Anyone know the design rationale for recording IQ wave files as floats? It looks like the code will playback wave files that are 16 bit ints. I'd think we could code them as 24 bit ints without (significant?) loss as that's what the soundcards are feeding us.

Reason for the question is I'm looking at using FLAC (Free Lossless Audio Compression - http://flac.sourceforge.net/ ) to compress recorded wave files and unfortunately it does not handle wave files encoded as floats. It does apparently handle 24 and 16 bit ints. It looks like at one point PowerSDR recorded wave files as ints as there's at least one 16 bit int encoded IQ wave file (the Aurora file from the Dayton Demo CD) running around in the wild that PowerSDR will playback.

Also, a second question, where in the signal processing pipe is the data being recorded grabbed from -- is it after IQ fixup? That might explain the recording as floats. Will take a look @ the code and see if I can puzzle it out, but would be interested to hear from anyone that knows off the top of their head.

Cheers,

Bill (kd5tfd)



Reply via email to