Caution

You're reading the documentation for a development version. For the latest released version, please have a look at 0.9.1.

librosa.util.stack¶

librosa.util.stack(arrays, *, axis=0)[source]¶

Stack one or more arrays along a target axis.

This function is similar to np.stack, except that memory contiguity is retained when stacking along the first dimension.

This is useful when combining multiple monophonic audio signals into a multi-channel signal, or when stacking multiple feature representations to form a multi-dimensional array.

Parameters

arrayslist: one or more np.ndarray
axisinteger: The target axis along which to stack. axis=0 creates a new first axis, and axis=-1 creates a new last axis.

Returns

arr_stacknp.ndarray [shape=(len(arrays), array_shape) or shape=(array_shape, len(arrays))]

The input arrays, stacked along the target dimension.

If axis=0, then arr_stack will be F-contiguous. Otherwise, arr_stack will be C-contiguous by default, as computed by np.stack.

Raises

ParameterError

If arrays do not all have the same shape
If no arrays are given

See also

numpy.stack
numpy.ndarray.flags
frame

Examples

Combine two buffers into a contiguous arrays

>>> y_left = np.ones(5)
>>> y_right = -np.ones(5)
>>> y_stereo = librosa.util.stack([y_left, y_right], axis=0)
>>> y_stereo
array([[ 1.,  1.,  1.,  1.,  1.],
       [-1., -1., -1., -1., -1.]])
>>> y_stereo.flags
  C_CONTIGUOUS : False
  F_CONTIGUOUS : True
  OWNDATA : True
  WRITEABLE : True
  ALIGNED : True
  WRITEBACKIFCOPY : False
  UPDATEIFCOPY : False

Or along the trailing axis

>>> y_stereo = librosa.util.stack([y_left, y_right], axis=-1)
>>> y_stereo
array([[ 1., -1.],
       [ 1., -1.],
       [ 1., -1.],
       [ 1., -1.],
       [ 1., -1.]])
>>> y_stereo.flags
  C_CONTIGUOUS : True
  F_CONTIGUOUS : False
  OWNDATA : True
  WRITEABLE : True
  ALIGNED : True
  WRITEBACKIFCOPY : False
  UPDATEIFCOPY : False