mrw wrote: > At present, the buffer is set at 10 seconds of decoded audio, at 44,100 > Hz. That would be reduced to 1.125 seconds at 352,800 Hz. I don't know > what the underlying constraint is, but it seems 1.125 seconds just isn't > enough. I don't know what the optimum choice would be, perhaps there may > be something else in SqueezePlay that is waiting on a minimum amount of > available audio.
Well, in this case the underlying constraint is that 'high sample-rate' flac requires an output threshold of 2 seconds (20 tenths). So that kills 353/384 kHz streams, but didn't worry 176/192 kHz streams. https://github.com/Logitech/slimserver/blob/public/8.0/Slim/Player/Squeezebox.pm#L617 It seems to have been required for older Squeezeboxes and Transporter: https://github.com/Logitech/slimserver/commit/fbd203462fbe086dc3c905972a3dd756a8f60496 I wonder if it is still required for a vanilla Touch ? ralphy wrote: > What value worked for DECODE_AUDIO_BUFFER_SIZE on MacOS? > > For ALSA, I suspect what 'Triode did to handle 192Khz in EDO' > (https://github.com/ralph-irving/triode-squeezeplay/commit/52371b1b67a50ba81050c91091834d48846aa7a3) > just needs to be updated to support 384Khz rates as my squeezeplay > builds already include his change. > > The squeezeplay portaudio driver uses the same value of > DECODE_AUDIO_BUFFER_SIZE but doesn't include a similiar handler for high > samplerates like the ALSA driver. I don't know precisely what headroom is required. Each of the following two changes worked for portaudio: a) Simply doubling the size of the buffer (on the premise that if it worked for 176/192, double would work for 353/384). b) Implementing Triode's 1 second "arbitrary" override on higher bit rate rate streams. I haven't checked playback of 353/384 streams on an Alsa Squeezeplay (I don't have one available right now), my guess is it would work as is, because (b). I don't know what motivated Triode's change. Experience with EDO ? Why didn't LMS reduce the output threshold for SqueezePlay ? I can let you have a PR to adopt the Triode approach in portaudio. In the meantime, though, I'm investigating to see if a more "scientific" approach can be found. I.E. what minimum output buffer headroom is needed to guarantee that playback won't stall in this circumstance. ------------------------------------------------------------------------ mrw's Profile: http://forums.slimdevices.com/member.php?userid=38299 View this thread: http://forums.slimdevices.com/showthread.php?t=113234 _______________________________________________ Squeezecenter mailing list Squeezecenter@lists.slimdevices.com http://lists.slimdevices.com/mailman/listinfo/squeezecenter
