chunkup

Chunkup

Given an audio file, and a tab delimited file, split up audio file into smaller chunks.

Requirements & Caveats

Imports pysox, therefore has the same dependencies as pysox, namely:

Required prerequisite are the development libraries of sox at version 14.3.x, i.e. the header files and libraries to link against. Specifically you need sox.h in your include path and libsox.so and libsox.a in your link path. Pysox will not compile against any sox version prior to 14.3.0.
chunkup will only chunk up audio. Any other effects (resampling etc) should be applied to the audio file before running chunkup.
At the moment, chunkup will operate fairly blindly. If you give it start and end times for a chunk that actually lie outside of the provided audio file, it’ll return a bunch of sox errors that look like:

trim: End position is after expected end of audio.
trim: Last 1 position(s) not reached.

Installation

pip install chunkup

After installing chunkup, the chunkup command line script should be added to your path, and should be available from anywhere. Test it out by opening a new terminal and running

chunkup -h

Usage

This usage example pulls down Episode #20: I Want to Break Free of Reply All. The mp3 is 41.5M. You can get replyallcreak.txt from https://github.com/JoFrhwld/chunkup/blob/master/examples/replyallcreak.txt

curl -L -o examples/reply_all_podcast.mp3 http://bit.ly/chunkup
mkdir chunks

chunkup examples/reply_all_podcast.mp3 examples/replyallcreak.txt chunks/

Chunk naming

chunks/ now contains 104 chunks of speech I annotated, largely to separate out the three speakers in the first segment. The default naming conventions of chunks are:

[n]-[basename]-[col1]-[col3].wav

Where [n] is the numeric index of the chunk, [basename] is the base name of the original soundfile, [col1] is the value from column 1 in replyallcreak.txt (in this case, speaker ID), and [col3] is the value from column 3 in replyallcreak.txt (in this case, the onset time of the chunk in ss.ms). Available variables for naming chunks are:

naming variable	description
[n]	chunk number
[basename]	base name of original audio file
[col0-9]	value from any the given column in the chunk file

Configuration

You can change the chunk naming convention either at the command line usig the -n or --naming flag.

$ chunkup -n [col1]-[n].wav examples/reply_all_podcast.mp3 examples/replyallcreak.txt chunks/

Other config options include

option	description
-s, --start	Column index (starting with 1) for the start time, in ss.ms format
-e, --end	Column index (starting with 1) for the end time, in ss.ms format
--header	Include if chunk file has a header

All config options can be defined in a config file and passed to chunkup.py with the prefix +. A sample config file for this data would be:

--naming
[n]-[basename]-[col1]-[col3].wav
--start
3
--end
4

If saved to config.txt, it could be passed to chunkup.py like so:

$ chunkup +config.txt examples/reply_all_podcast.mp3 examples/replyallcreak.txt chunks/

Release files for chunkup 0.2.1

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for chunkup 0.2.1
File	Size	Uploaded
chunkup-0.2.1.tar.gz	4.6 kB	Apr 19, 2015	Details

Release files / chunkup-0.2.1.tar.gz

Download URL	chunkup-0.2.1.tar.gz
Size	4.6 kB
Tags	Source
SHA-256 checksum How to use checksums	`c592a7edac70dcc4d99bf990e8248807c4d0b6bd8be9e0f0ef0afb8308aac14e`
BLAKE2b-256 checksum How to use checksums	`a7d8892e174bda2efea401ea73e227b611aa7056f50d1499a0beb5dee0a5ed2b`
Upload date	Apr 19, 2015
Uploaded using Trusted Publishing? What is trusted publishing?	No