Digital signal processing

Module contents

DSP

This sub-package provides all dsp algorithms.

Classes

The main class is simsalabim.dsp.classes.Signal.

class simsalabim.dsp.classes.Signal(data, sampling_rate, comment=None, dtype=<class 'numpy.float64'>)[source]

Bases: object

Class for audio signals. Objects of this class contain data which is directly convertable between time and frequency domain (equally spaced samples and frequency bins).

property comment: Get comment.

property data: Get data.

property dtype: The data type of the audio object. This can be any data type and precision supported by numpy.

property duration: The duration of the signal in seconds.

property n_samples: The number of samples.

property sampling_rate: The sampling rate of the signal.

Utilities

Decibel

Provides

Basic calculations with levels.

simsalabim.dsp.decibel.dbadd(a, b)[source]

Energetic addition of levels.

Parameters

a (float) – Single level or sequence of levels.
b (float) – Single level or sequence of levels.

Notes

$L_{a+b} = 10 \log_{10}{10^{L_b/10}+10^{L_a/10}}$

Energetically adds b to a.

simsalabim.dsp.decibel.dbmean(levels, axis=None)[source]

Energetic average of levels.

Parameters: levels (list, nparray) – Sequence of levels.

Notes

$L_{mean} = 10 \log_{10}{\frac{1}{n}\sum_{i=0}^n{10^{L/10}}}$

simsalabim.dsp.decibel.dbsub(a, b)[source]

Energetically subtract level b from level a.

Parameters

a (float) – Level.
b (float) – Level.

Notes

$L_{a-b} = 10 \log_{10}{10^{L_a/10}-10^{L_b/10}}$

Energitally subtract b from a.

simsalabim.dsp.decibel.dbsum(levels, axis=None)[source]

Summation of levels.

Notes

$L_{sum} = 10 \log_{10}{\sum_{i=0}^n{10^{L/10}}}$

SNR

Provides

WADA (Waveform Amplitude Distribution Analysis) SNR calculation.

simsalabim.dsp.snr.wada_snr(wav)[source]

Direct blind estimation of the SNR of a speech signal.

Parameters: wav (str) – Filename of the wav file.

Notes

The function was adapted from public Matlab code by John Meade and can be found in this gist

References

1: Kim, Chanwoo, and Richard M. Stern. “Robust signal-to-noise ratio estimation based on waveform amplitude distribution analysis.” Ninth Annual Conference of the International Speech Communication Association. 2008.