octave-based frequency
"""
// If we don"t have a spectrogram, build one
if S is None:
// By default, use a magnitude spectrogram
S = np.abs(stft(y, n_fft=n_fft, hop_length=hop_length))
else:
// Infer n_fft from spectrogram shape
n_fft = (S.shape[0] - 1) * 2
// Compute the center frequencies of each bin
if freq is None:
freq = fft_frequencies(sr=sr, n_fft=n_fft)
// TODO: 2014-12-31 12:48:36 by Brian McFee <brian.mcfee@nyu.edu>
// shouldn"t this be scaled relative to the max frequency?
octa = np.zeros(n_bands + 2)
octa[1:] = 200 * (2.0**np.arange(0, n_bands + 1))
valley = np.zeros((n_bands + 1, S.shape[1]))
peak = np.zeros_like(valley)
for k, (f_low, f_high) in enumerate(zip(octa[:-1], octa[1:])):
current_band = np.logical_and(freq >= f_low, freq <= f_high)
idx = np.flatnonzero(current_band)
if k > 0:
current_band[idx[0] - 1] = True
if k == n_bands:
current_band[idx[-1] + 1:] = True
sub_band = S[current_band]
if k < n_bands:
sub_band = sub_band[:-1]
// FIXME: 2014-12-31 13:06:49 by Brian McFee <brian.mcfee@nyu.edu>
// why 50? what is this?
alph = int(max(1, np.rint(0.02 * np.sum(current_band))))
sortedr = np.sort(sub_band, axis=0)
valley[k] = np.mean(sortedr[:alph], axis=0)
peak[k] = np.mean(sortedr[-alph:], axis=0)
return peak - valley
