def render(play, voice_id, once, uno, bufs):
os.nice(10)
current = mp.current_process()
out = play(voice_id)
if once == True:
settings.voice(voice_id, 'once', False)
settings.buf(voice_id, value=out, bufs=bufs)
def out(generator, tick, bufs):
""" Master playback process spawned by play()
Manages render and playback processes
Params are collapsed to a key-value dict,
where the value is translated to the target
data type, and the key is expanded to the param
full name.
"""
voice_id = str(mp.current_process().name)
# Spawn a render process which will write generator output
# into the buf for this voice
r = mp.Process(name='r' + str(voice_id), target=render, args=(generator.play, voice_id, False, False, bufs))
r.start()
r.join()
def openpcm(device):
try:
out = alsaaudio.PCM(alsaaudio.PCM_PLAYBACK, alsaaudio.PCM_NORMAL, device)
except:
print 'Could not open an ALSA connection.'
return False
out.setchannels(2)
out.setrate(44100)
out.setformat(alsaaudio.PCM_FORMAT_S16_LE)
out.setperiodsize(10)
return out
# Open a connection to an ALSA PCM device
device = getattr(generator, 'device', 'default')
out = openpcm(device)
# On start of playback, check to see if we should be regenerating
# the sound. If so, spawn a new render thread.
# If there is a fresher render available, use that instead.
cooking = False # Flag set to true if a render subprocess has been spawned
volume = 1.0
next = False
buf = settings.buf(voice_id, bufs=bufs)
while settings.voice(voice_id, 'loop') == True:
regenerate = settings.voice(voice_id, 'regenerate')
once = settings.voice(voice_id, 'once')
uno = settings.voice(voice_id, 'uno')
quantize = settings.voice(voice_id, 'quantize')
target_volume = settings.voice(voice_id, 'target_volume')
if uno == True:
settings.voice(voice_id, 'loop', False)
if regenerate == True or once == True:
reload(generator)
device = getattr(generator, 'device', 'default')
out = openpcm(device)
r = mp.Process(name='r' + str(voice_id), target=render, args=(generator.play, voice_id, False, False, bufs))
try:
r.start()
r.join()
buf = settings.buf(voice_id, bufs=bufs)
except OSError:
dsp.log('failed to regenerate voice %s' % voice_id)
if quantize != False:
tick.wait()
dsp_loop(out, buf, voice_id)
settings.remove_voice(voice_id)