FFFT: Fast Full Fourier Transform

Experiments in spectral domain by taking frequency transform of an entire
fragment of music, changing spectral features and transforming back to the
time domain

Marc Groenewegen - marc.groenewegen@hku.nl
Pieter suurmond - pieter.suurmond@hku.nl

***********************************************************************
           N.B.: turn down the volume with these experiments!
         Serious risk of blowing up speakers or damaging your ears.
          Best played in combination with visual representation
	    such as sonic-visualiser or audacity.

    EDIT March 2025: in the C++ version, first and last 0.5 second
   of resulting audio are not written to the output file for this reason
***********************************************************************

Prerequisites:
- libfftw - C library for discrete Fourier transform in one or more dimensions
  http://fftw.org
- MFFM FFTw C++ wrapper
  http://sourceforge.net/projects/mffmfftwrapper
- libsndfile - sound file reading and creation
  http://www.mega-nerd.com/libsndfile

For Ubuntu and other Debian derivatives:
install libfftw3-dev, mffm-fftw-dev and libsndfile1-dev

The main file ffft.cpp is compiled by typing 'make'. If you installed the
mentioned prerequisites this should build the executable ffft

Call the program with two filenames on the command line: a file for input
and a file for output. The number of channels of the input file can be
greater than one. The program uses only the first channel and produces a
mono output file.

The unmodified program returns the original audio signal. This is useful
for testing.

When modifying the program, be sure you visually check the resulting sound
file before playing it and play at low volume. These spectral tricks can
result in accumulation of energy at certain points in the fragment.