def compute(self, x):
startstopsilence = es.StartStopSilence(threshold=-70)
for frame in es.FrameGenerator(x,
frameSize=2048,
hopSize=512,
startFromZero=True):
result = startstopsilence(esarray(frame))
solution = np.array([result[0], result[1]])
return solution * 512 / 44100.
def onsetDetection(self, samplePath):
# Get audio from file
loader = es.MonoLoader(filename=samplePath)
audio = loader()
startStop = es.StartStopSilence()
startFrame = 0
stopFrame = 0
for frame in es.FrameGenerator(audio, 64, 32):
startFrame, stopFrame = startStop(frame)
startTime = float(startFrame * 32) / 44100.0
stopTime = float((stopFrame * 32) + 64) / 44100.0
return startTime, stopTime
def find_start_stop(audio, sample_rate=44100, seconds=False, threshold=-60):
"""
Returns a tuple containing the start and the end of sound in an audio
array.
ARGUMENTS:
`audio` : essentia.array
an essentia array or numpy array containing the audio
`sample_rate` : int
sample rate
`seconds` : boolean
if True, results will be expressed in seconds (float)
`threshold` : int
a threshold as in `essentia.standard.StartStopSilence`
RETURNS:
`start` : int or float
the sample where sound starts or the corresponding second
`end` : int or float
the sample where sound ends or the corresponding second
"""
# reset parameters based on sample_rate
ratio = sample_rate / 44100
fs = round(1024 * ratio)
hs = round(128 * ratio)
processer = es.StartStopSilence(threshold=threshold)
for frame in es.FrameGenerator(audio,
frameSize=fs,
hopSize=hs,
startFromZero=True):
start, stop = processer(frame)
if seconds:
start = specframe2sec(start, sample_rate, hs, fs)
stop = specframe2sec(stop, sample_rate, hs, fs)
else:
start = int(specframe2sample(start, hs, fs))
stop = int(specframe2sample(stop, hs, fs))
if start == 2 * hs:
start = 0
return start, stop