def render(self, chunk_size=None): gain = getattr(self.track, 'gain', None) if chunk_size is None: # self has start and duration, so it is a valid index into track. output = self.track[self].data # Normalize volume if necessary if gain is not None: # limit expects a float32 vector output = limit(multiply(output, float32(gain))) yield output else: if isinstance(self.start, float): start = int(self.start * 44100) end = int((self.start + self.duration) * 44100) else: start, end = self.start, self.end for i in xrange(start, end, chunk_size): if gain is not None: yield limit(multiply( self.track[i:min(end, i + chunk_size)].data, float32(gain) )).astype(numpy.int16) else: yield self.track[i:min(end, i + chunk_size)].data
def render(self, chunk_size=None):
gain = getattr(self.track, 'gain', None)
if chunk_size is None:
# self has start and duration, so it is a valid index into track.
output = self.track[self].data
# Normalize volume if necessary
if gain is not None:
# limit expects a float32 vector
output = limit(multiply(output, float32(gain)))
yield output
else:
if isinstance(self.start, float):
start = int(self.start * 44100)
end = int((self.start + self.duration) * 44100)
else:
start, end = self.start, self.end
for i in xrange(start, end, chunk_size):
if gain is not None:
yield limit(multiply(
self.track[i:min(end, i + chunk_size)].data,
float32(gain)
)).astype(numpy.int16)
else:
yield self.track[i:min(end, i + chunk_size)].data
def g(self, d, gain, rate): s = 44100 if gain is not None: return limit(multiply(dirac.timeScale(d, rate, s, self.quality), float32(gain))) else: return dirac.timeScale(d, rate, s, self.quality)
def g(self, d, gain, rate):
s = 44100
if gain is not None:
return limit(multiply(dirac.timeScale(d, rate, s, self.quality),
float32(gain)))
else:
return dirac.timeScale(d, rate, s, self.quality)
def render(self):
# self has start and duration, so it is a valid index into track.
output = self.track[self]
# Normalize volume if necessary
gain = getattr(self.track, 'gain', None)
if gain != None:
# limit expects a float32 vector
output.data = limit(multiply(output.data, float32(gain)))
return output
def render(self):
# self has start and duration, so it is a valid index into track.
output = self.track[self]
# Normalize volume if necessary
gain = getattr(self.track, 'gain', None)
if gain != None:
# limit expects a float32 vector
output.data = limit(multiply(output.data, float32(gain)))
return output
def stretch(self, t, l):
"""t is a track, l is a list"""
signal_start = int(l[0][0] * t.sampleRate)
signal_duration = int((sum(l[-1]) - l[0][0]) * t.sampleRate)
vecin = t.data[signal_start:signal_start + signal_duration,:]
rates = []
for i in xrange(len(l)):
rates.append((int(l[i][0] * t.sampleRate) - signal_start, self.durations[i] / l[i][1]))
vecout = dirac.timeScale(vecin, rates, t.sampleRate, 0)
if hasattr(t, 'gain'):
vecout = limit(multiply(vecout, float32(t.gain)))
return AudioData(ndarray=vecout, shape=vecout.shape, sampleRate=t.sampleRate, numChannels=vecout.shape[1])
def stretch(self, t, l):
"""t is a track, l is a list"""
signal_start = int(l[0][0] * t.sampleRate)
signal_duration = int((sum(l[-1]) - l[0][0]) * t.sampleRate)
vecin = t.data[signal_start:signal_start + signal_duration,:]
rates = []
for i in xrange(len(l)):
rate = (int(l[i][0] * t.sampleRate) - signal_start,
self.durations[i] / l[i][1])
rates.append(rate)
vecout = dirac.timeScale(list(vecin), rates, t.sampleRate, 0)
if hasattr(t, 'gain'):
vecout = limit(multiply(vecout, float32(t.gain)))
audio_out = AudioData(ndarray=vecout, shape=vecout.shape,
sampleRate=t.sampleRate,
numChannels=vecout.shape[1])
return audio_out