process (using 175 samples)?
Thanks for the file!
As for the correction file, do I need to use it and re-run the calibration process (using 175 samples)?
Yes.
Are you using a specific image or set of images to judge skin tones? Is it a stock photo? I wonder if I could run a comparison with you and see if we arrive at the same conclusion.
Yeah stock photos, I just pick out a selection from image searches and discard anything with obvious color shifts.
