On Wed, 2013-07-10 at 11:28 +0300, Tanu Kaskinen wrote: > On Wed, 2013-07-10 at 11:12 +0300, Tanu Kaskinen wrote: > > On Wed, 2013-07-10 at 16:18 +1000, Luke Yelavich wrote: > > > Hey folks. > > > I've been experiencing some weird and infrequent crackling of audio > > > with pulse 4, and applications that use the simple pulse API. The > > > crackling tends to happen at the very start of the buffer, causing a > > > varying amount of audio to be audible as noise, or not at all. This > > > doesn't occur when other applications are playing audio at the same > > > time, and it comes and goes, and I can never predict when it might > > > start to happen, or when it will go away. Unfortunately I haven't been > > > able to reliably log this, since it seems to occur very infrequently > > > if not at all whilst being logged with full verbose output. > > > > If it doesn't happen when other applications are playing audio, it > > probably only happens when the device suspended prior to > > speech-dispatches starting to play it's audio. Reducing the suspend > > timeout from the default 5 seconds to 1 or 0 seconds may help with > > reproducing the issue more frequently. Setting the timeout to 0 seconds > > can be done by giving argument "timeout=0" to module-suspend-on-idle. > > I now tried speech-dispatcher myself, and at least on my machine > speech-dispatcher creates a persistent stream, so the device never > suspends. My setup is Ubuntu 13.04 without any changes to > speech-dispatcher configuration. I started speech-dispatcher without any > command line arguments, and I tested "spd-say hello", which seems to > result in sound remotely resembling the word "hello", but the audio is > terribly crackly (I presume I'm seeing the same bug as you). The audio > is never clean, it doesn't matter whether something else is playing at > the same time or not. > > My PulseAudio version is the current master branch head, I'll try next > with PulseAudio 3.0. PulseAudio 3.0 seems to work fine. I guess it's time for bisecting... -- Tanu
