Path: utzoo!attcan!uunet!snorkelwacker!mit-eddie!bu.edu!orc!oliveb!tymix!tardis!jms
From: jms@tardis.Tymnet.COM (Joe Smith)
Newsgroups: comp.sys.amiga.tech
Subject: Re: Sample Compression methods
Summary: try vertical compression
Message-ID: <1002@tardis.Tymnet.COM>
Date: 12 Apr 90 06:30:02 GMT
References: <1452@mindlink.UUCP>
Reply-To: jms@tardis.Tymnet.COM (Joe Smith)
Organization: BT Tymnet, San Jose, CA
Lines: 31

In article <1452@mindlink.UUCP> a186@mindlink.UUCP (Harvey Taylor) writes:
> various compression methods on sound samples. 
> arc=20%, lharc=24%, zip=18%
>    Is anybody aware of a better algorithm?

I overheard someone at BADGE talking about vertical compression.
Sound samples don't usually have repeated bytes.  Even steady tones don't
have repeated patterns of bytes (due to jitter in the LSB).  However,
there are usually a lot of consecutive negative bytes and consecutive
positive bytes.  For example:

	FC 11111100 -4
	FD 11111101 -3
	FE 11111110 -2
	FF 11111111 -1
	00 00000000  0
	01 00000001  1
	02 00000010  2
	03 00000011  3

Looking at the bits column by column instead of row by row, we have:
(previous 8 bytes)	FF FF FF FF FF F0 33 55
(the 8 bytes above)	F0 F0 F0 F0 F0 F0 33 55
(next 8 bytes)		00 00 00 00 00 F0 33 55
That is a pattern that should compress well with LZW.

-- 
Joe Smith (408)922-6220 | SMTP: jms@tardis.tymnet.com or jms@gemini.tymnet.com
BT Tymnet Tech Services | UUCP: ...!{ames,pyramid}!oliveb!tymix!tardis!jms
PO Box 49019, MS-C41    | BIX: smithjoe | 12 PDP-10s still running! "POPJ P,"
San Jose, CA 95161-9019 | humorous dislaimer: "My Amiga speaks for me."