After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 508029 - [speexresample] pops/glitches in some cases
[speexresample] pops/glitches in some cases
Status: RESOLVED FIXED
Product: GStreamer
Classification: Platform
Component: gst-plugins-bad
git master
Other Linux
: Normal major
: 0.10.10
Assigned To: Sebastian Dröge (slomo)
GStreamer Maintainers
Depends on:
Blocks:
 
 
Reported: 2008-01-08 11:40 UTC by Edward Hervey
Modified: 2008-10-29 16:40 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
Screenshot of resampled data and differences in Audacity (35.26 KB, image/png)
2008-01-08 11:42 UTC, Edward Hervey
Details

Description Edward Hervey 2008-01-08 11:40:15 UTC
There are some cases where speexresample introduces pops and glitches when resampling.

One example is the following pipeline:

gst-launch-0.10 audiotestsrc freq=220 samplesperbuffer=254 num-buffers=1000 ! audio/x-raw-int,rate=8000 ! speexresample  ! audio/x-raw-int,rate=11025 ! alsasink
Comment 1 Edward Hervey 2008-01-08 11:42:44 UTC
Created attachment 102387 [details]
Screenshot of resampled data and differences in Audacity

This is a screenshot of audacity showing:
* The difference between a 11025Hz rate 220Hz sine wav (audiotestsrc freq=220 ! audio/x-raw-int,rate=11025) and the pipeline specified in the comment above
* The output of the above pipeline

As you can see, it looks as though there's maybe an issue with the timestamp of the outgoing buffer, producing a pop and shift of the output in time.
Comment 2 Sebastian Dröge (slomo) 2008-01-23 11:16:52 UTC
gst-launch-0.10 -v audiotestsrc num-buffers=500 freq=220 samplesperbuffer=253 ! audio/x-raw-float,rate=16000 ! speexresample ! audio/x-raw-float,rate=8000 ! audioconvert ! wavenc ! filesink location=test.wav

This gives even worse results. There are always 64 or 65 samples containing something outside of [-1,1].
Comment 3 Jan Schmidt 2008-01-23 17:36:36 UTC
How is this a blocker?
Comment 4 Sebastian Dröge (slomo) 2008-01-24 09:45:41 UTC
Of course it isn't, should've been major instead. Sorry
Comment 5 Sebastian Dröge (slomo) 2008-10-28 19:29:18 UTC
I can't reproduce it anymore with latest CVS... so it's probably fixed by somethnig now ;) Edward?
Comment 6 Edward Hervey 2008-10-29 12:15:14 UTC
all good for me.