From: GStreamer (bugzilla.gnome.org) <bug...@bu...> - 2006-10-30 17:49:10
|
Do not reply to this via email (we are currently unable to handle email responses and they get discarded). You can add comments to this bug at http://bugzilla.gnome.org/show_bug.cgi?id=3D360673 GStreamer | gst-plugins-base | Ver: HEAD CVS ------- Comment #6 from Brian Cameron 2006-10-30 17:48 UTC ------- Created an attachment (id=3D75678) --> (http://bugzilla.gnome.org/attachment.cgi?id=3D75678&action=3Dview) gst.out1 Padriag O'Briain is looking into this bug, and here are some of his findi= ngs. =20 I experimented with two sets of values (882x20) and (882x600). The first= set of values causes stutter and the second set does not. I discovered that, in both cases, the time between successive calls to gst_audo_write is normally 0.00003 seconds but occasionally, about every = 300 times, the time is about 2.6 seconds. It looks to me that this time is being spent in the write call in gst_audio_write. My evidence for this is in the attachments. gst.out1 is for (882,20) and gst.out2 is for (882,600) and the output sug= gests that every so often a write call takes a long time to return. I had a look at prstat output while gst-launch was running and the only t= hread that appears there is thread 4 which consumes about 1% of CPU. Is the audio device fd is blovking or non-blocking mode? --=20 Configure bugmail: http://bugzilla.gnome.org/userprefs.cgi?tab=3Demail |