Kuro wrote: 
> On another occasion, I did see on the debug log that squeezlite tried to
> reopen the device with plughw, but on my setup, I saw something like
> "invalid parameters" and it fails to reopen using plughw.  This is with
> the Pogoplug.  With Cubox, it crashes with segmentation fault.
> 
> Not sure if it has to do with me forcing it to 16/44.1 (-r 44100).  Some
> how when it tried to reopen with plughw, it failed with invalid
> parameters.  FWIW, when I do a "-l" with squeezelite on the Pogoplug, I
> did not see plughw being listed.  I see the default device (which is
> plughw I think), plus all those surround devices and iec958, which is
> the same as hw.  So maybe if reopening plughw fails, squeezelite should
> try again with default:CARD instead for the CARD with the same name as
> the current connection.
> 
> I have libsoxr installed.  For 24/96 stream, the Pogoplug does not have
> enough juice to run the realtime downsampling to 16/44.1.  But with low
> bit rate stream like Internet radio, I think it should be fast enough
> with sox.  So maybe an input parameter on squeezelite not to reopen
> plughw for internet radio, but force the server to do the resampling
> with sox?
> 
> But obviously, it is best to run with plughw so CPU load is low.

Server based resampling can't be used by default for this issue as it
only does downsampling to allow high res formats to be played on low end
players.  There are plugins to allow upsampling on the server but that's
not the default operation.

Re the crash - some people have seen crashes in libmad on some arm
platforms.  Could you try with libmpg123 to rule this out.  Otherwise
can you run it inside gdb and post what the backtrace is at the crash?


------------------------------------------------------------------------
Triode's Profile: http://forums.slimdevices.com/member.php?userid=17
View this thread: http://forums.slimdevices.com/showthread.php?t=97046

_______________________________________________
unix mailing list
[email protected]
http://lists.slimdevices.com/mailman/listinfo/unix

Reply via email to