Why Bitwise Reproducibility Matters

Re:Why Bitwise Reproducibility Matters by ArkyBeagle on Saturday September 19, 2015 @07:03PM (#10245849)

So if you go get FFTW, build it into a program on a PeeCee, then round-trip a .wav file ( .wav=>FFT; FFT->.wav ) you get your original file back.

This was years ago, and I may still have been using 16 bit samples. It would be worth repeating this for 16 bit, 24 bit and 32-but float sample sets.

That's not bad.

[ Parent | on HN ]

Re:Why Bitwise Reproducibility Matters by dnautics on Saturday September 19, 2015 @07:08PM (#10245864)

.wav encodes amplitude data in what is essentially an integer representation (well, it's a fixed point float, but anyways).

[ Parent | on HN ]

Re:Why Bitwise Reproducibility Matters by ArkyBeagle on Saturday September 19, 2015 @09:39PM (#10246255)

Sure. They're constrained to 0...1 . But there are a lot of floating point operations per FFT, and it can be shown that they are invertible.

[ Parent | on HN ]