Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!uoft02!desire!w0033
From: w0033@desire.wright.edu
Newsgroups: comp.sys.apple
Subject: Re: ACER
Message-ID: <273.25b79d07@desire.wright.edu>
Date: 20 Jan 90 04:04:39 GMT
Lines: 26

>       From what I've been told, it's a very simple process.. Instead of
>having 8 bits represent from 0 to 255, they use 3 bits to represent
>0 to 8..
>
>       So 0 to 31 in the original system would be 000 in the new system,
>32-64 would be 001, etc...  With most sounds, after it's decompressed
>you can't tell much difference as even 8 graduations is usually enough.
>I think some programs actually unpack ACER sounds on the fly... It's a
>built in GS/OS (or ProDOS 16 or something) function I think.. BUT I'M NOT
SURE.

I don't think this is quite the way it works. The way I understand it is that
the ACE toolset simply stores the difference from one byte to the next in a
digitized sound file. Usually with a fairly high sample rate, this difference
should be pretty small since the data is supposed to be following some analog
source. The result is that these differences from one point to the next can be
represented by only 4 or even 3 bits. Another interesting effect is that if
there is a spike in the sound wave, it will be chopped since the model can only
represent small changes from one point to the next. The listening effect is
that some sound files which were incomprehensible before become clear and
useable. I have an article on this stuff from a users group I can upload if
anyone wants to know more. ACER Unpacker V1.0 is the only program I know of
currently that allows you to apply these tools to sounds. Many programs though
take advantage of this method already (Quadronome, Orbitzone etc.)
By the way, it is also available on the Apple Dayton BBS (513) 434-0077 24 hrs.
a day 2400 BAUD, in the GS Sound section.