Ricoh GR IV noise profiles

[andriiryzhkov]

Add noise profiles for Ricoh GR IV (ISO 100–204800, 34 profiles)

Tested locally on multiple files - looks good to me. In-camera noise reduction was turn-off.

noise_result.pdf

[kmilos]

Thanks! Any chance you can split this into separate pull requests?

[andriiryzhkov]

@kmilos done.

Noise profiles in this PR. White balance presets on #20361

[TurboGit]

The noise profile doesn't seems good to me. What was the actual images taken to create it?

[andriiryzhkov]

This is archive with processing result darktable-noiseprofile-20260216.tar.gz
I can also share RAW images if needed.

[TurboGit]

On the "standard deviation" graph the green line should be close to the brown line (mean = 1). As you can see this is not true and the deviation from the mean is already quite important at 300 ISO.

My bet is that you don't have enough transition on the bright side of your picture. You really want full black on the left, and a smooth transition from middle tones to the highlight.

Please try again with a better input. TIA.

[andriiryzhkov]

@TurboGit : Updated.
ricoh-gr-iv-noiseprofile-20260322.tar.gz
ricoh_gr_iv_noise_result.pdf

[TurboGit]

Still not good to me, see result for ISO 800:

This is what we should have (from a recently validated noise profile):

Nothing above mean (yellow line) on the right (past brightness 0.3).

[andriiryzhkov]

I would compare to the same camera type, like Ricoh GR IIIx #13953. It looks similar to me.

Anyway, I need my best according to the instructions - overexposed white, smooth gradient to complete dark. Though image was not too much blurry, but that was the best I can get on that fixed lens camera.

If that does not work, I have no other ideas what can be improved. The only thing I can do is to close this PR.

[TurboGit]

I would compare to the same camera type, like Ricoh GR IIIx #13953. It looks similar to me.

You're right, maybe inherent to Ricoh sensors, strange. I'll merge this then. Thanks.