Hiya all, I'm trying to write an ADPCM (VOX) encoder / decoder following Dialogic's algorithm as documented at: http://resource.intel.com/telecom/support/appnotes/adpcm.pdf (I've replaced the 12bit step size table with the equivalent 16bit table.) Things seem straight forward enough until I come to test it. I'm taking a wav file and running it through my encoder which I'm then comparing against the same wav but converted within SoundForge. SoundForge's output follows the form of the original wav very well, but the amplitudes in my program increase so much that they are actually clipped. I've tried using source code examples I've found on the web, but their results are just as bad. Has any one got experience creating working ADPCM encoders or has any suggestions what extra audio processing SoundForge might be doing? thanks Simon
adpcm vox...
Started by ●October 23, 2002