SoundSourceProxyTest.firstSoundTest fails with fdk_aac and FFmpeg 5.1
Author: daschuerCreated Sep 14, 2026Updated Sep 20, 2026
Labelsblockersoundsourcestems
Bug Description
This has been discovered here https://github.com/mixxxdj/mixxx/pull/16875
FFmpeg 4.4.1 works good it starts failing with FFmpeg 5.1 and commit n5.1-dev-348-gc69b1a12bb https://github.com/FFmpeg/FFmpeg/commit/c69b1a12bb
2026-09-09T04:30:34.6819269Z info [Qt mainThread] Opened file "/home/runner/work/mixxx/mixxx/src/test/id3-test-data/cover-test-øé~ł€˚-itunes-12.3.0-aac.m4a" using provider "FFmpeg 7.1"
2026-09-09T04:30:34.6820319Z warning [Qt mainThread] ReadAheadFrameBuffer - Missing range [0 -> 661) between output buffer [0 -> 2000) and input buffer [661 -> 960)
2026-09-09T04:30:34.6820988Z ./src/test/soundproxy_test.cpp:876: Failure
2026-09-09T04:30:34.6821308Z Expected equality of these values:
2026-09-09T04:30:34.6821574Z firstSoundSample
2026-09-09T04:30:34.6821786Z Which is: 2488
2026-09-09T04:30:34.6822002Z ref.firstSoundSample
2026-09-09T04:30:34.6822224Z Which is: 1166
2026-09-09T04:30:34.6823023Z /home/runner/work/mixxx/mixxx/src/test/id3-test-data/cover-test-øé~ł€˚-itunes-12.3.0-aac.m4a FFmpeg 7.1Unfortunately we us use this setup for our STEM decoder as well. That's the reason this is a blocker bug for the 2.6.0 release. User will start to create metadata that has an offset.
Version
2.6
OS
No response
Source: mixxxdj/mixxx