Hallo all, i currently play with speex settings. My conclusion is: speexenc is sort of bad with default options.
Here is my reconfigured pipeline: gst-launch pulsesrc device=0 ! audio/x-raw-int, rate=16000, channels=1 ! speexenc vbr=1 dtx=1 nframes=4 ! rtpspeexpay ! udpsink port=5555 host=192.168.0.155 In my tests, if i use speexenc without any option, network usage is 50 kbit/s. If i use "speexenc vbr=1 dtx=1 nframes=4", network usage is about 25 kbit/s without noticeable quality regression. Options description: vbr=1 # enable variable bitrate dtx=1 # do not send any packets on silence # see Bug https://bugzilla.gnome.org/show_bug.cgi?id=646474 nframes=4 # allow better compression and bigger packet size. It reduce network # traffic to 13 packets/sec. (with default options ~50 packets/sec). # the value bigger than 4 will introduce stuttering. ==================================================================== Pipes for testing: source: gst-launch pulsesrc device=0 ! audio/x-raw-int, rate=16000, channels=1 ! speexenc vbr=1 dtx=1 nframes=4 ! rtpspeexpay ! udpsink port=5555 host=192.168.0.155 sink: gst-launch udpsrc port=5555 caps="application/x-rtp, media=(string)audio, clock-rate=(int)16000, encoding-name=(string)SPEEX, encoding-params=(string)1, payload=(int)110" ! gstrtpjitterbuffer ! rtpspeexdepay ! speexdec ! pulsesink ==================================================================== Now the question: how can i change speexenc settings for empathy? Regards, Alexey _______________________________________________ telepathy mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/telepathy
