GNOME Bugzilla – Bug 169278
only part of the audio stream played back
Last modified: 2006-09-08 14:52:27 UTC
http://w2.tvtorrents.tv:8081/bt/alias.s04e09.hr.hdtv.ac3.5.1.xvid-ctu.torrent That file contains AC3 5.1 audio. Now, when i compile gst-plugins with a52dec support, i hear sound, but only the "music track" (without the a52dec plugin, i get no audio at all). I don't hear any voices. I've got gst-ffmpeg also installed (video plays fine). Using totem btw. This works fine with totem+xine.
Starting totem in debug mode and playing the above file: [jeroen@orcrist ~]$ totem --gst-debug=audioconvert:5 2>&1 | grep gstchannelmix.cDEBUG (0x9f77898 - 308337:46:44.753800000) audioconvert(19160) gstchannelmix.c(516):gst_audio_convert_setup_matrix: Matrix for 6 -> 6: { { 0.000000, 0.000000, 0.000000, 0.000000, 0.000000, 1.000000 }, { 1.000000, 0.000000, 0.000000, 0.000000, 0.000000, 0.000000 }, { 0.000000, 0.000000, 0.000000, 0.000000, 1.000000, 0.000000 }, { 0.000000, 1.000000, 0.000000, 0.000000, 0.000000, 0.000000 }, { 0.000000, 0.000000, 1.000000, 0.000000, 0.000000, 0.000000 }, { 0.000000, 0.000000, 0.000000, 1.000000, 0.000000, 0.000000 } } After seeking ahead and back to the start of the video, i get this: DEBUG (0x9c9d0d0 - 308337:47:33.458389000) audioconvert(19160) gstchannelmix.c(516):gst_audio_convert_setup_matrix: Matrix for 6 -> 2: { { -0.292893, 0.292893 }, { 0.292893, 0.000000 }, { -0.207107, 0.207107 }, { 0.000000, 0.292893 }, { -0.103553, 0.103553 }, { -0.103553, 0.103553 } } It now plays the audio (including voices) back properly. Audio Output is configured to "Stereo" in totem. From IRC: BBB (well, apparently alsa drops your audio, not us... alsa should simply not claim to support 6 channels...) BBB the fact that alsa claims to support 6 channels and then downsamples by dropping surround BBB a partial solution is to set your audiosink to "alsasink device=hw:0"
After seeking when the audio streams are properly merged, the volume is extremely low compared to other apps on my desktop (Muine for example). I have to turn the volume to MAX to be able to hear voices properly.
This is inside Totem btw, since applications will have to figure out a way of getting the amount of speakers attached and thereby configure output channels.
Jeroen, could you apply the patch from 326720 and check if this is resolved now with 0.10?
This should be fixed in -ugly CVS (as long as the audiosink either detects the number of channels right or is forced to 2 channels). *** This bug has been marked as a duplicate of 350761 ***