The Code Singer

The Code Singer is a sort of code obfuscator that inserts and removes insignificant white space in order to match the bytes of a target audio file, whilst not breaking or changing the semantics of the source code.

Conceived during Music Hackday London 2012.

Usage

./codesinger -t threshold target < infile > outfile

threshold is a number in the range [-1, 1] that defines at what target amplitude spaces will be inserted. The higher the value, the sparser the output waveform. Typically, threshold should be between [0, .3].

target is an audio file, in a format understood by libsndfile, e.g. WAV, AIFF, AU, etc.

infile is the input source code; outfile is the modified source code as a raw audio file with 8 bit depth and the same sample rate as target.

Syntax considerations

The Code Singer only works (as in, doesn't break the code) with C-like languages. The following syntactic constructs are considered:

• Single quote enclosed strings
• Double quote enclosed strings
• C-style, multi-line comments (/* */)
• C++-style, single-line comments (//)
• C preprocessor constructs (where # is the first character on the line)

Dependencies

The only dependency is libsndfile.

Bugs

Definitely!