music.normalize_stereo

music.normalize_stereo(sonic_vector, remove_bias=True, normalize_sep=False)

Normalize a stereo sonic vector.

The final array will have values only between -1 and 1.

Parameters:

sonic_vectorarray_like

A (2, nsamples) shaped array. A one-dimensional array is taken as mono and given two identical channels.

remove_biasbool

If True (default), subtract each channel’s mean before scaling. If False, map the sample range affinely onto [-1, 1], using each channel’s range when normalize_sep is True and the combined range otherwise.

normalize_sepbool

Set to True if each channel should be normalized separately. If False (default), the arrays will be rescaled in the same proportion (preserves loudness proportion).

Returns:

sv_normalizedndarray

A numpy array with values between -1 and 1.

Raises:

ValueError

If the sonic vector is empty; see normalize_mono() for what an empty render means.

Parameters:

• sonic_vector (ArrayLike)

• remove_bias (bool)

• normalize_sep (bool)

Return type:

NDArray[float64]

Examples

>>> quiet = np.vstack([note(220, 0.1) * 0.1, note(330, 0.1) * 0.4])
>>> loudest = float(abs(normalize_stereo(quiet)).max())
>>> round(loudest, 6)
1.0
>>> apart = normalize_stereo(quiet, normalize_sep=True)
>>> [round(float(abs(channel).max()), 6) for channel in apart]
[1.0, 1.0]