Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!rutgers!ames!ucbcad!ucbvax!INGRES.BERKELEY.EDU!hatcher
From: hatcher@INGRES.BERKELEY.EDU (Doug Merritt)
Newsgroups: comp.sys.amiga
Subject: DMCS..Not exactly standard.
Message-ID: <8706231113.AA11207@ingres.Berkeley.EDU>
Date: Tue, 23-Jun-87 07:13:26 EDT
Article-I.D.: ingres.8706231113.AA11207
Posted: Tue Jun 23 07:13:26 1987
Date-Received: Thu, 25-Jun-87 00:37:22 EDT
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: University of California, Berkeley
Lines: 65
Summary: Not exactly right; here's the scoop

In article <769@gryphon.CTS.COM> jdm@pnet02.CTS.COM (John Mesiavech) writes:
>One of the main reasons that the PS SMUSplay and other files can't do their
>thing with DMCS files is simple.  DMCS uses a variant of SMUS, with instrument
>locations at the END of the file, rather than at the beginning as required by
>the SMUS format.

This is not quite right; DMCS does *not* use a "variant" of SMUS. It has
its very own IFF type. There are lots of kinds of IFF standards, all of
which fit one broad convention to allow any IFF file to be read (if not
understood). The layout of the DMCS IFF files looks somewhat like the
SMUS IFF files, simply because they're both storing similar information.
In particular, the chunk names are all different, so even if the DMCS
"UIID" instrument chunk were moved to the top of the file, it wouldn't
help, because SMUS has instrument chunks named "INS1". A second example
is the melody itself, stored in "UBR1" chunks in DMCS and in "TRAK" chunks
in SMUS. None of the other chunks have the same names, either.

Also DMCS has a machine identifier "AMGA" (for Amiga) near the top of the
file; although unrecognized chunks will (should) be ignored by IFF
readers, that does mean there's going to be *some* difference between
DMCS files on the Amiga and on the Mac. Don't know if it implies actual 
incompatibilities or not (that chunk has no contents, it's just a flag).

>Now, DMCS CAN write out SMUS files..It's just that many
>people with DMCS don't bother to use the special menu selection "Save as SMUS
>file" when savig their work in DMCS.  A DMCS -> SMUS file converter would not
>be that hard to rite, and from what I've heard, some people are working on
>them.

All very true.

>It's rather interesting that EA, who *created* the IFF standard in the first
>place, would abandon it for their music program.

Not exactly. The IFF standard allows for developers to create their
own private unpublished chunk types within IFF files, and EA did exactly
that with the DMCS format. Although the precise contents are not public
information, it is still an IFF file and each chunk can be found,
read, and printed by a program like "iffdump", even though they haven't
told us what the contents of each chunk *means*. This is certainly a
regrettable way to do things, but is specifically allowed by the IFF
standard.

And it is preferable to having private *non-IFF* files...you
can never be sure what the h*ll a file is in the first place with
random binary. At least with IFF, there's always a chunk name to go
with the file; better than no information at all! Furthermore there are
typically a series of chunks, each of which is recognizable as a chunk,
and has a name to go with it, so there's a certain amount of structural
information available in any IFF file, even when it is a custom, unpublished
type. Which is how I managed to figure out all this stuff about DMCS
in the first place.

FYI there are other IFF types that are unpublished, including Deluxe
Video, Deluxe Print, and Flow outlines, and doubtless others, too.
The original IFF specs *named* maybe 15 or 20 types but only *documented*
the contents of a half dozen or so.

Sonix uses the published IFF SMUS type, but they do add an unpublished
SNX1 chunk buried in it, meaning that the piece can be played by
a public domain SMUS player, but there is some unknown information being
skipped in that one SNX1 chunk (I have no idea whether anything important
is stuck in there or not).
	Doug Merritt		ucbvax!ingres!hatcher (thru Jun 28)
			or	ucbvax!unisoft!certes!doug