You're reading the documentation for a development version. For the latest released version, please have a look at 0.10.2.


Array operations

frame(x, *, frame_length, hop_length[, ...])

Slice a data array into (overlapping) frames.

pad_center(data, *, size[, axis])

Pad an array to a target length along a target axis.

expand_to(x, *, ndim, axes)

Expand the dimensions of an input array with

fix_length(data, *, size[, axis])

Fix the length an array data to exactly size along a target axis.

fix_frames(frames, *[, x_min, x_max, pad])

Fix a list of frames to lie within [x_min, x_max]

index_to_slice(idx, *[, idx_min, idx_max, ...])

Generate a slice array from an index array.

softmask(X, X_ref, *[, power, split_zeros])

Robustly compute a soft-mask operation.

stack(arrays, *[, axis])

Stack one or more arrays along a target axis.

sync(data, idx, *[, aggregate, pad, axis])

Aggregate a multi-dimensional array between specified boundaries.

axis_sort(S, *[, axis, index, value])

Sort an array along its rows or columns.

normalize(S, *[, norm, axis, threshold, fill])

Normalize an array along a chosen axis.

shear(X, *[, factor, axis])

Shear a matrix by a given factor.

sparsify_rows(x, *[, quantile, dtype])

Return a row-sparse matrix approximating the input

buf_to_float(x, *[, n_bytes, dtype])

Convert an integer buffer to floating point values.


Compute the tiny-value corresponding to an input's data type.


match_intervals(intervals_from, intervals_to)

Match one set of time intervals to another.

match_events(events_from, events_to[, left, ...])

Match one set of events to another.


localmax(x, *[, axis])

Find local maxima in an array

localmin(x, *[, axis])

Find local minima in an array

peak_pick(x, *, pre_max, post_max, pre_avg, ...)

Use a flexible heuristic to pick peaks in a signal.

nnls(A, B, **kwargs)

Non-negative least squares.

cyclic_gradient(data, *[, edge_order, axis])

Estimate the gradient of a function over a uniformly sampled, periodic domain.

dtype_c2r(d, *[, default])

Find the real numpy dtype corresponding to a complex dtype.

dtype_r2c(d, *[, default])

Find the complex numpy dtype corresponding to a real dtype.

count_unique(data, *[, axis])

Count the number of unique values in a multi-dimensional array along a given axis.

is_unique(data, *[, axis])

Determine if the input array consists of all unique values along a given axis.

abs2(x[, dtype])

Compute the squared magnitude of a real or complex array.

phasor(angles, *[, mag])

Construct a complex phasor representation from angles.

Input validation

valid_audio(y, *[, mono])

Determine whether a variable contains valid audio data.

valid_int(x, *[, cast])

Ensure that an input value is integer-typed.


Ensure that an array is a valid representation of time intervals:


Check that x is a positive integer, i.e. 1 or greater.

File operations

example(key, *[, hq])

Retrieve the example recording identified by 'key'.


Display licensing and metadata information for the given example recording.


List the available audio recordings included with librosa.

find_files(directory, *[, ext, recurse, ...])

Get a sorted list of (audio) files in a directory or directory sub-tree.


Print the citation information for librosa.