Infinite contrast with ICC profile without black point data

[yZibi]

mpv Information

mpv v0.41.0 Copyright © 2000-2025 mpv/MPlayer/mplayer2 projects
 built on Dec 25 2025 15:27:46
libplacebo version: v7.358.0 (v7.351.0-106-gbc90ef9-dirty)
FFmpeg version: N-122267-gabb152413
FFmpeg library versions:
   libavcodec      62.23.102
   libavdevice     62.2.100
   libavfilter     11.10.101
   libavformat     62.8.101
   libavutil       60.20.100
   libswresample   6.2.100
   libswscale      9.3.100

Other Information

- Windows version: 11 24h2 26100.4061
- GPU model, driver and version: GTX1070, 582.28
- Monitor: Dell h2421hs [SDR]
- Source of mpv: winget
- vo=gpu-next

Reproduction Steps

Enabling ICC profile makes near-black regions noticeably darker than without it.

Tested with Dell factory ICC profile, and with the standard Windows "sRGB IEC61966-2.1" ICC profile.

The default (non-ICC) rendering closely matches madVR configured as:

• This display is already calibrated
• Pure power curve 2.20

The issue is identical with Vulkan and D3D11.

The following options made no visible difference:
--target-trc=gamma2.2
--icc-use-luma=no
--icc-force-contrast=1000

logfile says:
[v][vo/gpu-next/libplacebo] Opened ICC profile:
Raw black point: X=0.000000% Y=0.000000% Z=-0.000000%
Contrast = 203 cd/m^2 : 0.001 mcd/m^2 ≈ 203000000 : 1
Measured primaries:
White: x=0.312696, y=0.329020
Red: x=0.649, y=0.336
Green: x=0.310, y=0.618
Blue: x=0.146, y=0.055
Containing primaries: ITU-R Rec. BT.2020 (Ultra HD)
Transfer function: IEC 61966-2-4 sRGB (CRT emulation)

Even when using:

--icc-force-contrast=1000

the log still reports the same contrast value.

Q: Is the reported black point of zero and the resulting ~203,000,000:1 contrast expected for a normal SDR ICC profile?
P.S. --vo=gpu - no image, just grey screen. --vo=direct3d has no issue (maybe cuz it doesn't load profile?)

Expected Behavior

mpv --no-config

Result:

• correct black level
• visually matches madVR

Actual Behavior

mpv --no-config --icc-profile-auto=yes
or

mpv --no-config --icc-profile="C:\Windows\System32\spool\drivers\color\sRGB Color Space Profile.icm"

Result:

• darker near-blacks / shadow detail
• visible difference from baseline

Log File

output.txt

I carefully read all instruction and confirm that I did the following:

• I tested and confirmed that the issue exists with the latest release version or newer.
• I provided all required information including system and mpv version.
• I produced the log file with the exact same set of files, parameters, and conditions used in "Reproduction Steps", with the addition of --log-file=output.txt.
• I produced the log file while the behaviors described in "Actual Behavior" were actively observed.
• I attached the full, untruncated log file.

[kasper93]

A) Why do you use ICC profile?
B) Use --target-contrast.

[yZibi]

A) Why do you use ICC profile?
B) Use --target-contrast.

Why not?
Same with --no-config --icc-profile-auto=yes

[kasper93]

A) Why do you use ICC profile?
B) Use --target-contrast.

Why not? Same with --no-config --icc-profile-auto=yes

Raw black point: X=0.000000% Y=0.000000% Z=-0.000000%

Because your profile reports infinite contrast and you don't want infinite contrast, therefore why use profile which is not valid for your use case? And why did you ignore my note about --target-contrast?

[yZibi]

A) Why do you use ICC profile?
B) Use --target-contrast.

Why not? Same with --no-config --icc-profile-auto=yes

Raw black point: X=0.000000% Y=0.000000% Z=-0.000000%

Because your profile reports infinite contrast and you don't want infinite contrast, therefore why use profile which is not valid for your use case? And why did you ignore my note about --target-contrast?

Not only my profile, but i think every generic profile has no black point data.
i repeat: --no-config --icc-profile-auto=yes --target-contrast=1000 -> same isssue.