Skip to content

RGBA blending (software fallback?) seems to produce wrong results sometimes #17188

@afishhh

Description

@afishhh

mpv Information

mpv v0.41.0-UNKNOWN Copyright © 2000-2025 mpv/MPlayer/mplayer2 projects
 built on Jan  1 1980 00:00:00
libplacebo version: v7.351.0
FFmpeg version: 7.1.1
FFmpeg library versions:
   libavcodec      61.19.101
   libavdevice     61.3.100
   libavfilter     10.4.100
   libavformat     61.7.100
   libavutil       59.39.100
   libswresample   5.3.100
   libswscale      8.3.100

Other Information

- Linux version: NixOS 25.11 (Xantusia)
- Kernel Version: 6.17.7
- GPU Model: NVIDIA Corporation TU104 [GeForce RTX 2060]
- Mesa/GPU Driver Version: 580.95.05
- Window Manager and Version: dwm 6.3
- Source of mpv: git
- Latest known working version: not known, did not check
- Issue started after the following happened: I ran into it

Reproduction Steps

RGBA blending software fallback seems to be kind of broken?
For some reason this issue did not show up when I tried to capture it via screenshot window but it is there and can be captured with an external screenshot.
It shows up when using vo=x11 or vo=kitty though so I assume it's some software fallback that's broken here.

Steps:

  1. Open any video with vo=x11 or some other vo which won't use gpu blending
  2. Load attached test.ass file or any other ass file really
  3. Open console
  4. set force-rgba-osd-rendering yes
  5. Observe weirdness around places where there is overlapping text

Expected Behavior

Image

Actual Behavior

Image

Bonus demonic version, with #17187 in https://www.youtube.com/watch?v=ksdvNgqOToQ:
Image

Log File

Plugins still in log because nix wrapper adds them and it's too late at night for me to bother for a clearly non-log-useful issue.

Also I kind of lie in I produced the log file while the behaviors described in "Actual Behavior" were actively observed.* because I used a different video to make smaller log file without dav1d errors and dumping of video metadata.

log.txt

Sample Files

test.ass.txt

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

  • I tested with the latest mpv version to validate that the issue is not already fixed.
  • 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.
  • I attached the backtrace in the case of a crash.

Metadata

Metadata

Assignees

No one assigned

    Type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions