[comp.sys.amiga] Frame Buffers

SANDNEGD@UREGINA1.BITNET (01/26/88)

Does anyone have any information about the ReaSyn frame buffer board
from Mimetics?  I would like some professional-level color flexibility
in my Amiga 1000, but I have an annoying suspicion that this board
is intended for 2000's only, in the video slot.

If this is the case, how about sources for frame buffer peripherals in
general that might be interface-able?  I don't need digitizing capability,
just display.  I am almost on the verge of designing one myself by
plumbing into the video connector and talking to the memory through the
expansion port, but methinks therein madness lies.

                                             Glen D. Sandness

king@dretor.DRETOR.UUCP (Stephen King) (06/22/88)

I would like to ask about people's ideas for an auxiliary frame buffer for
the Amiga. I have a few ideas of my own which I would like to incorporate
into a hacker project, but I would like to know what other people think
before I start designing the thing. Here is a short list of capabilities I
would like to endow upon my creation:
1) Will work as FAST expansion RAM.
2) 15 bits per pixel, 5 each RGB and one ignored on 16 bit bus.
3) Will accept digital RGB input - can grab frames.
4) tied to Amiga timing - genlockable with appropriate adaptor.
5) will de-interlace with double-buffering.
6) will work as scan doubler for flicker-free 3D applications.

Comments? Criticisms?
Mail me and I will post a summary.

...{utzoo|mnetor}!dciem!dretor!king		Stephen J King

R38@psuvm.psu.edu (Marc Rifkin) (02/22/90)

I am working in an Amiga graphics visualization lab at Penn State, and
am looking for information from people who have had experience with-
Amiga Frame Buffers, etc... and who know about the V-Machine, Video
Toaster, Video Transputer or similar device.  We need hi-resolution,
hi-color, hi-quality graphics, preferably compatible with Sculpt-4d,
and offering at least some image processing or painting support.
We are additionally seeking any new 3d software that may employ
pseudo-ray tracing (ie- fast), textures, (We have many other wishes,
anyone willing to listen?)

What we can offer is a testing ground for any such above hard/software
(DEVELOPERS are you listening??)  We are already putting existing
technology to the test. How about something new?

Marc Rifkin r38@psuvm
 Appt. 302 626 E. College Ave
 State College, Pa. 16801  (814) 867-4837

JKT100@psuvm.psu.edu (JKT) (02/23/90)

Marc Rifkin <R38@psuvm.psu.edu> says:
>
>I am working in an Amiga graphics visualization lab at Penn State...

I'd like to add to this:  Marc and I are in a group which is working
on a large computer animated video project, the results of which will
be widely seen - Nationally if not internationally.  We're using
Amigas for nearly everything.

>What we can offer is a testing ground for any such above hard/software
>(DEVELOPERS are you listening??)  We are already putting existing
>technology to the test. How about something new?

This is the chance for developers to get recognition!!  If you have
something still only in TESTING, this is the chance to prove its
capabilities!!  Essentially, we've already reached the limits of the
available technology (both hardware and software) and need more.

>
>Marc Rifkin r38@psuvm
> Appt. 302 626 E. College Ave
> State College, Pa. 16801  (814) 867-4837
                                                     Kurt
--
 ----------------------------------------------------------------------
|| Kurt Tappe   (814) 862-8630 ||  "This town needs an enema."        ||
|| 600 E. Pollock Rd., #5705   ||                   - Joker, "Batman" ||
|| State College, PA 16801      --------------------------------------||
||   jkt100@psuvm.bitnet  or  jkt100@psuvm.psu.edu                    ||
||        or  jkt100%psuvm.bitnet@psuvax1           QLink: KurtTappe  ||
 ----------------------------------------------------------------------

seanc@pro-party.cts.com (Sean Cunningham) (02/24/90)

In-Reply-To: message from R38@psuvm.psu.edu

 
I have experience on the Mimetics FrameBuffer w/FrameCapture...it's output is
pretty good, though it doesn't have RGB out, only a couple of BNC connectors. 
It's got 764 x 484 resolution (depending on your application), and there are
drivers to allow Sculpt/Animate 4D to render to it.  It gives you 2.1M colors
onscreen out of a 16.7M color pallette.  I've tested out Sculpt and Turbo
Silver files on it, and apart from some aspect ratio tweaking that needed to
be done, they looked excellent.
 
There's some new rendering package that's supposed to offer ray-tracing
realism, but is supposed to be alot faster.  It's called RENDITION.  It's not
available to the public yet, but it sounds nice.  Another one to look into is
PageRender 3D...
 
hope this helped alittle...
 
Sean

tbz@sactoh0.UUCP (Terry B. Zweigenbaum) (02/24/90)

I have been using the Mimetics Framebuffer with Turbo Silver and
Sculpt4D.  It seems that 24 bit Sculpt4D images still have bands
while 24 bit Turbo Silver images are nicely dithered.  Is there any
way to get rid of these bands?
 
--Bob   ..!pacbell!sactoh0!tbz

-- 
#############################################################
#         PRIVATE             #  SAC-UNIX, Sacramento, Ca.  #  
#         PARKING             #  UUCP=...pacbell!sactoh0    #
#############################################################

seanc@pro-party.cts.com (Sean Cunningham) (02/28/90)

In-Reply-To: message from tbz@sactoh0.UUCP

I noticed this as well!
 
It appears that Sculpt/Animate ignores all dithering and antialiasing
settings when you use 24-bit mode...I don't know if they had a purpose for
this, or if it's a bug...lack of foresight, or what.
 
A friend of mine sent a *BIG* file away to get rendered on a hires
filmrecorder...this sucker was 1280 X 1500 X 24-bits!  It looked awsome, but
you could still see (very) slight jaggies.
 
Have you seen output from Caligari-Broadcast?
 
Sean

bobl@pro-graphics.cts.com (Bob Lindabury) (02/28/90)

In-Reply-To: message from tbz@sactoh0.UUCP

> I have been using the Mimetics Framebuffer with Turbo Silver and
> Sculpt4D.  It seems that 24 bit Sculpt4D images still have bands
> while 24 bit Turbo Silver images are nicely dithered.  Is there any
> way to get rid of these bands?
> 
> --Bob   ..!pacbell!sactoh0!tbz

There is no way to get rid of the bands unless you can coherse Byte By Byte
into doing some more work on their Amiga product line rather than jumping over
to the Mac product line entirely.

Seems the boys at Byte by Byte could care less about us Amiga users because
they are going after the Mac carrot (or is that Apple?).  At any rate, it
seems that not much forethought was put into the Sculpt product and they opted
to optimize the redering routines for the Amiga's display modes and didn't
even consider 24 bit rendering to any large degree.  Seems the software will
produce these bands so that another routine can dither between them to get
those nice HAM images.  Too bad that the bands are not skipped when rendering
in 24 bit mode.  So, an otherwise perfectly good product is "rendered" next to
useless for 24 bit images.  Yeah, doesn't everyone want to drop $500 for a 3d
ray tracer that doesn't do 24 bit well?  How about we all drop our bucks on
the product so that Byte by Byte can do more Mac developement?  These ideas
dosn't sound good to you?  Hmm...

I'm still waiting for a "good" 3D program with an interface like Caligary,
Editing easy like Sculpt and rendering/mapping capabilities like Turbo Silver.

If all the best of each of these programs were incorporated into a new 3d
program that also supported the Renderman Spec, I would (as others would) jump
on it in a second even if it did cost in the $1000-$2000 range.

You mention nice dithering in Turbo for 24 bit images.  I'm not sure what you
mean by that but with a 24 bit image (over 16 million colors available) you do
not need to do any dithering.

-- Bob
_______________________ Pro-Graphics BBS  201/469-0049 ________________________
                                             
InterNet: bobl@pro-graphics.cts.com          |       ProLine: bobl@pro-graphics
    UUCP: ..crash!pro-graphics!bobl          |        CServe: 70347,2344
ARPA/DDN: ..crash!pro-graphics!bobl@nosc.mil |  Amer. Online: Graphics3D
___________                                                        ____________
            Raven Enterprises - 25 Raven Ave. Piscataway, NJ 08854

wayneck@tekig5.PEN.TEK.COM (Wayne C Knapp) (03/01/90)

In article <1689@crash.cts.com>, bobl@pro-graphics.cts.com (Bob Lindabury) writes:
> Seems the boys at Byte by Byte could care less about us Amiga users because
> they are going after the Mac carrot (or is that Apple?).  At any rate, it
> seems that not much forethought was put into the Sculpt product and they opted
> to optimize the redering routines for the Amiga's display modes and didn't
> even consider 24 bit rendering to any large degree.  Seems the software will
> produce these bands so that another routine can dither between them to get
> those nice HAM images.  Too bad that the bands are not skipped when rendering
> in 24 bit mode.  So, an otherwise perfectly good product is "rendered" next to
> useless for 24 bit images.  Yeah, doesn't everyone want to drop $500 for a 3d
> ray tracer that doesn't do 24 bit well?  How about we all drop our bucks on
> the product so that Byte by Byte can do more Mac developement?  These ideas
> dosn't sound good to you?  Hmm...

The problem isn't rendering to 24 bits.  It sounds like that works just fine.
Once you have 24 bits per pixel rendered and displayed it isn't possible to 
dither anymore.  What data can be used to control the dithering?  One could
take the 24 bit image and dither it to say an 18 bit images (6 bits per red,
green, blue).  This would reduce the number of colors in the image and reduce
the resolution to a lesser degree.  It would be similar to running the image
though a low-pass filter.  Anyway the picture would become kind of HAMish, that
is it would appear softer.  If you want to dither to 24 bits, you have render
something like 30 bits per pixel or 36 bits per pixel would be even better.
Then you have some data to dither on. 

> I'm still waiting for a "good" 3D program with an interface like Caligary,
> Editing easy like Sculpt and rendering/mapping capabilities like Turbo Silver.
> If all the best of each of these programs were incorporated into a new 3d
> program that also supported the Renderman Spec, I would (as others would) jump
> on it in a second even if it did cost in the $1000-$2000 range.

I think you realize what Renderman is all about.  Anyway if you what to use it
be prepared to pay a lot.  The main feature the Pixar uses is called 'shaders'.
Only there is a problem --- 'shaders' are not a part of the Renderman Spec..
This means if you want those Pixar quality pictures you better buy Pixar's
Renderman Render.  Also shaders are pieces of code so if you aren't hot at
physics and programming you end up paying a lot to get them written.  (Pixar
has to make money after all.)  However, the end product can be amazing.  The
sharders can turn a run of the mill image into a near photo-realistic image.
What will this cost?  Well the Pixar Renderman render sells for $4650, plus
you need a Mac II or PC with a transputer board if you want images in less 
than an hour.  You are looking a $10,000 to $15,000.  Still hot on Renderman?

Well if you want a true Amiga solution and $1000 doesn't scare you off, give
Martin Hash a call.  Ask him about Animation:Jorneyman. (206) 573-9427. 
  
> You mention nice dithering in Turbo for 24 bit images.  I'm not sure what you
> mean by that but with a 24 bit image (over 16 million colors available) you do
> not need to do any dithering.

This is simply not true.  It is very possible for 24 bit images to have 
banding.  Isn`t that what you were complaining about at the biginning of 
this article.
                                               Wayne Knapp

bobl@pro-graphics.cts.com (Bob Lindabury) (03/03/90)

In-Reply-To: message from seanc@pro-party.cts.com

> It appears that Sculpt/Animate ignores all dithering and antialiasing
> settings when you use 24-bit mode...I don't know if they had a purpose for
> this, or if it's a bug...lack of foresight, or what.
> 
> Sean

Sculpt/Animate ignores all dithering routines as well it should when rendering
a 24bit image.  After all, with over 16million colors available, you don't
need to dither to get a gradiation.  However, they optimize the ray-trace for
Ham previous to this or something resulting in a 16 color base image which
does not translate well to 24 bit at all.   Basically Sculpt/Animate is a
waste of money if you want to render to 24 bit IMHO.  I can't live with the
bands.

-- Bob
_______________________ Pro-Graphics BBS  201/469-0049 ________________________
                                             
InterNet: bobl@pro-graphics.cts.com          |       ProLine: bobl@pro-graphics
    UUCP: ..crash!pro-graphics!bobl          |        CServe: 70347,2344
ARPA/DDN: ..crash!pro-graphics!bobl@nosc.mil |  Amer. Online: Graphics3D
___________                                                        ____________
            Raven Enterprises - 25 Raven Ave. Piscataway, NJ 08854

bobl@pro-graphics.cts.com (Bob Lindabury) (03/06/90)

In-Reply-To: message from wayneck@tekig5.PEN.TEK.COM

> In article <1689@crash.cts.com>, bobl@pro-graphics.cts.com (Bob Lindabury) writes:
>> Seems the boys at Byte by Byte could care less about us Amiga users because
>> they are going after the Mac carrot (or is that Apple?).  At any rate, it
>> seems that not much forethought was put into the Sculpt product and they opted
>> to optimize the redering routines for the Amiga's display modes and didn't
>> even consider 24 bit rendering to any large degree.  Seems the software will
>> produce these bands so that another routine can dither between them to get
>> those nice HAM images.  Too bad that the bands are not skipped when rendering
>> in 24 bit mode.  So, an otherwise perfectly good product is "rendered" next to
>> useless for 24 bit images.  Yeah, doesn't everyone want to drop $500 for a 3d
>> ray tracer that doesn't do 24 bit well?  How about we all drop our bucks on
>> the product so that Byte by Byte can do more Mac developement?  These ideas
>> dosn't sound good to you?  Hmm...
>
> The problem isn't rendering to 24 bits.  It sounds like that works just fine.
> Once you have 24 bits per pixel rendered and displayed it isn't possible to
> dither anymore.  What data can be used to control the dithering?  One could
> take the 24 bit image and dither it to say an 18 bit images (6 bits per red,
> green, blue).  This would reduce the number of colors in the image and reduce
> the resolution to a lesser degree.  It would be similar to running the image
> though a low-pass filter.  Anyway the picture would become kind of HAMish, that
> is it would appear softer.  If you want to dither to 24 bits, you have render
> something like 30 bits per pixel or 36 bits per pixel would be even better.
> Then you have some data to dither on.

Wayne, it seems you've missed my point.  If you (or Byte by Byte) were
rendering TRUE 24 bit images, they would have no need to do ANY dithering at
all!  I don't want my 24 bit images dithered!!  Sculpt does NOT render 24
bits well and that is the whole point.  They go through a PRE-dithering
sequence even when doing 24 bit images.  This in not require, needed or even
wanted.  It is unacceptable.

>> I'm still waiting for a "good" 3D program with an interface like Caligary,
>> Editing easy like Sculpt and rendering/mapping capabilities like Turbo Silver.
>> If all the best of each of these programs were incorporated into a new 3d
>> program that also supported the Renderman Spec, I would (as others would) jump
>> on it in a second even if it did cost in the $1000-$2000 range.
> 
> I think you realize what Renderman is all about.  Anyway if you what to use it
> be prepared to pay a lot.  The main feature the Pixar uses is called 'shaders'.
> Only there is a problem --- 'shaders' are not a part of the Renderman Spec..
> This means if you want those Pixar quality pictures you better buy Pixar's
> Renderman Render.  Also shaders are pieces of code so if you aren't hot at
> physics and programming you end up paying a lot to get them written.  (Pixar
> has to make money after all.)  However, the end product can be amazing.  The
> sharders can turn a run of the mill image into a near photo-realistic image.
> What will this cost?  Well the Pixar Renderman render sells for $4650, plus
> you need a Mac II or PC with a transputer board if you want images in less
> than an hour.  You are looking a $10,000 to $15,000.  Still hot on Renderman?

You miss the point here again Wayne.  Obviously I don't want to do Renderman
rendering on my Amiga...that is not the purpose for which I intended the
Renderman spec be used.  What I WOULD expect is that the higher priced 3D
programs support the output of Renderman definition files (or whatever they
call them) so that the final output could be rendered on a Pixar engine or
some other machine that is JUST a renderer.  Hey, who expects to render these
type of things on the Amiga anyway?  You?  Certainly not me!  If I had my
druthers, I would like to do quick pencil test type rendering on the Amiga to
make sure that the movement and such were fine..I would also probably render
at least a few key frames to make sure lighting was good.  This could be done
in either Ham or maybe some other appropriate mode on the Amiga (preferably 8
bit if we ever get an 8bit card) that would be relatively quick and still give
us a good indication of the "look" of our scenes.  From there it would be
rendered frame by frame to either video or film on a Pixar or some other such
rendering engine.  Of course this is going to cost some money.  Those willint
to pay will do it because most of the cost is in the creation..not the
rendering.

>> You mention nice dithering in Turbo for 24 bit images.  I'm not sure what you
>> mean by that but with a 24 bit image (over 16 million colors available) you do
>> not need to do any dithering.
> 
> This is simply not true.  It is very possible for 24 bit images to have
> banding.  Isn`t that what you were complaining about at the biginning of
> this article.
> Wayne Knapp

Again, you are not reading my message correctly.  Refer to my first paragraph
on the problems with dithering and 24 bit images.  We are NOT coming from 32
bit images.  We are generating 24 bit images from scratch..dithering shouldn't
be done as can be evidenced by the horrible output from the Sculpt program.

-- Bob
_______________________ Pro-Graphics BBS  201/469-0049 ________________________
                                             
InterNet: bobl@pro-graphics.cts.com          |       ProLine: bobl@pro-graphics
    UUCP: ..crash!pro-graphics!bobl          |        CServe: 70347,2344
ARPA/DDN: ..crash!pro-graphics!bobl@nosc.mil |  Amer. Online: Graphics3D
___________                                                        ____________
            Raven Enterprises - 25 Raven Ave. Piscataway, NJ 08854