A Python module for reading and writing WAV files using numpy arrays.
Project description
wavio is a Python module that defines two functions:
wavio.read reads a WAV file and returns an object that holds the sampling rate, sample width (in bytes), and a numpy array containing the data.
wavio.write writes a numpy array to a WAV file, optionally using a specified sample width.
The functions can read and write 8-, 16-, 24- and 32-bit integer WAV files.
The module uses the wave module in Python’s standard library, so it has the same limitations as that module. In particular, the wave module does not support compressed WAV files, and it does not handle floating point WAV files. When floating point data is passed to wavio.write it is converted to integers before being written to the WAV file.
wavio requires Python 3.6 or later.
wavio depends on numpy (http://www.numpy.org). NumPy version 1.19.0 or later is required. The unit tests in wavio require pytest.
The API of the functions in wavio should not be considered stable. There may be backwards-incompatible API changes between releases.
Important notice
In version 0.0.5, the data handling in wavio.write has been changed in a backwards-incompatible way. The API for scaling the input in 0.0.4 was a flexible interface that only its creator could love. The new API is simpler, and it is hoped that it does the right thing by default in most cases. In particular:
When the input data is an integer type, the values are not scaled or shifted. The only changed that might happen is the data might be clipped if the values do not fit in the output integer type.
By default, floating point input is scaled to the full width of the output integer type, with the constraint that 0.0 in the input is mapped to the midpoint of the output integer type. The scale parameter allows that behavior to be changed–it gives the upper bound of the float values that are mapped to the maximum of the output integer type. That is, it defines what the user considers the “full scale” amplitude of the input to be. Regardless of the value of scale, the float input 0.0 is always mapped to the midpoint of the output type; wavio.write will not translate the values up or down.
A warning is now generated if any data values are clipped. A parameter allows the generation of the warning to be disabled or converted to an exception.
Example
The following code (also found in the docstring of wavio.write) writes a three second 440 Hz sine wave to a 24-bit WAV file:
import numpy as np import wavio rate = 22050 # samples per second T = 3 # sample duration (seconds) n = int(rate*T) # number of samples t = np.arange(n)/rate # grid of time values f = 440.0 # sound frequency (Hz) x = np.sin(2*np.pi * f * t) wavio.write("sine24.wav", x, rate, sampwidth=3)
- Author:
Warren Weckesser
- Repository:
- License:
BSD 2-clause (http://opensource.org/licenses/BSD-2-Clause)
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for wavio-0.0.6-py2.py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 3eb78c893a190db838c682eb914fc322e2bddc6afe45533660d9ec82df44d235 |
|
MD5 | 58195b9daa9ae15311f58fb04332198b |
|
BLAKE2b-256 | cc8fad64bd39f0596109af84d25f5c3945823c2c0e2450a12d5855baf73b4324 |