RE: Cinematronics/Sega universal sound board...

From: Ozdemir, Steve <steve.ozdemir_at_plpt.com>
Date: Thu Apr 23 1998 - 17:31:57 EDT

G'day Clay,

At first glance this does seem to be an elegant solution to the simpler
sounds. I question how well it would handle continuous shifting
background sounds like in Star Castle. But hey, if the solution even
handles 2/3's of Cinematronics sounds that alot closer than we are now
to a universal sound board!

               Steven S Ozdemir
               sso@plpt.com (my company renamed itself in Feb)
               sso@dsc.com (good for a few more months)
               ozdemir@xenon.stanford.edu (permanent...weekly)

> ----------
> From: Clay Cowgill[SMTP:ClayC@diamondmm.com]
> Sent: Thursday, April 23, 1998 1:41 PM
> To: 'vectorlist@spies.com'
> Cc: Clay Cowgill
> Subject: Cinematronics/Sega universal sound board...
>
> I had a revelation yesterday...
>
> There's an off-the-shelf part that would fit 99% of our requirements
> for
> making a universal sound board for Cinematronics and Sega G80
> multigames.
>
> It's already been used on tons of arcade game boards-- the OKI 6295.
>
> Very neat little gizmo. It's an ADPCM playback device that supports
> up
> to 127 "phrases" stored in an external ROM. 4 bit ADPCM with rates up
> to 32KHz. Just over 16 seconds of playback at that speed from an
> external 27C020. It's about what we're "wishing" for-- four channels,
> low parts count (ROM + 6295 + little micro to bridge interfaces),
> compression, automatic playback...
>
> So here's the cool part. You write a two 8 bit words to the device.
> This tells it what "phrase" to play, on what channel, and how loud (16
> steps). You can have up to 127 phrases per ROM, and can bank-select
> ROMs for more. (All sounds played simultaneously would have to be in
> the same bank.)
>
> The first 0x3ff bytes of the ROM are a phrase table (8 bytes per
> phrase). If you tell the chip to play "Phrase 3" it goes to address
> 0x010 and grabs 8 bytes-- the first three are the start address of the
> sound, the next three are the ending address, the other two bytes do
> something I don't recall off the top of my head. Then you're done.
> The
> chip does the rest-- fetches samples, decodes, plays the sound.
>
> For multivoice polyphony you just send it more "Phrase" addresses
> assigned to different channels.
>
> If you wanted more than 4 voices, just use 2 of 'em.
>
> We'd just put a little uController (like a 2051 or Z8 or Pic or
> whatever) in front of the 6295 to receive the sound commands from the
> game hardware and have it do the decode to trigger playback of the
> samples. Pretty clean solution.
>
> Some trap code with the emulators should be able to give us an
> un-biased
> assessment of how many simultaneous sounds are active at any given
> time.
> Long-term repeating sounds are possible (but not optimal for the
> technique), but the rest should be a walk-in-the-park.
>
> Comments?
>
> -Clay
>
> Clayton N. Cowgill Engineering Manager
> -------------------------------------------------
> /\ Diamond Multimedia System, Inc.
> \/ Communications Division
>
Received on Thu Apr 23 14:30:53 1998

This archive was generated by hypermail 2.1.8 : Thu Jul 31 2003 - 23:00:44 EDT