You're reading an old version of this documentation. If you want up-to-date information, please have a look at 0.10.2.


librosa.samples_like(X, hop_length=512, n_fft=None, axis=-1)[source]

Return an array of sample indices to match the time axis from a feature matrix.

Xnp.ndarray or scalar
  • If ndarray, X is a feature matrix, e.g. STFT, chromagram, or mel spectrogram.

  • If scalar, X represents the number of frames.

hop_lengthint > 0 [scalar]

number of samples between successive frames

n_fftNone or int > 0 [scalar]

Optional: length of the FFT window. If given, time conversion will include an offset of n_fft // 2 to counteract windowing effects when using a non-centered STFT.

axisint [scalar]

The axis representing the time axis of X. By default, the last axis (-1) is taken.

samplesnp.ndarray [shape=(n,)]

ndarray of sample indices corresponding to each frame of X.

See also


Return an array of time values to match the time axis from a feature matrix.


Provide a feature matrix input:

>>> y, sr = librosa.load(librosa.ex('trumpet'))
>>> X = librosa.stft(y)
>>> samples = librosa.samples_like(X)
>>> samples
array([     0,    512, ..., 116736, 117248])

Provide a scalar input:

>>> n_frames = 2647
>>> samples = librosa.samples_like(n_frames)
>>> samples
array([      0,     512,    1024, ..., 1353728, 1354240, 1354752])