On Mon, 2007-02-05 at 12:17 -1000, david wrote: > Paul Davis wrote: > > On Mon, 2007-02-05 at 16:41 -0500, carmen wrote: > >> On Mon Feb 05, 2007 at 01:59:29PM -0600, Victor Roetman wrote: > >>> (and I really wish there was a jack backend for gstreamer) > >> there is. it didnt work very well for me (worse than alsa+plug:jack, just like XINE) but you might want to give it a shot > > > > carmen, judging from the rest of your post, CPU usage was the main > > issue. > > > > this is actually not a consequence of the JACK sink directly, but the > > appalling unoptimized resampling code in gstreamer. if you connect the > > JACK sink up to a stream that is already at the JACK sample rate, i have > > found you see the expected (low) CPU utilization. > > > > the gstreamer guys know about this, but there is nobody available or > > with the know-how to speed it up. libsamplerate is not an option, > > apparently, because of the license :(( > > I've been warned by a number of Linux folk to avoid GStreamer. i think thats a mistake. gstreamer has quite a complex API and a very complex internal design, but it solves a number of problems that nobody else has bothered to solve.
