[gst-devel] oss audio pipeline doesn't work when debug on

[David Schleef]

On Mon, Jun 09, 2008 at 03:35:13PM -0500, Liu, Bin wrote:
> Hi,
> 
> I am new to gstreamer. I cross-compiled gst to my ARM board. I have no
> problem to play the mp3 audio stream from a movie clip using the
> following pipeline:
> 
> 	gst-launch -v --gst-debug-level=2 \
>       filesrc location=t.avi ! avidemux ! queue ! mad ! osssink
> 
> But if I turn on the debug level higher than level 2 for any component,
> I cannot hear anything from the speakers, for example using the
> following pipeline:

This is not surprising.  Audio is automatically dropped when the
buffers arrive at the sink late.  Debugging produces a lot of output,
and if the output device is slow or has a small buffer, it will
likely fill up and cause all GStreamer processing to stop.  On
a desktop system, output to an xterm is both fast and has a large
buffer, so you're unlikely to ever see the problem.  The best way to
solve this on an embedded system is using ssh to log into the system
and/or pipe the debug output out to a file.

It is also possible to write a custom debug handler that manages the
output in a specific way.  For one client, I created a custom debug
handler that emulated a simple web server, and connecting to this
web server gave a constant stream of gstreamer debug output in a
web browser.



dave...