def extract(self):
Features.extract(self)
self._hop = 1 if self._hop is None else self._hop
window, hop = self._window, self._hop
if window and hop is not None:
fp = self.feature_params
num_frames = int(
(window * fp['sample_rate']) / (1000.0 * fp['nhop']))
num_hop = int((hop * fp['sample_rate']) / (1000.0 * fp['nhop']))
print(num_frames, num_hop)
if not num_frames and num_hop:
raise ValueError(
"num_frames and num_hop too small for FFT window / hop")
else:
Y = []
for k in range(0, self.X.shape[1] - window + 1, num_hop):
X = log(
self.X[:, np.arange(k, k + num_frames)] +
np.finfo(np.float32).eps
) if self._log else self.X[:,
np.arange(k, k + num_frames)]
Y.append(
np.fft.fftshift(np.absolute(np.fft.fft2(X))).flatten())
self.X = np.array(Y)
else:
self.X = log(self.X +
np.finfo(np.float32).eps) if self._log else self.X
self.X = np.fft.fftshift(np.absolute(np.fft.fft2(self.X)))
def extract(self):
Features.extract(self)
self._hop = 1 if self._hop is None else self._hop
window, hop = self._window, self._hop
if window and hop is not None:
fp = self.feature_params
num_frames = int((window*fp['sample_rate'])/(1000.0*fp['nhop']))
num_hop = int((hop*fp['sample_rate'])/(1000.0*fp['nhop']))
print num_frames, num_hop
if not num_frames and num_hop :
raise ValueError("num_frames and num_hop too small for FFT window / hop")
else :
Y = []
for k in range(0,self.X.shape[1]-window+1,num_hop):
X = log(self.X[:,np.arange(k,k+num_frames)]+np.finfo(np.float32).eps) if self._log else self.X[:,np.arange(k,k+num_frames)]
Y.append(np.fft.fftshift(np.absolute(np.fft.fft2(X))).flatten())
self.X = np.array(Y)
else:
self.X = log(self.X+np.finfo(np.float32).eps) if self._log else self.X
self.X = np.fft.fftshift(np.absolute(np.fft.fft2(self.X)))
def extract(self):
Features.extract(self)
self.X = P.sqrt((P.diff(self.X)**2).sum(0)) / self.X.shape[0]
def extract(self):
Features.extract(self)
mf = (self.X.T * self._logfrqs).sum(1) / self.X.T.sum(1)
self.X = (((self.X / self.X.T.sum(1)).T *
((P.atleast_2d(self._logfrqs).T - mf)).T)**2).sum(1)
def extract(self):
Features.extract(self)
self.X = (self.X.T * self._logfrqs).sum(1) / self.X.T.sum(1)
def extract(self):
Features.extract(self)
self.X = P.sqrt((P.diff(self.X)**2).sum(0))/self.X.shape[0]
def extract(self):
Features.extract(self)
mf = (self.X.T * self._logfrqs).sum(1) / self.X.T.sum(1)
self.X = (((self.X / self.X.T.sum(1)).T * ((P.atleast_2d(self._logfrqs).T - mf)).T)**2).sum(1)
def extract(self):
Features.extract(self)
self.X = (self.X.T * self._logfrqs).sum(1) / self.X.T.sum(1)