Robert_Salesas@mindlink.UUCP (Robert Salesas) (02/12/91)
You might want to try the fibonici delta method of compression. Its a lossy method but gives good results on sampled data. Check the Amiga areas for info as it is used in the IFF 8SVX standard. Rob
sbarnhar%MAILBOX.MAIL.UMN.EDU@uga.cc.uga.edu ( Shawn Barnhart) (02/13/91)
I don't know much about digitized sound or compression algorithms, but my guess is that sound compression involves sampling the sampled sound. IE, say you sample a sound at 22 Khz and you want to compress it. Wouldn't you just sample the sample? Of course this means that you can't go back to the original sample rate. Maybe you could save the information that you don't sample from the sample and try to compress that using the standard compression utilities/algorithms (Lempl-Ziv, Huffman, et al). I've never tried, so your mileage may vary. There are (supposedly) "advanced" compression utilities for the Mac that can obtain 20-30% compression of digitized sound (as opposed to resampling). I don't know what algorithms they may use, though. I imagine they are proprietary versions of some of the more standard algorithms. -Shawn
00csgunn@bsu-ucs.uucp (II INFINATUM) (02/13/91)
In article <25947@adm.brl.mil>, sbarnhar%MAILBOX.MAIL.UMN.EDU@uga.cc.uga.edu (Shawn Barnhart) writes: > There are (supposedly) "advanced" compression utilities for the Mac that can > obtain 20-30% compression of digitized sound (as opposed to resampling). > I don't know what algorithms they may use, though. I imagine they are > proprietary versions of some of the more standard algorithms. > -Shawn The routines used to compress digitized sound flies on the Mac, and the Apple IIgs for that matter, use a rather weird compression algorythm. They look at the sound as that... Sound. And predict what the next few bytes of the waveform will look like. If the prediction is right, then the data can be compressed. Im know this is rather vague, but this is how it was described to me. There is also another compression on the IIgs that gives either 3:8 or 4:8 compression. Im not really sure how that works either. Hope this helps. -- -------------------------------------------------------------------------------- Steve Gunn Internet --> hangtime@bsu-cs.bsu.edu The Indiana Academy 00csgunn@bsu-ucs.bsu.edu Ball State University UUCP --> {BackBones}!iuvax!bsu-cs!hangtime Muncie, IN 47306-0655 {BackBones}!iuvax!bsu-cs!bsu-ucs!00csgunn