adzialocha/spectral-centroid-splitter

Splits a wav file into small fractions and sorts them after a spectral-centroid analysis

spectral-centroid-splitter

Splits a uncompressed wav file into small fractions and sorts them after a spectral-centroid analysis.

Writes fractions as wave audio file, json data for further processing and png fft analysis plot figure in specified output directory.

Requirements

• python 2.7
• matplotlib 1.4.3
• numpy 1.9.2

Usage

usage: spectral-centroid-splitter.py [-h] [--duration [ms]] [--sort-grid [hz]]
                                     [--output-dir path]
                                     input-wav-file

positional arguments:
  input-wav-file     input wave file path (uncompressed)

optional arguments:
  -h, --help         show this help message and exit
  --duration [ms]    duration of one wave fraction (in milliseconds)
  --sort-grid [hz]   sort grid after spectral analysis (in hz steps)
  --output-dir path  output directory path

Example

python spectral-centroid-splitter.py test.wav --duration 50 --sort-grid 100

Name                  "test"
Duration              4.595021s
Sample-Rate           16bit
Frame-Rate            48000hz
Fraction-Size-Frames  2397
Fraction-Sort-Grid    100
Fraction-Duration     0.050000s
Fractions-Count       92

Example output file @ ./out/1300/test_out_2.png (spectral centroid at around 1300hz)