How to split channels

[Filippo]

Hi!

I am loading a WAV file with sf::InputSoundFile in C++ and I want to extract the samples for each channel separately. As far as I know the method sf::InputSoundFile::read(Int16 *samples, Uint64 maxCount) writes an array of interleaved samples to samples. Each sample is then made up by 16 bits, the first 8 belonging to a channel, and the remaining 8 belonging to the other. My silly question is: should I consider each 8-bits part as an Int8 or as a Uint8? I am very much confused.

A code snippet that performs the conversion from interleaved to separate channels would be also appreciated! 

What follows is my attempt to do this conversion.

sf::Int16 samples[1024];
sf::Int16 mono[1024];
file.read(samplesData, 1024);
if (file.getChannelCount() == 2){
  for(int k = 0; k < 1024; ++k){
    sf::Int16 right = samples[k] & 0x00FF;
    sf::Int16 left = (samples[k] & 0xFF00) >> 8;
    mono[k] = (right+left)/2; // mean of the two channels
  }
}

[eXpl0it3r]

writes an array of interleaved samples to samples.

This is true.

Each sample is then made up by 16 bits, the first 8 belonging to a channel, and the remaining 8 belonging to the other.

I have no idea where you heard that and it totally contradicts your previous statement.

Interleaved means, the first sample is from channel 1, the second sample is from channel 2, the third sample is from channel 1, the forth sample is from channel 2 and so on... (if you have stereo).

Hope that helps.

[Filippo]

I have no idea where you heard that

Actually, this is the bad source, though probably I misread it
https://stackoverflow.com/questions/48804523/is-it-possible-to-split-left-and-right-channel-from-audio-sample-in-sfml

Hope that helps.

It really helps a lot!! Thank you!! I've been stuck for hours thinking about how to split those 16 bits. 

[Hapax]

Since the stackexchange question also asks how to split the channels, I posted an answer (and example) there.  It should help you

https://stackoverflow.com/a/50078065/6158263