Skip to main content

Tools for generating noisy and reverberant audio files

Project description

Build Status

maracas is a library for corrupting audio files with additive and convolutive noise. Its objective is to simplify reproducible dataset generation for speech processing (mainly enhancement and ASR).

The usage is really simple and based on the maracas.dataset.Dataset class. Here is a short example:

from maracas.dataset import Dataset
import numpy as np

# Make sure this is reproducible

d = Dataset()

# All files can be added one by one or by folder. Adding a folder will add
# all speech files inside that folder recursively if recursive=True.
d.add_speech_files('/home/jfsantos/data/speech_files/', recursive=True)

# When adding noises, you can give a "nickname" to each noise file. If you do not
# give it a name, the name will be the file name without the '.wav' extension
d.add_noise_files('/home/jfsantos/data/multichannel_noises/restaurant_ch01.wav', name='restaurant')
d.add_noise_files('/home/jfsantos/data/multichannel_noises/cafeteria_ch01.wav', name='cafeteria')
d.add_noise_files('/home/jfsantos/data/multichannel_noises/traffic_ch01.wav', name='traffic')

# Adding reverb files works like adding noise files

# When generating a dataset, you can choose which SNRs will be used and how many
# files per condition you want to be generated. 
d.generate_dataset([-6, -3, 0, 3, 6], '/tmp/noise_plus_reverb_dataset', files_per_condition=5)

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

maracas-0.0.1.tar.gz (5.1 kB view hashes)

Uploaded source

Built Distribution

maracas-0.0.1-py3-none-any.whl (6.7 kB view hashes)

Uploaded py3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page