------- Additional Comments From richard@... 2007-12-31 06:07 -------
Please explain the last comment further - it makes no sense to me at all. I
wouldn't expect the number of samples exported to be the same as the number in
the project, because of the resampling. I would expect the duration (number of
samples x rate) of the exported file to be the same as the source selection.
* I create a 12 second section of tone, with a sample rate of 48000Hz.
* With snap-to on I select from 3 seconds 0 CDDA frames to 7 seconds 0 frames,
and create a label. The length of this selection is 192000 samples, or exactly 4
* I change the project rate to 44100Hz. The selection duration is still 4
seconds, but represents 176400 samples at 44100Hz
* I run export multiple, to a WAV file using labels
* I get a file which contains 175,904 samples when re-imported into audacity
(also confirmed by sndfile-info).
Whilst incorrect, this is much less than the errors that were exhibited
previously (when the exported file would have had 192000 samples at 44100 Hz,
and lasted 4.35 seconds). A new bug should be opened for this, as the cause is
almost certainly different, and a different test case is needed to detect it (a
length error of 96 samples or 0.002 of a second is not normally audible).
Note that testing this is only meaningful with audio formats that support
sample-length accurate encoding. To my knowlege these are Uncompressed PCM, FLAC
and Ogg Vorbis (all of which exhibit the problem currently). MPEG audio files
will always be padded to a fixed frame size during encoding.
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.