transport : bpm changes and jackd: first beat played double
When using jackd and the bpm change table: Once the song ran over a bpm change and I reset the song position to start, the first beat is played double and the hydrogen timeline is out of sync with the jack timeline.
Leave a comment
file:cqijZcSSKr4kr8eJe4gwI3: patch for proposed fix.
retested and could not reproduce this, however the bug description does not say if h2 is jack master or if another app is also working in jacktrans mode or if it's just qjackctl that is used to control the transport ..
so need more info here
I did find something interesting :
- create a song that has 4 patterns in it (make sure you have a kick on 'the 1')
- on bar 1 place a bmp marker of 120bpm
- on bar 3 place a marker of 180bmp
play the song > all OK
now change the bmp marker on bar 3 to 60bpm and play the song again
> you can clearly hear that the first note of bar 3 (where the tempo changes to 60) is played double
(the velocity is higher)
sounds like this can be related to the initial bug report
so need more info here
I did find something interesting :
- create a song that has 4 patterns in it (make sure you have a kick on 'the 1')
- on bar 1 place a bmp marker of 120bpm
- on bar 3 place a marker of 180bmp
play the song > all OK
now change the bmp marker on bar 3 to 60bpm and play the song again
> you can clearly hear that the first note of bar 3 (where the tempo changes to 60) is played double
(the velocity is higher)
sounds like this can be related to the initial bug report
testing some more
- create a pattern (kick on the 1 and 3, snare on the 2 and 4)
- make a song that contains this pattern 2 times (so the song only has 2 patterns)
- add a tempo marker on bar 1 (120bpm)
- add a tempo marker on bar 2 (100bpm)
remark:
some interesting things :
1) the double played notes dont seem to happen all the time, but the chanse if it happening increases if the difference between the 2 bpm markers increases (25% for 120bpm > 110bmp , 100% for 120bpm > 50bpm)
2) when a 'double note' is played you can hear it, but it really becomes clear if you look at the mixer VU meters
BUT: you can only see the master VU meter clip, the instrument (kick) VU meter shows the same level for 'good' kicks and for 'clipped' kicks (???)
it almost looks as if the sound is added to the master twice (POST instrument fader)
2) the clipped sounds happen at 2 places in the song : first beat of bar 1, or first beat of bar 2
if the tempo* drops* (eg 120bpm > 50bpm) the clipping will be on the first beat of bar 2
if the tempo_ goes up_ (eg 120bpm > 180bpm) the clipping in on beat1 of bar1
sounds like some kind of race condition
- create a pattern (kick on the 1 and 3, snare on the 2 and 4)
- make a song that contains this pattern 2 times (so the song only has 2 patterns)
- add a tempo marker on bar 1 (120bpm)
- add a tempo marker on bar 2 (100bpm)
remark:
some interesting things :
1) the double played notes dont seem to happen all the time, but the chanse if it happening increases if the difference between the 2 bpm markers increases (25% for 120bpm > 110bmp , 100% for 120bpm > 50bpm)
2) when a 'double note' is played you can hear it, but it really becomes clear if you look at the mixer VU meters
BUT: you can only see the master VU meter clip, the instrument (kick) VU meter shows the same level for 'good' kicks and for 'clipped' kicks (???)
it almost looks as if the sound is added to the master twice (POST instrument fader)
2) the clipped sounds happen at 2 places in the song : first beat of bar 1, or first beat of bar 2
if the tempo* drops* (eg 120bpm > 50bpm) the clipping will be on the first beat of bar 2
if the tempo_ goes up_ (eg 120bpm > 180bpm) the clipping in on beat1 of bar1
sounds like some kind of race condition
file:aQfcmAFPSr4ynFacwqjQXA
screencast of the comment added on april 4th
screencast of the comment added on april 4th