#40 0.5.7 Follow playhead marker not in sync with metronome


A world update pulled in 0.5.7 and I have been tracking down a strange problem. I just downgraded to 0.5.6 and the problem went away.

Note: using only MIDI, no digital audio, using external USB MIDI device, no softsynths, setup has been fairly stable for several years. the only change here was qtractor version as reported by genlop listing of emerge logs

Initial symptom was session playback metronome was inconsistent, MIDI events would slowdown after several measures, drop notes and then jump forward to catch up to metronome.

Created a new session with simple MIDI track, just single quarter notes, the symptom was playback of the single note "sounds" correct. Meaning to my ear the timing was ok. But the GUI playhead in 0.5.7 is constantly falling behind. About a half note behind after 8 measures, a whole note behind after 16. Note: Lag time is just estimated.

After downgrading, I see the following message opening every session:
07:45:16.828 The original session sample rate (48000 Hz) is not the same of the current audio engine (44100 Hz). Saving and reloading from a new session file is highly recommended.

My jackd setup is at 44100 Hz. However, I am not doing any digital audio in this current situation so am unsure this is related or not.

After downgrade all test sessions run in a constant time and playback all notes performed.


  • Rui Nuno Capela

    Rui Nuno Capela - 2013-02-17

    please, try svn trunk and build for debug (./configure --enable-debug) ...

    iit is not clear on your report whether the issue is visual or metronome timing is audibly falling behind (drifting)

    i suspect there's some relation to sample-rate mismatch--try running jackd -r48000 and check whether it makes any difference.

    also, an example session file (.qtz) for evidence is most helpful.


  • Rui Nuno Capela

    Rui Nuno Capela - 2013-02-17

    update: svn trunk rev.3205+ (aka. v0.5.7.26+)

    • Reinstated and fixed (old) warning and impending re-conversion on loading session files which the original sample-rate differs from current audio device engine (aka. JACK).

    might just have it fixed as it were on v0.5.6.

    please check, test && tell.

  • Andrew Hamlin

    Andrew Hamlin - 2013-02-17

    Wow! Fast turn around... I will try to check this out by the end of the day. Will post a reply as soon as I have feedback for you on all questions.


    Last edit: Rui Nuno Capela 2013-02-17
  • Andrew Hamlin

    Andrew Hamlin - 2013-02-17

    I confirmed two things:
    1) Upgraded back to 0.5.7 and opened my test session that 0.5.6 had converted down from 48000 to 44100. With jackd still running at 44100 Hz, 0.5.7 played just fine.
    2) Restarted jackd at 48000 Hz. Created new session with 0.5.7. Timing for recording and playback were all correct.
    I haven't confirmed the svn_trunk build yet, but wanted to get you this feedback. Later tonight I will try to build the new source if time allows.

    • Andrew Hamlin

      Andrew Hamlin - 2013-02-18

      I have built svn trunk and tested session with jackd at 44100 and 48000, conversion occurs back and forth. Also, tested 0.5.7 alternating jackd Hz and confirmed the problem existsin 0.5.7 and has been fixed on trunk. Attached is the quarter note test session.

  • Rui Nuno Capela

    Rui Nuno Capela - 2013-03-05
    • status: open --> pending
    • milestone: 0.5.7 --> svn_trunk
  • Rui Nuno Capela

    Rui Nuno Capela - 2013-05-06
    • status: pending --> closed

Log in to post a comment.