bsespede / speech-compression

Octave implementation of the FFT to compress audio

Speech compression

Using FFT to compress audio.

#HOW-TO

• To compress, uncompress, generate audio and make plots for different epsilons (100 number between 0.002 and 0.2): plot_fixed_bits('11',0.0002, 10, 2, 8) genera: 11_epsi_compression.png 11_epsi_distortion.png

• To compress, uncompress, generate audio and make plots for different number of bits (2 to 8 in this case): plot_fixed_epsilon('11', 0.2, 2, 8) genera: 11_bits_compression.png 11_bits_distortion.png

Credits

Used @PGryllos huffmandict implementation: https://github.com/PGryllos/nhuff