View Issue Details

IDCategoryLast Update
0004834bugs2013-04-06 23:28
ReportercolinfAssigned To 
Reproducibilityalways 
Status confirmedResolutionopen 
Product Version 
Fixed in Version 
Summary0004834: Regions recorded via alsa_in have slightly different lengths to those recorded in the same pass from hardware inputs
DescriptionI ran out of channels on my 8-channel sound card, so I was using a USB interface to get a couple extra via alsa_in. It works nicely enough, but regions recorded simultaneously from the main sound card and via alsa_in have slightly different lengths. The alsa_in regions seem to be 122 samples longer.

This of course messes up edit group operations.
Additional InformationI'm using jack 1, the last revision from the svn repository before it was moved to git, if that has any bearing on things.
TagsNo tags attached.

Activities

colinf

2012-07-07 12:20

updater   ~0013802

Last edited: 2012-07-07 18:10

I haven't measured it properly yet, but it seems too as if the regions are slightly late (maybe by the same amount they're too long?).

[EDIT] Actually, it seems to my ears that nudging them earlier by the capture offset brings them in time. I'll try to make some measurements when I'm not in the middle of a session...

colinf

2013-04-05 12:54

updater   ~0014812

A further observation:

I originally spotted this with jackd running on a UA-1000 USB2 sound card, and alsa_in capturing from a Zoom H4. I've now seen the same behaviour on a different system, with jackd running on a Delta 1010 and Fons Adriaensen's zita-a2j capturing from a Delta 44, so I reckon it's definitely an ardour bug.

paul

2013-04-06 14:06

administrator   ~0014831

probably caused by latency reporting being different for each signal path.

nettings

2013-04-06 23:28

manager   ~0014837

to narrow it down: are the extra 122 samples happening regardless of the duration of the recording, or are there more samples the longer you have been recording? the latter might indicate that the resampling on alsa_in is not doing its job properly... but then it should go poof once the delta reaches the jack period size.

Issue History

Date Modified Username Field Change
2012-04-09 21:53 colinf New Issue
2012-04-09 23:43 cth103 cost => 0.00
2012-04-09 23:43 cth103 Target Version => 3.0 beta4
2012-05-23 15:08 cth103 Target Version 3.0 beta4 => 3.0
2012-06-26 00:56 cth103 Status new => confirmed
2012-07-07 12:20 colinf Note Added: 0013802
2012-07-07 18:10 colinf Note Edited: 0013802
2013-04-05 12:54 colinf Note Added: 0014812
2013-04-06 14:06 paul Note Added: 0014831
2013-04-06 23:28 nettings Note Added: 0014837