waveform and transport out of sync for mp3

[Daniel Schürmann]

Waveform and transport are out of sync.

My test track has according to Audacity 0,4 s silence at the beginning.

This is displayed in Mixxx correctly, the waveform starts, when the track time is on 0,40.

But the sound output starts actually starts at 0,36 and the first beat marker of the non const beat grid is at 0,36

[Daniel Schürmann (daschuer)]

The original issue appears using a mp3 file. If I save the same file as wave using Audacity, it sound output aligns perfectly with the waveform.
This was tested with 64 bit Linux build of 1.12.0-alpha (build master r5041) on Ubuntu Trusty.
the original track is an an mp3float.

I have also recoded the file in 192 kbps variable and fixed rate.
Now the Audio starts at 0:45 according to the waveforms im Mixxx and according to Audacity, but the sound starts later at 0.47 for both tracks in Mixxx.

Is that regression?

[Daniel Schürmann (daschuer)]

> Is that regression?
Tested with 1.11 release and it suffers the same issue.

[Daniel Schürmann (daschuer)]

http://hirntier.blogspot.de/2010/08/getting-serious-with-sample-accuracy.html
The test tool source can be found here:
https://github.com/Distrotech/gmerlin/blob/distrotech-gmerlin/gmerlin_avdecoder/tests/seektest.c

This seams to be interesting as well:
http://www.hydrogenaud.io/forums/index.php?showtopic=85125

[RJ Skerry-Ryan (rryan)]

Have you discovered the root cause? What's your planned fix?

[Daniel Schürmann (daschuer)]

https://github.com/mixxxdj/mixxx/pull/461

On cause are simple bugs in seek calculations.
An other cause are mad_frame_decode calls that to not advance to the next frame.

[Daniel Schürmann (daschuer)]

The fix is now in Master. Not in 1.12 alpha :-(

[Be (be.ing)]

Fixed by Uwe: https://github.com/mixxxdj/mixxx/pull/411

[Swiftb0y (swiftb0y)]

Mixxx now uses GitHub for bug tracking. This bug has been migrated to:
https://github.com/mixxxdj/mixxx/issues/7794