Caution

You're reading an old version of this documentation. If you want up-to-date information, please have a look at 0.10.1.

librosa.estimate_tuning

librosa.estimate_tuning(*, y=None, sr=22050, S=None, n_fft=2048, resolution=0.01, bins_per_octave=12, **kwargs)[source]

Estimate the tuning of an audio time series or spectrogram input.

Parameters:
ynp.ndarray [shape=(…, n)] or None

audio signal. Multi-channel is supported..

srnumber > 0 [scalar]

audio sampling rate of y

Snp.ndarray [shape=(…, d, t)] or None

magnitude or power spectrogram

n_fftint > 0 [scalar] or None

number of FFT bins to use, if y is provided.

resolutionfloat in (0, 1)

Resolution of the tuning as a fraction of a bin. 0.01 corresponds to measurements in cents.

bins_per_octaveint > 0 [scalar]

How many frequency bins per octave

**kwargsadditional keyword arguments

Additional arguments passed to piptrack

Returns:
tuning: float in [-0.5, 0.5)

estimated tuning deviation (fractions of a bin).

Note that if multichannel input is provided, a single tuning estimate is provided spanning all channels.

See also

piptrack

Pitch tracking by parabolic interpolation

Examples

With time-series input

>>> y, sr = librosa.load(librosa.ex('trumpet'))
>>> librosa.estimate_tuning(y=y, sr=sr)
-0.08000000000000002

In tenths of a cent

>>> librosa.estimate_tuning(y=y, sr=sr, resolution=1e-3)
-0.016000000000000014

Using spectrogram input

>>> S = np.abs(librosa.stft(y))
>>> librosa.estimate_tuning(S=S, sr=sr)
-0.08000000000000002

Using pass-through arguments to librosa.piptrack

>>> librosa.estimate_tuning(y=y, sr=sr, n_fft=8192,
...                         fmax=librosa.note_to_hz('G#9'))
-0.08000000000000002