GNOME Bugzilla – Bug 506557
Convertion from wma to ogg introduce hatches in the sound
Last modified: 2011-05-19 06:14:51 UTC
Please describe the problem: Direct convertion from wma to ogg introduce hatches in the sound. Pipelines used are bellow. Steps to reproduce: 1. Chose a wma file 2. $ gst-launch-0.10 filesrc location=<file.wma> ! decodebin ! audioconvert ! vorbisenc ! oggmux ! decodebin ! audioconvert ! alsasink Actual results: Sound is hatch Expected results: Sound should not be hatch Does this happen every time? yes Other information: $ gst-launch-0.10 filesrc location=<file.wma> ! decodebin ! audioconvert ! wavenc ! decodebin ! audioconvert ! vorbisenc ! oggmux ! decodebin ! audioconvert ! alsasink => With one stage of decoding/encoding (in wav), there is not trouble.
- does it work fine if you use alsasink device=front instead? - does it work fine if you use audioconvert ! audiorate ! vorbisenc ?
alsasink device=front => No, it doesn't work. Error is : gstalsasink.c(622): gst_alsasink_open (): /pipeline0/alsasink0: Device 'front' is busy audioconvert ! audiorate ! vorbisenc => The sound is largely better with audiorate, but it still contain hatch which aren't present using wavenc
I can reproduce this (with pitfdll being the decoder). The timestamps are messed up, vorbisenc reacts by dropping buffers.
That would be either a corrupted file then or bug #549940. I'll close this as a duplicate of bug #549940. *** This bug has been marked as a duplicate of bug 549940 ***