NetNews Usenet Archive 1992 #30/NN_1992_30.iso/spool/comp/sys/atari/st/18591

Text File | 1992-12-20 | 5.9 KB | 128 lines

Newsgroups: comp.sys.atari.st
Path: sparky!uunet!elroy.jpl.nasa.gov!hanauma.jpl.nasa.gov!hyc
From: [e-mail] (Howard Chu)
Subject: Re: Falcon030, resolutions, TRUE color, and DSP
Message-ID: <[e-mail]>
Sender: [e-mail] (Usenet)
Nntp-Posting-Host: hanauma.jpl.nasa.gov
Organization: SAR Systems Development & Processing, JPL
References: <[e-mail]>
Date: Fri, 18 Dec 1992 07:42:13 GMT
Lines: 115

In article <[e-mail]> [e-mail] (Len Stys) writes:
>Hi, everyone.

Howdy, all
>
>I would like to get a better grasp at HOW good the Falcon030 could
>be as a game system.
>
>First of all, I'm told that 640x480 TRUE COLOR would require 4 Megs
>of memory for NICE games.  I'm told that 2 Megs would not really be
>enough since there must be room for other stuff like sound.

At 600K per screen, figure 1.2meg for double-buffered graphics, yeah, it
starts eating up memory pretty fast...
>
>Secondly, no one seems to know if this resolution (interlaced of course)
>could be displayed on TV without flicker.  Does anyone know?

Just what exactly are you asking here? Of *course* it will flicker, that is
the nature of interlaced video. Will you notice it? Depends on the TV, and
on what's being displayed. A single-pixel horizontal line will flicker
terribly, but that's true regardless of whether the video is computer-generated
or just happened to appear in some TV or film production. The desktop has
a lot of flicker, because it's covered with icons and windows that all have
single-pixel horizontal lines for borders. A game with moving images,
background scenery and such probably won't have very noticable flicker
(unless it's got some pretty boxy-looking graphics, eh?)
>
>Thirdly, I have talked with someone that said that 640x200 (I'm not sure
>if you can get true color in this resolution) would also produce some
>pretty good games.  I guess that could be pushed to 640x240, but
>anyways, I'm told that 300k for ONE screen.  And you'll need at least
>TWO screens.  This would allow this game to produce with only 2 Megs
>easily--MAYBE even 1 Meg.

True color is available in all resolutions on a TV, up to 768x480.
>
>I'm NOT sure if that would flicker or not either.  If someone knows,
>please tell me.

The 200 (or 240) vertical-resolution modes are non-interlaced, and therefore
will not flicker at all.
>
>I have been also informed that the Blitter REALLY would not be fast
>enough to move graphics around.  And someone suggested that Atari should
>have used the TI30410.  My question is, would the DSP work just as well?
>The Atari Advantage article states these things:

The DSP can only communicate with the main memory at up to 1 megabyte/second.
This is nowhere near the speed required to pass as a blitter. The items
mentioned in the Atari Advantage article are aimed purely at computation -
the DSP can be used to process the geometric formulas required to define
shapes and 3-D objects, rotate them, shade them, etc., but all that computation
can only be done in the DSP memory, and you would then have to transfer
the results to main memory before you could do anything like draw them onscreen.
>
>Will the DSP be able to produce CD quality sound AND generate graphics
>in a game?

The DSP is not needed for sound generation. For simple stereo sample playback,
only the on-board CODECs need to be involved. So, yes, the DSP could do number
crunching while CD-quality sound is playing back. However, I would expect that
most games would have done their serious graphics number crunching in advance,
and would simply do frame playback during an actual game.
>
>What is the 32Kx24 bit (96K bytes) of local zero wait state SRAM for?
>What does the DSP do with it?

What *doesn't* the DSP do with it? The DSP is not a tightly coupled
coprocessor, like a 68882 FPU is coupled to the 68030. It is a totally
independent computer, and that 96K of memory is its entire workspace.
(Well almost; there's some on-chip ROM as well.) When you want the DSP
to do something for you, you download a DSP program from ST memory to
DSP memory, then you download the data to be operated on, and then you
upload the results back from DSP memory into ST memory. All of these
memory transfers are initiated by the 68030 and run at 1 MB/sec.
>
>Is there any chance that Atari is only including 1 Meg in the Falcon030
>and expecting the 640x480 TRUE color mode to be used as a standard because
>of a video technique which will allow them to do so?

I don't think Atari really expects many true-color games to run in
640x480 resolution on 1 megabyte machines. I think they really expect to
see a lot of games support the 256 color modes, but ya never know fer sure.
>
>How was the Tina Turner video able to run at 4 frames per second
>in TRUE COLOR and what resolution was that?

Dunno, never saw it.
>
>Is a Motorola 68030 16 MHz powerful enough to be used as a game system?
>As a game system that can compete with future 32-bit game systems?

The 68030 at 16 MHz is certainly more powerful than, say, the 68000 or 65816
found in most home systems right now. It is a full-fledged 32 bit processor,
as well. The 65816 is only 16 bit, and we've already had debates over what to
call the 68000...
>
>And lastly, the Jaguar is rumored to run off of a RISC 64-bit chip.
>We don't know what type of chip this is.
>
>Could the Jaguar POSSIBLY be a Motorola 68030 at 16 MHz and have a
>RISC DSP?

"run off" is a misleading phrase. "RISC DSP" is almost redundant. I'd bet
that the Jaguar (which I've never seen, and heard only a very little about)
is 68030 based with a 64 bit graphics coprocessor. I would also bet that
this 64-bit chip is called a RISC chip merely because of the tiny size of
its instruction set, an instruction set devoted entirely to describing a
graphics display, (just like the 8-bit ANTIC) and having very little other
functionality.
--
  -- Howard Chu @ Jet Propulsion Laboratory, Pasadena, CA

All true wisdom is conveyed in one-line witticisms.