Author: red (05-051.026.popsite.net)
Date: 01-13-2002 21:27
Elliot,
The method used by WinRec for compression is a very simple delta method. It stores the changes from sample to sample instead of absolute values and uses an 8-bit lookup table that contains the changes from sample to sample. I think Cubase Audio 16track mode uses a similar method. It works very well and is very easy to implement (the DSP would have loads of time left for other stuff).
Here some compression info for ya:
http://www.wavpack.com/technical.htm
http://www.openmash.org/lxr/source/codec/audio/adpcm.c
I think Adaptive Difference Pulse Code Modulation (ADPCM) sounds ideal for your application. There is DSP code available for decompressing ADPCM (Motorola had it on their site at a while back). However, the C source code I listed above for an ADPCM coder looks pretty simple!
