Light ROM 1/LIGHT-ROM 1 (Amiga Library Services)(1994).iso/imagine/text/archive.05

Text File | 1994-10-19 | 176.0 KB | 4,189 lines

IMAGINE archive: collected off of [e-mail]

ARCHIVE V
Apr 18 '91 - May 9 '91

If you have questions or problems with this file, email Marvin Landis
at [e-mail]

Many thanks to Doug Dyer ([e-mail]) for the previous
versions of this archive

note: each message seperated by a '##'

&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&

Date: Thu, 18 Apr 91 00:26:51 -0400
From: [e-mail]

Hello there, this is my first post, here I go! Someon e was asking about
testing Imamgine out on an '040-based Amiga. Well, we will be doing just
that in about a week's time. We will be using RCS Management's 040
board for the Amiga 2000 and will be testing all of our software with
it (we have lots--> Imagine, Lightwave, Sculpt, TAD, Dpaint etc...) I
am very intersting to see how it will compare to our Amiga 3000. I
will definetly post the results for Imagine.fp here...

                            Stef.


##

Subject: AmigaWhirld 3D Comparison
Date: Thu, 18 Apr 91 06:15:27 EDT
From: [e-mail]

I was amused to note that the Lightwave 3D and Imagine pictures in the
scene-test comparison were accidentally exchanged. Whoops.

Also, the major blue cast in Animation Journeyman is a major bug to one of
the early versions (I think it didn't save the green field in 24-bit
renders.) This has since been fixed.

Sorry, I don't have the actual scene description. Write to the author
care of AmigaWhirld.

-Steve

---------------------------------------------------------------------------
Steve Worley                                        [e-mail]
---------------------------------------------------------------------------


##

Subject: World Size
Date: Thu, 18 Apr 91 06:00:58 EST
From: John J. Rosner <[e-mail]>

2 Questions about world size:

Giving 8000 x 8000 x 8000 in the Globals size bar work great, but how do I
know what the default size is?  It is not configured and the size bar starts
out 0  Knowing this would take some of the guesswork out of scaling up
objects.

Is the limit set at 8000?  I thought the Imagine doc said the world size was
infinite (obviously one runs into "word" size problems first).  I set the size
to 10000 and it barfed, in a polite kind of way.

Thanks in advance,

John Rosner


##

Subject: '040 Comparisons
Date: Thu, 18 Apr 91 09:36:41 EDT
From: Mark Thompson <[e-mail]>

> testing an '040-based Amiga with
> Imagine, Lightwave, Sculpt, TAD, Dpaint etc...)
> I will definetly post the results for Imagine.fp here...

Please post the Lightwave comparisons also, I would really like to see them.
What is RCS Management's 040 board going for (street price)? Is it currently
available or is this a beta unit? Oh, and how much/what kind of RAM does it
use.
|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
|      `       '        Mark Thompson                 CONCURRENT COMPUTER  |
| --==* RADIANT *==--   [e-mail]        Principal Graphics   |
|      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   |
|     Productions       [phone] [phone]   & General Nuisance   |
|                                                                          |
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

##

Subject: Buddy System & Object Disks...
Date: 18 Apr 91 10:24:00 EST
From: "SYSTEM MANAGER" <[e-mail]>

I am a bit surprised there has been nary a comment about the semi-review
I wrote about the Buddy System for Imagine, nor has there been a comment
about the object disk idea.

Did my posting get out???  If yes, then I guess my posting did not
strike the fancy of the 3ders out here.   If that is the case, ok
then, I will work on a better posting!  :-)

I will resend the posting if it did not get out!

-mark=
[e-mail]


##

Subject: Re:  The "State of 3-D Art" in Amiga World
Date: Thu, 18 Apr 91 11:25:12 EDT
From: [e-mail] (Brian Reynolds)

This is a follow up (and correction) to my previous post.

As Glenn Lewis pointed out I meant to say I am using Rayshade
V4.0(beta) to do my tracing.  I use rayscene to create scene
descriptions and script files for animation.

(As I was writing this I read a reply from Steve Worley mentioning that
the Imagine (captioned as upper left) and LightWave 3D (captioned as
lower left) images were accidently exchanged, so some of the following
may make no sense at all.  Steve, how did you know the images were
switched?)

For those of you who have seen the article, what do you think of the
four test images?  I thought that the blue light was too powerful in
the upper left image.  It completely over powered the green coloring of
the glass. (If this is the Lightwave image, it makes sense.  You can't
do refraction mapping they way you can do reflection mapping.)  The
reflection in the chrome ball also looked a little strange.  I would
have expected more curvature in the reflected image.  (Another
consequence of not ray tracing.)  The image in the upper right (3D
Professional I think) looked much more like what I expected.  The glass
was green, but you could see the blue light on it.  (Maybe both images
are wrong.  What color do green and blue yield, yellow?)  The
reflections in the chrome ball curved up towards the edges.  The lower
left image was also nice.  Without using raytracing they did a lot of
the tricks used by Renderman (and others) to simulate reflections and
shadows.  (Of course if this is the Imagine image, they cheated on the
reflection map by using raytracing.  :)  The article mentions that
getting shadows takes a lot of time.  Is there any reason shadows
should take more time than reflections?  To get the reflections he
rendered a reflection map for the chrome ball.  To get shadows I would
think that you would render a shadow map.  I guess if you have to do
one for each light source (there were three in the scene) it could take
some time.  I took the RenderMan course at SIGGRAPH last year, but it
has been a while and I don't quite remember how shadow mapping works.
I won't talk about the Animation: Journeyman image since Steve mentions
that its bad showing was due to a bug that has been fixed.  Personally
I like the idea behind Animation:  Journeyman.  Splne based modelling
makes more sense (to me) than polygon based modelling.  Does anybody
think they will add raytracing splines (or NURBS, I don't really know
what they are using) to their product?  Will the competition do it?

Brian Reynolds "... a drone from sector 7G."
Fusion Systems Group
[e-mail] -or- ...!uupsi!fsg!reynolds

##

Subject: Re: The "State of 3-D Art" in Amiga World
Date: Thu, 18 Apr 91 12:59:03 EDT
From: Mark Thompson <[e-mail]>

Brian Reynolds writes:
> the Imagine (captioned as upper left) and LightWave 3D (captioned as
> lower left) images were accidently exchanged, so some of the following
> may make no sense at all.  Steve, how did you know the images were
> switched?)

Imagine does refraction and Lightwave does not. The images show exactly the
opposite.

> I thought that the blue light was too powerful in
> the upper left image.  It completely over powered the green coloring of
> the glass. (If this is the Lightwave image, it makes sense. You can't
> do refraction mapping they way you can do reflection mapping.)

Refraction would not make the glass look any greener. Since I don't know
exactly how this scene was setup, I cannot account for the very blue
appearance of the glass. I do know however that a PURE blue light on a
green object will reflect only blue. It was possibly a mistake on the part
of the person setting up the scene.

> The reflection in the chrome ball also looked a little strange.  I would
> have expected more curvature in the reflected image.

The curvature is goimg to be a function of both Lightwave's reflection
mapping and the source reflection image. If the image used to simulate
reflection does not exhibit any fresnel distortion (the fish-eye look),
the mapped reflection will not look totally correct. Pixar alleviates
this problem in environment mapping by wrapping 6 images (one for each
direction in the environment) onto a reflective surface. Since reflection
mapping only uses one image to capture all views, it is somewhat limited.

> What color do green and blue yield, yellow?)

It helps to think of the surface as a filter. Green is made up of blue and
yellow. If blue light passes through a filter that transmits blue and yellow,
then blue comes out (attenuated by whatever the blue component of the surface
is: x/255).

> The article mentions that
> getting shadows takes a lot of time.  Is there any reason shadows
> should take more time than reflections?

Reflection mapping is a very simple geometric transform, very similar to
image texture mapping. Shadows however are done with ray-casting (in
Lightwave) which is nearly identical to ray-tracing except all the additional
lighting and reflection/refraction calculations are not necessary. It is
still far more computationally intensive than 3D surface transforms (reflection
mapping).

> think that you would render a shadow map.

Nope, see above.

> I took the RenderMan course at SIGGRAPH last year, but it
> has been a while and I don't quite remember how shadow mapping works.

Shadow mapping is equivalent to depth mapping. For each light in the scene,
a distance calculation is made from the light to the nearest surface at
every pixel. This is used to determine what is visible from each lights
perspective.

> I like the idea behind Animation:  Journeyman. (Spline based modelling)

Me to!!!

> Does anybody
> think they will add raytracing splines (or NURBS, I don't really know
> what they are using) to their product?

I have asked Allen in the past to add spline patches to Lightwave and
although he is interested in doing it, it might be a little while.

Phew, thats the most my slow hands have typed all week. Hope this helps.
|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
|      `       '        Mark Thompson                 CONCURRENT COMPUTER  |
| --==* RADIANT *==--   [e-mail]        Principal Graphics   |
|      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   |
|     Productions       [phone] [phone]   & General Nuisance   |
|                                                                          |
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

##

Subject: Re: Repeating Brushmaps
Date:    Thu, 18 Apr 91 14:38 EDT
From: "Doug Bischoff" <[e-mail]>

     Well, still no luck on those darn repeating brushes with a ground.
Here's the setup:

     Ground
     Brush is a black-and-white 100x100 checkerboard pattern.
     Applying the brush just by re-sizing the front view and moving the top
view so the axis isn't right on the ground axis in any view.  (Box still inter-
sects the axis, tho.)  All done in Local mode.
     Not rotating the ground at all.

     I render it... and STILL get just a left to right stripe of the brush pat-
tern.  NO top to bottom (as you look at a perspective view.)  This happens in
both Scanline AND Ray-trace... no difference.

     Second question: I have an indoor scene with walls, floor, and roof.  I pu
t the camera in there and it looks like everything's a dollhouse.  Either I'm
outside the structure, or anything I try to look at fills the screen entirely.
     I tried changing the focal length of the camera, and that just gave me a
VERY strange looking fish-eye view.  Should I scale the entire scene up a few
HUNDRED and change the world size to 8000 x 8000 x 8000 and see if that's enou-
gh?  Thanks for the help.

/---------------------------------------------------------------------\
| -Doug  Bischoff- |    *** ***    ====--\         | "I'm not God...  |
| -DEB110 @ PSUVM- |   *  ***  *     ==|<>\___     |    I was just    |
| -The Black Ring- |    *** ***        |______\    |       misquoted!"|
| --- "Wheels" --- |      ***           O   O      |   -Dave Lister   |
| Corwyn Blakwolfe |     T.R.I.     -------------  |    RED DWARF     |
\---- [e-mail]  D.BISCHOFF on GEnie  THIRDMAN on PAN -----/

##

Subject: ground brush maps
Date: Thu, 18 Apr 91 11:52:12 -0700
From: [e-mail] (Ed Chadez)

Would someone mind reposting the instructions on how to apply a brush map
to the ground?  Obviously Doug Bischoff (and myself for that matter) don't
know how.  (Actually, I lost the instructions that someone placed before.)

Thanks, in advance.
-Ed

--
--//---------------------------------------------------------------------------
\X/ [e-mail]/Edward Chadez                    CARL Systems(303)861-5319
-------------------------------------------------------------------------------

##

Subject: "State of the Art"
Date: Thu, 18 Apr 91 19:04:03 EDT
From: [e-mail]

Brian Reynolds asks how I knew the Imagine & Lightwave pix were mixed:


It immediatedly lept out at me. It's easy to see there is no refraction in
the glass cylinder, and the chrome ball does not reflect it's base, just the
mapped checkerboard. These effects are in the Lightwave image, which it cannot
do.



-Steve

---------------------------------------------------------------------------
Steve Worley                                        [e-mail]
---------------------------------------------------------------------------


##

Date: Thu, 18 Apr 91 20:59:05 -0400
From: [e-mail]

RCS will be loaning us a unit for a few shows/demonstrations/evaluations
in the comming weeks/months. We will receive it in less than 2 weeks. they
are based here (in Montreal) so it makes things easier.

For those who don't know, RCS Management has been advertising an 040
board which pumps out over 20 MIPS for the Amiga 2000 for awhile now.
I don't have the specs on the thing but my boss says that you can put
32MB RAM on the board (SIMMS I expect). The price, I was told approx
$3000 canadian (about 2500 american), but don't count on it. Anyhow,
everyone is very excited (especially the Macintosh club...)

As for evaluations, I will do Lightwave and Imagine 1.1 for sure, as
for the rest, we don't have latest versions and don't use them. I
*would* like to use Journeyman as this type of program is *my* type
of animation tool... Salut les gars!
                            Stef.
BTW: Has anyone seen the latest AWOLRD animation video II yet? Better
than the first I hope!

##

Subject: Re: Buddy System & Object Disks...
Date: Thu, 18 Apr 91 18:22:14 PDT
From: "Jim Lange" <[e-mail]>

In-Reply-To: WRPYR:[e-mail]'s message of 04-18-91 10:24

[e-mail] writes:

> I am a bit surprised there has been nary a comment about the semi-review
> I wrote about the Buddy System for Imagine, nor has there been a comment
> about the object disk idea.

Thanks for the review.  I'm glad products have arrived to fill the void created
by Impulse's poor documentation.  However, now it's possible to spend almost as
much money on learning aids as most of us spent to upgrade in the first place!

    Surface Master        $30    (prices are approximate)
    Buddy System        $30
    Imagine Companion book    $30
    Guided Tour video    $30
                ---
                   $120

Perhaps others can post a review of the book and video in the near future.  I
suspect that there is much overlap between these products.

As far as the objects disk idea, many people are already posting objects to ftp
sites and this is working very well.  There may not be enough demand for PD
disks (at least within this mailing list), but I'm sure there are many users
out there struggling with Imagine that don't have the fine resources we do.

-------------------------------------------------------------------------------
Jim Lange                [e-mail]
Oracle Corporation            {uunet,apple,hplabs}!oracle!jlange
-------------------------------------------------------------------------------


##

Subject: Your point of view...
Date: Fri, 19 Apr 91 1:37:19 MET DST
From: [e-mail]

Ok, here is an easy one for you...


If I want a raytracer that should be able to do NICE objects and NICE
texturing and overall quality pictures, then which one is better,
Imagine, Lightwave or JourneyMan? Or put it this way: Which one
makes pictures most resembling Pixar's (Renderman)?

I know this is a very blurry question :), but maybe someone
could tell me what that picture in AmigaWorld (may) looks like?
(You know that one about "The 3-d state of the art" or something.)



   -- Henrik


Internet: [e-mail]




##

Subject: Imagine files at ftp sites
Date: Thu, 18 Apr 91 18:34:21 PDT
From: "Jim Lange" <[e-mail]>

Could someone volunteer to periodically post a file on abcfd20 and/or hubcap
that describes the contents of everything under the Imagine subdirectory?  For
those of us who only have BITFTP access, trying to determine what is out there
can be tedious.

-------------------------------------------------------------------------------
Jim Lange                [e-mail]
Oracle Corporation            {uunet,apple,hplabs}!oracle!jlange
-------------------------------------------------------------------------------


##

Date: Thu, 18 Apr 91 20:48:02 -0500
From: Donald Richard Tillery Jr <[e-mail]>

Regarding the "comparison" of ray-tracers in the May Amiga World, I would
like to point out that no fair comparison could be made by making the
green cylinder have any index of refraction other than 1.00.  Making the
Imagine, 3D Professional and Animation:Journeyman pictures render it with
an index of refraction, the render time is increased and this makes me
wonder about the accuracy of the "comparison."  That coupled with the
switch of at least two of the images, taints this article for me.  Even
other than that, I found the May issue to be one of the least accurate
issues I've seen in a long time (sorry, I can't quote much now, but I'm
thinking of giving it the once over again and writing Amiga World a
concerned letter).

Rick Tillery ([e-mail])


##

Date: Thu, 18 Apr 91 21:39:45 -0500
From: [e-mail] (Matt Feifarek)

    Can someone explain changing the camera's focal length again?

I tried scaling all 3 axes of the camera (both interactively and in Action)
but, no dice.  Any suggestions?

Also, how do you rotate something? <DUH!>  This sounds stupid, but here's what's happening to me:
    I made a sphere (my corroded sphere that I sent mail about...),
    made 30 frames....
    set frame 30's rotation to 360 (which reset to zero, so I picked 350)
    rendered a wireframe anim...
    It rotates -10 degrees instead of +350.
I tried splitting it in parts (0-90, 100-180, 190-270, 280-360) etc, but I
keep getting the negative resistance.

I know that I can now use the rotation f/x, but I shouldn't have to... this
should be easy!

My only theory:  When morphing position/alignment/size/objects etc., Imagine
ALWAYS takes the shortest path... ie -10 degrees instead of 350 degrees.

Can anybody see the flaw in my porcedure??

Thanks!

P.S.  Has anybody tried my corrosion thing yet?

    MJF

##

Subject: Rotating Objects
Date: Fri, 19 Apr 91 01:05:46 -0400
From: Udo K Schuermann <[e-mail]>

[e-mail] (Matt Feifarek) writes:
>     Can someone explain changing the camera's focal length again?
>
> I tried scaling all 3 axes of the camera (both interactively and in Action)
> but, no dice.  Any suggestions?

Focal length is a function of the X/Y ratio.  If you change both X and Y
simultaneously by the same amount, the ratio is unchanged.

> Also, how do you rotate something? <DUH!> This sounds stupid, but
> here's what's happening to me:

>    set frame 30's rotation to 360 (which reset to zero, so I picked 350)

It's ALIGNMENT and not ROTATION.  And yes, morphing the alignment
(orientation) will use "shortest distance between two angles."
    Breaking up the rotation the way you did should work.  I have
had success with something like 0-175, 180-355 on a 72 frame rotation,
so I know it can be done.  Try 0-170, 180-350.
    If you're not using any other F/X on the object, then use the
Rotate F/X -- it's VERY easy to use and can save you a lot of work.

HEY, IMPULSE: may I be so bold as to suggest giving us multiple F/X
per object?  How about information on writing our own F/X and
Textures?  -- just some suggestions, not meant as a flame!

Cheers!

 ._.  Udo Schuermann      "Does he talk?  DOES HE TALK?" -- "Of course I talk.
 ( )  [e-mail]   I'm the Minister for Internal Affairs." -- M.Python

##

Subject: Re: "State of 3-D Art" in Amiga World (Journeyman, Rayshade)
Date: Fri, 19 Apr 91 09:20:48 +0200
From: [e-mail] (Helge Egelund Rasmussen)

Brian Reynolds (fsg.com!reynolds) wrote:

> As Glenn Lewis pointed out I meant to say I am using Rayshade
> V4.0(beta) to do my tracing.  I use rayscene to create scene
> descriptions and script files for animation.

How good is Rayshade 4.0, I played around with 3.?, but I thought that
DKB was much better.

>  Personally
> I like the idea behind Animation:  Journeyman.  Splne based modelling
> makes more sense (to me) than polygon based modelling.  Does anybody
> think they will add raytracing splines (or NURBS, I don't really know
> what they are using) to their product?  Will the competition do it?

I haven't read the Amiga World article, so I can't comment on it, but
"Animation: Journeyman" can raytrace spline patches in the current version,
didn't the article mention that??

Helge
---
Helge E. Rasmussen  .  PHONE + 45 36 72 33 00  .  E-mail:  [e-mail]
Compel A/S          .  FAX   + 45 36 72 43 00  .
Copenhagen, Denmark

##

Subject: World Size (again)
Date: Fri, 19 Apr 91 06:32:18 EST
From: John J. Rosner <[e-mail]>

Rick, thanks for the reply of 32,000 per axis.  I have 6 meg (A3000), nothing
else running and Imagine would not take 10,000.  Does this seem correct?  I
could use an excuse to buy more memory.

Later,

John Rosner


##

Subject: Ram, Ram, Ram
Date: Fri, 19 Apr 91 09:06:58 EDT
From: [e-mail]

Remember my low RAM troubles? I just got 8M more ram (now at 2M Chip, 14M
Fast) so I shouldn't be overly concerned about low RAM anymore. Anyone
have any ideas how to exploit this? It's nice to have enough RAM to be
rendering a complex project while working on another at the same time, and
still be able to bring up the memory-monger Art Department.

Steve Menzies- you have an 18M A3000-- what do you do with your RAM?

Also, who's going to SIGGRAPH this summer? I am, as well as Scott Busse and
Mark Thompson. It's early, but we should definately plan to meet.

-Steve

---------------------------------------------------------------------------
Steve Worley                                        [e-mail]
---------------------------------------------------------------------------


##

Subject: Siggraph... where is it this year?
Date: Fri, 19 Apr 91 09:22:05 EDT
From: Sandy Antunes <[e-mail]>

Subject line says it all...
Where is SIGGRAPH being held this summer?  Hopefully the
New England area?  I always like an excuse to visit...
                             sandy
-------------------------------------------------------------------
[e-mail]     Sandy Antunes, El Loco d'Waupelani
  it's 2am... do you know what time it is?      -------------------

##

Subject: Re: Siggraph... where is it this year?
Date: Fri, 19 Apr 91 09:30:19 EDT
From: Mark Thompson <[e-mail]>

> Subject line says it all...
> Where is SIGGRAPH being held this summer?  Hopefully the
> New England area?  I always like an excuse to visit...

Nope, the other side of the world......Las Vegas.
I'm working on an animation for the film and video show as we speak.
%~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~%
%      `       '        Mark Thompson                 CONCURRENT COMPUTER  %
% --==* RADIANT *==--   [e-mail]        Principal Graphics   %
%      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   %
%     Productions       [phone] [phone]   & General Nuisance   %
%                                                                          %
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

##

Subject: Re: Your point of view...
Date: Fri, 19 Apr 91 10:08:51 EDT
From: Mark Thompson <[e-mail]>

>If I want a raytracer that should be able to do NICE objects
            ^^^^^^^^^
I assume you are using raytracer to generically mean 3D renderer. Neither
Lightwave or Renderman are raytracers (athough I don't believe the Renderman
spec precludes the use of ray-tracing for generating images; it merely suggests
other approaches).

>and NICE texturing and overall quality pictures, then which one is better,
>Imagine, Lightwave or JourneyMan? Or put it this way: Which one
>makes pictures most resembling Pixar's (Renderman)?

I'd rather not answer which is better because this is very subjective and
each has its own merits. None of them completely fulfill all my desires in a
3D program but they are getting there. As to which one makes pictures most
resembling Renderman, I would have to say Lightwave. Renderman like Lightwave
uses various rendering 'tricks' to achieve a photorealistic look rather
than resorting to the time consuming task of ray-tracing. With the imminent
release of Lightwave 2.0, you will be able to do Renderman-like effects
such as motion blur, soft shadows, and depth of field as well as all the other
existing rendering features including: soft edge spotlights, fractal noise
textures (clouds, smoke, fire), glows, fog/haze, water waves, etc. Most of
these features are not currently available in the other Amiga 3D packages.
What Lightwave lacks is a high powered modeler like Imagine and spline
surface patches like JourneyMan. But it is arguably the best renderer
available (IMHO). One note about JourneyMan, while the animation capabilities
of this program absolutely thrill me, I have yet to see an image it created
that showed any semblence of photorealism. Maybe Helge Rasmussen could post
a JourneyMan image that will knock our sock off?

>I know this is a very blurry question :), but maybe someone
>could tell me what that picture in AmigaWorld (may) looks like?
>(You know that one about "The 3-d state of the art" or something.)

Well not only were some of the AmigaWorld pictures mixed up (as Steve pointed
out), but none of them truly represented the full rendering capabilities of
these programs. So I wouldn't base your opinions on just those picures.
Determine what capablities you need and look at many sample images. Then
decide which suits you best, or just buy them all :-)
%~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~%
%      `       '        Mark Thompson                 CONCURRENT COMPUTER  %
% --==* RADIANT *==--   [e-mail]        Principal Graphics   %
%      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   %
%     Productions       [phone] [phone]   & General Nuisance   %
%                                                                          %
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

##

Subject: Re: Ram, Ram, Ram
Date: Fri, 19 Apr 91 10:18:59 EDT
From: Mark Thompson <[e-mail]>

> Remember my low RAM troubles? I just got 8M more ram (now at 2M Chip, 14M
> Fast) so I shouldn't be overly concerned about low RAM anymore. Anyone
> have any ideas how to exploit this?
> Steve Menzies- you have an 18M A3000--

OOOOOHHHHHH, you guys make me so MAD!!!!  ;-}
I wish not knowing what to do with it was a problem for me (I only have 7M
and can't go above 9M).

> what do you do with your RAM?

Many, many polygons and lots of 24bit image maps.

> Also, who's going to SIGGRAPH this summer? I am, as well as Scott Busse and
> Mark Thompson. It's early, but we should definately plan to meet.

Yes, definately.
|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
|      `       '        Mark Thompson                 CONCURRENT COMPUTER  |
| --==* RADIANT *==--   [e-mail]        Principal Graphics   |
|      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   |
|     Productions       [phone] [phone]   & General Nuisance   |
|                                                                          |
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

##

Subject: Re: Siggraph... where is it this year?
Date: Fri, 19 Apr 91 11:52:04 EDT
From: Mark Thompson <[e-mail]>

> WHEN is Siggraph, and do you have any ticket info??

July 28 - August 2
Admission into the exhibition is typically around $15 and the Film & Video
show is usually $20. Both are included if you take a coures or register
for the papers/panels. Below is info on them that was posted to comp.graphics:

             ------------------------------------------------

First of all, the SIGGRAPH '91 advance program is now being mailed,
to the following:

- SIGGRAPH members
- SIGGRAPH '90 technical program attendees
- people who sent inquiries about SIGGRAPH '91

It takes a while to get all the programs mailed, so don't panic if
you haven't gotten it yet.

If you aren't already receiving the advance program per above, you may
request it via mail, email, phone or fax.  Note that the postal and email
addresses have changed from last year.  Be sure to include your postal
address in your request; the advance program is 60 pages long and cannot
be sent via fax or email (or read to you over the phone :-).

SIGGRAPH '91
401 N. Michigan
Chicago, IL  60611
[phone] (voice)
[phone] (fax)
[e-mail]

>       Does anyone out there have any idea how much conference
>       fees are likely to be for a non-ACM non-SIGGRAPH member type
>       person this year? Just a ball park figure would be fine.
>       Last years fees would be ok.

Okay, but I won't be responsible for any typographical or transmission
errors, and SIGGRAPH is certainly not responsible for them either.
All prices are in US dollars (as all payments must be).  See the advance
program for details regarding who qualifies for the various rates.

Two days of courses                Member  Non-member  Student
Postmarked on or before 21 June     435       500        205
Postmarked after 21 June            540       625        255

One day of courses                 Member  Non-member  Student
Postmarked on or before 21 June     255       295        120
Postmarked after 21 June            320       365        150

Papers/Panels                      Member  Non-member  Student
Postmarked on or before 21 June     255       390        130
Postmarked after 21 June            425       490        215

Educators' Program:  $125  (free with paid papers/panels registration)

>       Secondly how difficult is it to find lower to middle
>       (backpacking up to motor inn) type accom during
>       the conference period? Is it ok to leave getting
>       (that sort of) accom to chance at that time? or is
>       it imperative to get something organised before
>       then (how long before?).

I don't know of any city that has a greater variety or quantity
of hotel rooms than Las Vegas.  It's hard to imagine that you
couldn't just show up and find *something* (but having said that,
it will now happen to somebody ;-).  On the other hand, conference
hotels are reasonably priced ($45-$95 per night, plus 7% tax,
single or double), and have the advantage of free SIGGRAPH busing to
all conference activities.

|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
|      `       '        Mark Thompson                 CONCURRENT COMPUTER  |
| --==* RADIANT *==--   [e-mail]        Principal Graphics   |
|      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   |
|     Productions       [phone] [phone]   & General Nuisance   |
|                                                                          |
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

##

Subject: Marble objects.
Date: Fri, 19 Apr 91 18:38:40 MET DST
From: Henrik Harmsen <[e-mail]>

Hi!

I'm still waiting for my Imagine package to show up..:( But in the
meantime I would like to ask if it is possible to render Marble
with Imagine. I know it can render wood and lots of other stuff, but
I have never seen any mentioning if marble. Maybe the wood texture
could be modified into rendering marble?


  -- Henrik (waiting idly by for Imagine to enter his life .. :)


##

Subject: Re: Siggraph... where is it this year?
Date: Fri, 19 Apr 91 11:57:52 EDT
From: Mark Thompson <[e-mail]>

I forgot to mention that every year I usually get several free tickets
for admission to the exhibition from various vendors. If anyone who
is definately going could use one of these, let me know when it gets
closer to that time (around June or so because I won't have them till then).
|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
|      `       '        Mark Thompson                 CONCURRENT COMPUTER  |
| --==* RADIANT *==--   [e-mail]        Principal Graphics   |
|      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   |
|     Productions       [phone] [phone]   & General Nuisance   |
|                                                                          |
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

##

Subject: Re: Imagine files at ftp sites
Date: Fri, 19 Apr 91 11:59:18 PDT
From: "Jim Lange" <[e-mail]>

Doug <WRPYR:[e-mail]> writes:

> Some of them have associated readmes on hubcap, and Ill (ah, after finals and
> orals) do the honors of adding readmes for each file.
>     -Doug

I was actually thinking of a single file that lists the directory structure and
contents such as the output from an 'ls -RC'.  Descriptions and readme files
would be nice, but I just need to know what's there.

-------------------------------------------------------------------------------
Jim Lange                [e-mail]
Oracle Corporation            {uunet,apple,hplabs}!oracle!jlange
-------------------------------------------------------------------------------


##

Subject: Re: Ram, Ram, Ram
Date: Fri, 19 Apr 91 13:55:51 EDT
From: [e-mail] (Bob Lindabury - SysAdm)

Mark Thompson <rutgers!westford.ccur.com!mark> writes:

> > Also, who's going to SIGGRAPH this summer? I am, as well as Scott Busse and
> > Mark Thompson. It's early, but we should definately plan to meet.
>
> Yes, definately.

Well, if you guys can buy me travel tickets and hotel rooms, I'll go!
<grin>

Really, if it was on the east coast, I would be able to hack it but
alas, it is just too much moola for my wallet.  Unless I get a real
good 3D animation job....yeah..that's the ticket!

-- Bob

 The Graphics BBS  908/469-0049  "It's better than a sharp stick in the eye!"
 ============================================================================
  InterNet: [e-mail]                | Raven Enterprises
      UUCP: ...rutgers!bobsbox!graphics!bobl      | 25 Raven Avenue
    BitNet: bobl%graphics.rent.com@pucc           | Piscataway, NJ 08854
    Home #: 908/560-7353                          | 908/271-8878

##

Subject: Re: Your point of view...
Date: Fri, 19 Apr 91 13:53:01 EDT
From: [e-mail] (Bob Lindabury - SysAdm)

Mark Thompson <rutgers!westford.ccur.com!mark> writes:

> Determine what capablities you need and look at many sample images. Then
> decide which suits you best, or just buy them all :-)

Funny you should mention that.  Seems that you and I have done just
that!  Sometimes you have to purchase them all to figure out what is
going to work best for you.  And in some cases, you find that 2
packages can compliment each other as in the Imagine Detail Editor
(Modeler) and Lightwave 3D.

I hate buying all the packages however.  <grin>

-- Bob

PS: Haven't gotten Journeyman yet as I have yet to see a decent image
come out of it.

 The Graphics BBS  908/469-0049  "It's better than a sharp stick in the eye!"
 ============================================================================
  InterNet: [e-mail]                | Raven Enterprises
      UUCP: ...rutgers!bobsbox!graphics!bobl      | 25 Raven Avenue
    BitNet: bobl%graphics.rent.com@pucc           | Piscataway, NJ 08854
    Home #: 908/560-7353                          | 908/271-8878

##

Subject: Re: Imagine files at ftp sites
Date: Fri, 19 Apr 91 17:34:39 EDT
From: [e-mail] (Doug)

> contents such as the output from an 'ls -RC'.  Descriptions and readme files
> would be nice, but I just need to know what's there.


Sorry, I misunderstood.  There is a new file "files" in the IMAGINE
subdirectory that is just a ls -l -RC (-l will give size and date).
    -Doug

##

Subject: A treatise on Texture
Date: Fri, 19 Apr 91 20:43:46 EDT
From: [e-mail]

This is the first article of a pair describing the use of textures and
brush maps. There is a lot of confusion on how to use both of these
powerful features, and if you learn to use them properly you can create
some astounding images. This first article is a general description of
the use of texture.

-----------------------------------------------------------------------

                             TEXTURE


No matter how good you are at creating objects, they are never detailed
enough, especially in surface appearance. Textures allow you to quickly
give your objects a complex, detailed surface without much work. A
woodgrained picture frame has a lot more character than a flat brown one.

Textures are just algorithms that Imagine uses to decide how to color
a particular point on the surface of an object. Textures can also determine
the transparency, reflection, and surface angle at each point. They
can even be layered on top of each other for more complex effects.

Each texture has its own particular inputs. Most of the time, you need
to enter a color, transparency, and reflection value for whatever detail the
texture is applying to your object. These are raw numbers- no sliders. I
often use the color sliders in the attribute requester to chose the texture
color or whatever, then write the numbers I want down. [Remember to reset
the default color!]

Note that for most textures, when I say "color" I mean surface color,
reflection, and transparency. Most textures can set all three!
Some textures, like disturb, will affect surface light reflection like
altitude brush maps do. More on altitude maps in the next article.

There are often a few extra parameters to set dealing with the way the texture
is applied. This might be check spacing, wood grain thickness, or brick
size. These are pretty straightforward, especially since the photocopied
Imagine Manual addenda have OK descriptions of what each parameter does.
One very important note: Any raw size measurement, like check size, (anything
that measures a distance) is measured in STAGE EDITOR units. Huh? What I
mean is that if your check size is 100, every 100 units in the STAGE editor
you'll get a new check. This won't matter if you don't resize your objects
in the stage editor, but if you design a checkerboard thats 80 units wide,
and set check distance to 10, things might work out great. If, however,
you scale the object in the STAGE editor to 160, you're going to get 16
checks across. -->Texture parameters do NOT scale with objects<--

The only other parameter is the texture axis, which can be manually edited.
The texture axis is pretty important. For most textures, you need a
"base" location and orientation to give the details a reference. For
example, the linear texture needs to know where the "fade" starts and
what direction to fade in. What you do is just place the texture axis
where you want the fade, and point the Z axis in the direction you
want it to go. The wood texture at it's simplest is a bunch of
concentric cylinders of coloring. Where should the center be, and
which direction should the cylinder point? The texture axis will tell
you. Some textures don't care about the axis, though, like Camo.

An important point- if your texture axis is RIGHT on a face, you might
get some funky effects, since for a texture like checks, the surface of the
object is EXACTLY where the checks change. The algorithm does not know what
color to return, so you get what Impulse calls a "digital bounce." This is
most common when you're texture mapping a flat plane. Fix: move the texture
axis just a tiny bit.

A tip- wood looks best when the axis is nearly, though not quite, parallel
to the longest object dimension. This gets you nice grain cross sections,
and looks more realistic (who ever cut two-by-fours across the grain?)

I could go into a discourse about each texture, but that'd take a
lot of time. The trick is to play with them! Mike Halvorson loves
to say "just play with the feature, you'll get it", which I find
true, but very vague and annoying to users who are already confused.
However, for people who are comfortable with Imagine, its the best
way to extend your mastery  and produce some truly delightful scenes.

Many, if not all, textures only affect some parts of an object.
The camouflage texture is an excellent example. You set the default color
of the object from the attributes requester. The Camo texture then
layers its spots ON TOP of this default. If there is no spot on a
particular location, the default color will show through. This is
true with most textures. Wood only adds the "grain" and lets the object's
default color become the normal woody non-grain parts. Linear gradually
fades from the default to another color. Checks adds color on its checks
and lets the default attributes stay in the opposing checks.

Why is this important? Well, this can be used to our great advantage. YOU
CAN ADD UP TO FOUR TEXTURES SIMULTAINOUSLY. They are added in order
from 1 to 4. What can you do with this? Well, you can take a desk, and
with Texture 1 add a wood texture. Then you can add a camo texture as #2,
and the spots will cover up the wood, but you'll see GRAINED wood where
there are no spots and NONGRAINED, solid color spots where there's a
camo spot. This can be used up to four times.

I have an island (I'm working on "Ocean Sunset")  and I've given the
Vista-created terrain three textures. First, a Radial texture which
varies the base color from two subtle shades of sandy-brown. Second,
my sandstone texture (I posted long  ago) to simulate the sandy shores.
[Maybe camo with TINY spots would work.. Hmmm] Then, a linear which
fades the beach into a nice vegetation green color once you get past
a certain distance from the beach. These give the island a nice,
detailed character that I'd never be able to match by picking and
coloring individual polygons.

Louis Markoya's Surface Master has some example combinations of textures.
He has pictures that show how the different texture parameters affect the
final object appearance. I believe he showed all the textures except Camo.
He also had a selection of parameters for the wood texture for different
pieces of wood- it was VERY nice. The examples of combinations of textures
(like dots on brick) were pretty cheesy, though.

The most useful textures are probably wood and linear. Wood can
do a lot of powerful effects, and linear is useful everywhere. The
other textures are useful, too, of course, but I use linear and wood
the most. There are a lot of impressive things you can do by abusing
textures :-) Here's a fun one:

 o  Create an object. A long logo works great. Color it and texture it
    any way you want.

 o  Add a linear texture, set the Z transition width to about 20% of the
    object length. Put the texture axis way over to one end, oriented
    towards the center of the logo or whatever. Make the color of the
    texture be black, no reflection, and 255 255 255 filter. Yes, completely
    transparent. Make sure the linear texture is the last one if you
    already have some other textures on the object.

 o  Render. You should have basically an invisible object, since the
    linear texture is completely transparent and covers the whole
    logo. Fix the axis if its pointed the wrong way.

 o  Copy the object. Move the texture axis way to the other side, oriented
    the same way. Save it with a DIFFERENT filename. Test render. It should
    look just like your normal object without a funky linear texture.
    It should certainly NOT be transparent.

 o  The fun part. In the action editor, morph object one into object two.
    The only change is the texture axis, so Imagine will interpolate
    its location from one end of the logo to the other for each frame.
    Make the animation at least 10 frames, preferably 20. You can render
    in scan- it'll work just fine.

    What happens is the linear transition band "flies" across the logo,
    fading the logo in as it moves from one side to the other. It's an
    impressive way to introduce an object into a scene! It is also pretty
    easy to do... 10 minutes tops.


Textures are really powerful, and if you haven't played with them, START!

-------------------------------------------------------------

The companion article on brush maps will follow Real Soon Now.


-Steve

---------------------------------------------------------------------------
Steve Worley                                        [e-mail]
---------------------------------------------------------------------------


##

Subject: Flaming Lasers
Date: Sat, 20 Apr 91 12:33:59 EDT
From: Sandy Antunes <[e-mail]>

Well, I finally mastered an obvious technique, i.e.
How to make flaming lasers  in Imagine!
This technique is useful for lots of effects, actually.

Just make your animation as usual, making the frames in ILBM
format.  HAM or HiRes is okay (heck, use lores if you want!)

Figure out which frames you want the lasers in.  Load them one
by one into DPaint or DigiPaint or whatever your favorite paint
program is that handles that format.  PD programs like "paint"
are useful enough for lasering, but it's nice being able to do
more complicated stuff.

Then, just draw in the lasers.  I suggest they be slightly blurry
rather then having crisp edges, but it's easy enough to find one
that looks nice.  Save each file back to its original name, then
hop back into imagine and have it make the animation for you!

I suppose if you're doing 24bit images you'll need MacroPaint or
such.  I don't know... I scrape along with an A500 w/only 3 megs.

Another great use here is deleting things from the pictures...
like individual stars and such.  Think of it like a real movie...
Imagine is the film crew and actors, the paint program is the special
effects company making final touches.

                                             sandy
-------------------------------------------------------------------
[e-mail]     Sandy Antunes, El Loco d'Waupelani
  it's 2am... do you know what time it is?      -------------------

##

Subject: Re: Flaming Lasers
Date: Sat, 20 Apr 91 10:17:41 -0700
From: [e-mail] (Ed Chadez)

Another use of Sandy's clever technique is to add special effects into the
ILBM pictures via a paint program just like Sandy did for the lasers.

I prefer Photon Paint or DigiPaint III (I don't own DeluxPaint...no flames
please) because most of my Imagine animations are HAM.  I added things
like shields and lasers and stuff like that much the same way Sandy
described.

However, it may be more convincing to actually build these objects in
Imagine.  With the abality to create translucent objects with transparent
bit-mapping and textures that can be seen thru, it shouldn't be too
difficult. Especially now since you can add brightness to objects.
Imagine (no pun) a long narrow cylinder that is bright, which is at the
center of another cylinder that is partially transparent with either a
texture applied to it (wood or disturbed, I suppose) or a bit map made
with your favorite paint package.  With enough work, you could produce
conviencing effects!


-Ed.

--
--//---------------------------------------------------------------------------
\X/ [e-mail]/Edward Chadez                    CARL Systems(303)861-5319
-------------------------------------------------------------------------------

##

Subject: Part Two- Brushmaps. (long)
Date: Sat, 20 Apr 91 14:21:49 EDT
From: [e-mail]

Brush Maps. Second in a series about adding personal touches to
your Imagine objects. I'm sure this is the one that is more useful,
since brush maps are inherently confusing.

------------------------------------------------------------------------------


Textures are not the only way to add detail to an object. A more
direct, less elegant, but more personallizable method is to use brush
maps. Brush maps are ways of taking standard Amiga pictures (sometimes
called IFFs, though pictures are a subset of the Interchange File
Format) that can be placed "by hand" on your object. In its easiest
incarnation, you could brush map a picture of your face onto a flat
plane, put a frame (using wood texture! ) around it, and you have a
virtual art piece. In its most complex incarnation, you could take a
set of 40 256-level intensity maps saved as IFF24s and tile them
endlessly on a plane during a 40 frame anim, and have the map pixel
intensity create reflections and highlights from the flat plane just
like it was really an animated, wind-wave covered ocean.

Brush maps can give objects the same four characteristics that
textures can- surface color, reflection, transparency, and surface
orientation. Going back to the "my face in a frame" example, a color
map is straightforward. A reflection map will reflect the color and
intensity of light corresponding to the map- in other words, a black
map would make the picture in the frame reflect no light, a white map
would make the picture a mirror, and a yellow map would make only
yellow light reflect. [The yellow example is not strictly true.. the
yellow-mirror would reflect green and red light, which combine to make
yellow]. Transparency (really filter) is similar. Black is opaque,
pure white is the clearest crystal, and yellow would let yellow light
though.  An example given by Rick Rodriguez in his manual is that a
filter map of your face applied to a plane would be like a
stained-glass window. You could have fun with that. Note that
transparent objects have an index of refraction, which is set in the
attributes requester. See my article, "The Art of Glass" for a much
more detailed discussion on transparency.

The last brush type is altitude, which is a bit more tricky. Here, the
IMAGINE MANUAL IS COMPLETELY WRONG! [Sorry, Rick!] The manual
describes a "displacement map", which is somewhat similar. An altitude
map just tells Imagine that light hitting the object's surface should
be reflected, refracted, and specularated (!) as if it hit a surface
that had a certain shape to it- the shape described by the brushmap's
intensity. If you mapped a picture with lots of small fuzzy grey dots
onto a sphere, you would get reflections and light highlights as if
the sphere had tiny pits in it like an orange. NOTE! The altitude map
does NOT change the real surface height of your object at all. THIS is
the difference between a displacement map and an altitude map.

One option you should use if you're using transparency map is the
"Full Scale Value" in the brush requester. This allows you to tell
Imagine how transparent a pure white image should be. If you want pure
white to be COMPLETELY transparent (invisible) set this number to 255
(full scale).  If you set the number lower, you get full white to
become opaque to whatever degree you wish. If you're making a stained
glass window, you might want to set the number to 200 so that people
can SEE that there is "glass" even where there is a clear pane.
Impulse's old default was 245, I think.

Any standard Amiga IFF can be used as a map. This includes all
pictures saved by Deluxe Paint, the most common picture editing
program on the Amiga. A format of high color resolutions known as
IFF24 saves three bytes of color information per pixel and produces
beautiful color reproduction. These IFF24s can be produced from the
ToasterPaint, Digi-view RGB saves (only 21 bits, but it is saved as 24
for compatibility), The Art Department, and many other graphics
programs. The Art Department is the penultimate tool for manipulating
24-bit (and other) images- it can scale, compost, and manipulate
24-bit pictures painlessly, as long as you have RAM. [It is NOT a
drawing program, though] Whatever type of brush you use, remember that
Imagine can't do magic to your pictures- a 16 color picture of
yourself is going to look positively cheesy compared to a 24-bit
picture. For some applications, though, you don't need much more.
Applying a 4-color logo to an object won't benefit by using a 16
million color brush - just use 4. Also, higher resolution is obviously
higher quality, though it's pointless to use a 1000 by 1000 picture of
a logo- even in a close-up shot, it would be indistinguishable from a
640 by 400 picture. Too low a resolution, though will be painfully
obvious when you render. Depending on how close of a view you use in
Imagine, you usually don't need more than 320 by 200. Again, this is
very dependent on how close your camera view is. I've used 100 by 100
brushes to great effect.

A fun trick- Imagine can easily _OUTPUT_ IFF24s... you can use a previous
rendering as a brush map.

If you are truly interested in making high quality brush maps, you
should definately use IFF24s as brushes. If you have DCTV, you have a
terrific paintbox for editing pictures. Toasterpaint also works, but
is not pleasant. If you don't have either (like me!), you can use a
trick that works extremely well.  I take a picture to be edited (often
grabbed with Digiview!) and in Art Department size it to about 900 by
600 (this is very RAM limited... grey images can get 3 times as
large). Then I display the picture in overscan dithered EHB, and save
this image. Then I power up Deluxe Paint, and edit the picture. EHB
gives you loads of colors to deal with, so you don't lose too much
color resolution. [You do lose some!] DPaint can take brushmaps of any
size [RAM limited!], and editing them is quite easy. With the very
large scale, individual pixels don't matter much, so I use large
brushes and the airbrush especially. When you're done editing, I save
the picture, and use Art Department to resize the image back to its
original size. For 24 bit pictures of a size of about 128 by 128, the
quality is terrific! You can see some of the brushes I made posted to
ab20 and hubcap (the sidewalk and latticework, and my latest brick
wall).

Sources of hi-quality pictures are everywhere. Digi-view is your
friend- it can make 768 by 480 resolution 24-bit pix. Toaster camera
frame grabs will give you 24-bit pix at a resolution of 768 by 240.  A
fun source of varying quality images are GIFs. GIFs are a graphic
picture format, most popular on MS-DOG computers. There are THOUSANDS
of pictures, most with 256 colors out of 16 million.  You can find
about 1000 with a nice index on the anonymous FTP site
wuarchive.wustl.edu.  I often take these pix and shrink them by 1/2
and save them as 24-bit IFFs- there's no color loss and the file size
is much smaller.

Once you have your map and know what you want to do with it, you have
to place it on your object. The placement determines the size and
orientation of the map, as well as how much of the object is
influenced, and in the case of altitude maps, how much surface light
is distorted.

There are three basic types of wrap- a "flat" wrap (Flat X Flat Z), a
"sphere" wrap (Wrap X Wrap Z), and a "cylinder" wrap (Flat X, Wrap Z
and Wrap X, Flat Z). Flat will ignore any surface bumps and features
and just apply itself directly, much like a slide projector would
project onto a bumpy screen. A sphere wrap tries to encase the object
in the brush, then shrinkwrap the map onto all of the surface features
of the object. The cylinder wrap tries to follow contours in one
direction, but ignore them in another. Think of taking a piece of gift
wrap, and bending it around so its a hollow cylinder.  Then place the
object in the center of this vertical gift wrap cylinder and push IN
(but not up or down!) to follow the object contours.

Placing the maps is sometimes tricky. Mike Halvorson wrote a
description that was posted about 6 weeks ago, you can find it in the
Imagine list archives. Also, the green Impulse Winter 1991 bulletin
had two very useful diagrams. What I'm about to describe is not as
simple and foolproof as just looking at the 4 simple pictures they
printed, due to the mere fact that ASCII graphics suck. But I'll try!

FLAT "WRAPS"

  Flat wraps are the most common and certainly the most controllable
of the brush map types. Think of having a decal or poster that you
want to stick onto a wall. Flat wrapping will do just this. A good
example is trying to put a logo onto the side of a truck- an excellent
example of where brush maps shine.

First, you should obviously have your logo picture and truck designed
and ready. Now, to place the brush onto an object, you should probably
use the "Edit Axes" mode. This lets you move the axes with the same
mouse and keyboard commands that you normally use for objects... m for
move, s for scale, r for rotate, x,y,and z to toggle a direction.

A danger with editing brush axes is that you want to be in LOCAL mode,
especially when you are scaling the brush axis in just one direction
(like you were trying to increase its height, keeping the same width
and depth.)  If you don't scale in LOCAL mode, sometimes your changes
don't stick. You can check by selecting "Edit Axes" again to make sure
nothing changes after you're done.

The axis you are editing has a bit yellow bounding box that is very
deceiving. THE AREA WHERE THE BRUSH IS ACTUALLY MAPPED IS JUST THE
UPPER RIGHT QUADRANT OF THIS BOX. The brush is placed with its lower
left corner right at the center of the axes, and its upper right
corner at a point defined by the X and Z axes of the brush map axis.
Got that?

            Z
 +----------^------------+     Front View
 |          |xxxxxxxxxxxx|
 |          |x Picture xx|
 |          |xx Area xxxx|
 |          |xxxxxxxxxxxx|
 |          +------------> X
 |                       |
 |                       |
 |                       |
 |                       | <-brush map bounding box
 +-----------------------+


You want to position the axes so that the upper-right quadrant lies
exactly where you want your brush to lie. If you want your brush to
cover the entire side of the truck, you'd probably want to make the
brush a few extra pixels high and wide so that you don't accidentally
get a border around the edge of your logo.

The Y axis of the brush map is pretty important as well. It tells
Imagine how DEEP to apply your brush. Basically, any part of the
object that falls between the axis origin and the tip of the Y axis
WILL be colored (or reflected, or whatever). For the truck, you'd want
to move and scale the brush axis in the Y direction so that the Y axis
line INTERSECTS one side of the truck but NOT the other. The
intersected side of the truck would be within the influence of the
brush map, whereas the other side of the truck would be left alone. If
you scaled the Y axis to include both truck sides, the other side of
the truck would get the brush map applied to it as well. [In fact,
you'd see a mirror image of the brush on the other side, since you'd
be looking at it in the other direction.]

Here's a terrible ASCII drawing showing how you'd position a brush so
that it puts a very small logo on the side of a big rectangular solid,
like a truck body.

                 FRONT VIEW


        +-----------------------------------------+
        |                                         | <-Truck body
        |                                         |
        |                                         |
        |                  ^ X                    |
        |                  |                      |
        |        Brush     |                      |
        |        axis->    +-----> Z              |
        |                                         |
        |                                         |
        |                                         |
        |                                         |
        +-----------------------------------------+


===============================================================


                  TOP VIEW

        +-----------------------------------------+ <-Truck Body
        |                                         |
        |                                         |
        |                                         |
        |                                         |
        |                                         |
        |                  ^ Y                    |
        |                  |                      |
        +------------------+----------------------+
                           |
              Brush->      +-----> Z
               Axis



See how the Y axis only intersects one side?

That's about it for flat wrapping. It is pretty easy to control where
the brush gets applied to any object.


WRAP X WRAP Z  WRAPS

-----

These are the most complex wraps. The Y axis here isn't really used,
since the wrap is applied to the whole object surface, not just part.

You want to position the axes so that the Z axis covers the entire
height of the object... its height should be slightly TALLER than the
object. The X axis should be thought as a "radius of influence"... it
should be a little bit bigger than HALF the width of the object.  The
axis should be placed (LOOKING DOWN!!) at the center of the object,
and looking from the side, at the bottom of the object. The Z axis
should pass through and be slightly bigger than the object's height.
The X axis should be slightly bigger than the object's maximum radius
away from the center. The Y axis size doesn't matter. Keep it small,
suggests Impulse.


OK, Here's the diagram. The object is a sphere. I swear!
===========================================================
FRONT VIEW
                     ^ Z <- brush axis
                     |
                   __|__
                  /  |  \
                 -   |   -
                /    |    \
               /     |     \
              |      |      |
              |      |      |
              |      |      | <-Sphere-like object
              |      |      |
              \      |     /
               \     |    /
                -    |   -
                 \   |  /
                  ---+--
                     |
                     +--------> X



===========================================================
TOP VIEW
                   _____
                  /     \
                 -       -   <-Sphere-like object
                /         \
               /           \
              |      ^ Y    |
              |      |      |
              |      +------+-> X
              |             |   Brush Axes
              |             |
              \            /
               \          /
                -        -
                 \      /
                  ------



That's it! If you wrap a picture of a grid of lines, it will come
out looking like latitude and longitude lines on a globe.



CYLINDER WRAPS (WRAP X, FLAT Z, FLAT X, WRAP Z)

The brush placement is identical to the placement used for the
spherical wrap. The effect is quite different, however. If you tried
cylinder wrapping the grid picture onto a sphere, you'd get OK
latitude lines (going North-South) but the longitude lines would get
further apart the closer you were to the poles, due to the flat
projection of the horizontal lines. The lines themselves would also be
wider at the poles.


One last note for brush axis placement- for adding ALTITUDE maps, the
Y axis depth is used to measure how much indentation a full scale
range of intensity (0-255) should simulate. For cylinder and sphere
wraps, just scale the Y axis. For orange pits, the axis might be 1% of
the sphere's size. For an eroded planet, you might use 10%. More than
this would make really stupid looking reflections. Altitude maps are
subtle.

The Y depth is also used for flat altitude wraps, which might limit
you if you want to indent both sides of a truck. You'd have to use two
brushes in this case. I would have preferred a number gadget in the
brush requester.

----------

Once you know how to place individual brushes, you can start with the
fancy tricks. Brushes overlay each other just like textures. You can
put up to 4 brushes on an object, and they are applied in order. Many
maps don't interfere, though- you could have a color map and a
reflection map on the same object in the same place. Both will work
just fine.

I am not positive, but I'm pretty sure brushes are put on AFTER
textures are applied. Otherwise you'd get wood grain on your face. :-)

Repeating brushmaps are a joy. They "tile" an object with an endless
succession of images both side to side and top to bottom. The brush
will repeat all the way out the the end of the object. If you tile a
ground, the brushes will go to infinity. The size of each tile is set
by the brush axes, just like a non-tiled map. The brushes are placed
next to each other with no space between them. You could draw a VERY
detailed picture of a bathroom tile, and map it onto a wall.  when
rendered, the wall would be (surprise) tiled! I've used this to GREAT
effect for making very detailed sidewalks, rose trellises, brick
walls, roof shingles, and golf greens. All of these are flat wraps. To
be honest, I haven't tried infinite tiled wrap wraps...  I'm not sure
what they'd do!

Placing repeating brushmaps is just the same as a regular flat wrap.
The size of the brush is determined by the X and Z axis, and the
"depth of influence" is determined by the Y axis.

One additional option that is very useful is the "mirror" option. This
makes every tile be a mirror reflection of its neighbor. The great
advantage of this is that the edge colors ALWAYS match, just like if
your finger touches a mirror, your twin in the mirror will reach out
and touch your finger at the exact same place. This might hide
discontinuities in your brushmap if you want to hide the seams, or it
might be a special effect you're looking for.

Repeating brushmaps aren't just for covering an infinite plain with
your face. They can be an extremely powerful way to get very complex
textures on an object. You can imagine drawing one very high
resolution, high quality brick with scratches, pits, chips and tiny
detail, then tiling it onto a wall. Presto! You have a brick wall with
a lot of character, unlike the Brick texture which is too plain to
fool anyone up close. This is probably the most useful aspect of
infinite tilings. You can find my infinite sidewalk, infinite golf
green, infinite brick wall (very nice!) and infinite rose trellis on
ab20 and hubcap.

When I refer to "ab20" and "hubcap" I mean the anonymous FTP sites
ab20.larc.nasa.gov and hubcap.clemson.edu, both of which have a lot of
Imagine files, objects, brush maps and goodies that I've uploaded.
There is also an archive of this list on both sites. If you need to
know if you have FTP access and how to use it, ask one of your local
computer experts- they should be able to help you.

The last brush map ability is very useful. You can actually have
ANIMATED brush maps which change every frame.  Note that these
pictures are not "AnimBrushes" that DPaint will save. These are
individual pictures, which means you can have a 24 bit "animbrush."
To use this feature, you should save the sequence of pictures you wish
to show in a format like

  Mypic.0001
  Mypic.0002
  Mypic.0003
  Mypic.0004
  Mypic.0005

and so on up to however many pictures you have. Make sure to have FOUR
numbers in the extension. Mypic.01 will NOT work. To use this sequence
of pictures as an animation, you should use "Mypic" as the brush file
name [without the quotes], then set the "Max sequence #" to the number
of pictures you have. I wrote a LONG article about how you can make
animations of yourself using Digi-View and a VCR. You might have seen
the animation with me "trapped" inside of this giant sphere rolling
around on a plane, with my pitiful attempts to break out recorded in
16 glorious colors. You too can do this with YOUR setup. It's worth
playing with! You can find the article in the Imagine archives on
hubcap.

That's all there is to brush maps. They can add a lot of detail to
your objects, and they're not hard to use if you know how to place the
axes. Go out and create!



-Steve "Boy, are my fingers tired" Worley

---------------------------------------------------------------------------
Steve Worley                                        [e-mail]
---------------------------------------------------------------------------


##

Subject: new picture available ftp
Date: Sat, 20 Apr 91 22:07:36 EDT
From: [e-mail] (Doug)

I put an interesting ilbm on hubcap in the PICTURES directory called
"metal.dreams.lzh"


Its a hires overscan of a crumpled mass of metal with light rays beaming from
one side.  Simple trace, but the light effects (totaly unplanned, of course)
were real neat.

    -Doug

##

Subject: brushmaps on ground
Date: Sun, 21 Apr 91 14:32:32 CDT
From: [e-mail] (Randy Carris)

I'm also having troubles with frustrations with mapping a brush on the ground.  I tried to
map a simple rectangular brush with repeating on.  I tried it many different ways and got
results ranging from nothing to the single stripe that someone else was getting.  Could
someone who knows how to do this either describe this correct position or better yet ftp a
screen shot of it to ab20?  By the way I was doing a quarter screen ham scanline for
testing purposes, if this could possible be what is fouling it up. (It shouldn't)

Thanks,
Randy Carris
Second Look Mutimedia Computing Studio

##

Subject: Archive #4 on hubcap
Date: Sun, 21 Apr 91 15:59:56 EDT
From: [e-mail] (Doug)

Imagine archive #4 is now in the pub/amiga/3D/IMAGINE/TEXT directory.

hubcap:
hubcap.clemson.edu or 130.127.8.1



Lot of interesting postings in this one!


Bye
    -Doug
     [e-mail]


##

Subject: Imagine 2.0
Date: Sun, 21 Apr 91 15:09:17 EDT
From: [e-mail]

No, this isn't about what will be in 2.0, this is just an idea of
mine that SHOULD be in 2.0.

Mike Halvorson said that AREXX would be far to difficult to include in
Imagine. This statement pained me, since AREXX is the most beautiful
way to coordinate different programs together. Imagine (!) being able
to write scripts to place objects, automatically load and merge them,
create Vista-like landcapes by automatically using magnetism on a plane
a few hundred times... Automatically change the attributes of 100 different
objects simultaneously.. Geez. There are hundreds of uses.

Unfortunately, I have to agree with Mike... it would take a lot of
effort to make Imagine completely AREXX drivable.

How about just making the PROJECT editor drivable, though?? You would
want AREXX commands to do the following:

o  Open Project
o  Open Subproject
o  Select/Deselect frames
o  Delete selected frames
o  Generate selected frames
o  Make Movie with selected frames

Thats it! If you've seen the MinRexx interface code by Thomas Rokiki
(Amiga and Tex Dude!) you might know adding an interface for simple commands
like these would take about an hour of coding time.

But think of what you could then do!

You could generate a frame, have The Art Department display it (The Art
Department has far better dithering, plus color/contrast correction),
save the HAM pic, then delete the original Imagine pic for disk space.
Then, you could call MakeAnim to build an anim of the higher-quality
HAM pix.

You can AUTOMATICALLY do this! This would work perfectly for DCTV anims,
too. It's a pain to manually assemble them!

For power users, you could generate a frame, display it on your 24-bit board,
then signal your single-frame recorder, then go on to the next frame.

For my current project, "Ocean Sunset", I use a sequence of 40 450x450
24 bit pictures generated by a C program. Each is half a megabyte. 20
Megs total. I would love to generate the 24-bit sea brush, tell Imagine
to render, then delete the now-useless brush. What convienience!

Again, a FULL AREXX interface is too much to ask for, but I think a
Project Menu only version would be easy to implement and use and be
invaluable.  Would anyone else use this minimal AREXX control? If I'm
not alone, we might be able to insure Impulse includes it.


-Steve

---------------------------------------------------------------------------
Steve Worley                                        [e-mail]
---------------------------------------------------------------------------



##

Date: Sun, 21 Apr 91 21:25:39 -0500
From: [e-mail] (Matt Feifarek)

    I just saw the Todd Rundgren Video done with the Toaster.  Pretty Neat,
there were quite a few cool effects.  There was one part where a bump map in the
shape of his face pushed out of a plane and then faded into color...

Anyway, there was one effect that stumped me...

There was this flying street light ?!? ( stop light --- not street light)

that emitted a beam of light that you could see... almost like a mist coming
out of it.

Can anyone think of a good way to do this in imagine?  A plain cylinder would
be okay, but the edges would be hard, and it wouldn't look quite right...

I think that the linear or radial textures would work... one attribute on
transparent the other normal, but these cannot me cylindrical... only spherical
and linear.  I suppose you could use cylindrical and stretch one of the axes
WAAAAAAY out to approximate a cylinder... or would that work??  Any ideas??

Try and catch the video... it's nothing novel in terms of Computer Graphics,
but it is a whole lot of non-novel things in a short space of time ( nothing
that we haven't seen before from the biggies like Wavefront and Pixar, but
lots and lots of effects packed into a single short video)  PLUS, it's done
with our 'game machine' friend.

    Matt Feifarek    [e-mail]


##

Subject: Is there a ProVista News Group?
Date: Mon, 22 Apr 91 00:17 EDT
From: [e-mail]

I've been reading the _Imagine_ line for some time now, and much of the
terminology is the same as that found in the software _ProVista_.  Since
_ProVista_ and _Imagine_ can be integrated, I assume some of you have the
program.  If there is no such line, may I ask _ProVista_ questions on this
line?

                        Thanks!
                                                - [e-mail]

##

Subject: RobotCop Object
Date: Sun, 21 Apr 91 23:49:37 PDT
From: [e-mail] (Dave Coughran)

I got Mike Halvorson's RobotCop object from the Impulse club on
People/Link.  I have uploaded it to ab20 in the directory
incoming/amiga/3d/Imagine/OBJECTS.  I couldn't seem to send it
to hubcap, but some nice person might move it(?)

This is the chrome RoboCop that was in one of the pictures in the
Amiga World article on 3D package.  It's a big object, I know you can
render it in 6 megs, I don't know about less.

enjoy.


##

Subject: Re: Flaming Lasers and Rundgren Video
Date: Mon, 22 Apr 91 11:01:41 EDT
From: Mark Thompson <[e-mail]>

Matt Feifarek writes:
>I just saw the Todd Rundgren Video done with the Toaster.  Pretty Neat,
>there were quite a few cool effects.There was one part where a bump map in the
>shape of his face pushed out of a plane and then faded into color...

If you mean the cube that his face pulsated out of, that was an image map
on a morphed object, not a bump map.

> Anyway, there was one effect that stumped me...
> There was this flying street light ?!? ( stop light --- not street light)
> that emitted a beam of light that you could see... almost like a mist coming
> out of it.

This was done using semi-transparent cylinders with clear edges. Transparency
control on object edges is very handy.

> Can anyone think of a good way to do this in imagine?  A plain cylinder would
> be okay, but the edges would be hard, and it wouldn't look quite right...

and Ed Chadez writes (regarding lasers):
>However, it may be more convincing to actually build these objects in
>Imagine.  With the abality to create translucent objects with transparent
>bit-mapping and textures that can be seen thru, it shouldn't be too
>difficult.

Identical problems. Since imagine doesn't support edge transparency, you
must simulate it with transparency mapping. The problem is that is that
since you are dealing with a surface and not a volume, you cannot define
a cylindrical density function that will be opaque at the center and clear
at the edge. Therefore, even if you created a transparency map that defined
the cylinder edges as clear and the inside opaque, what you would see is a
totally transparent cylinder (ie. nothing) because it is the surface that is
rendered and the surface is clear. I ran into this problem in Lightwave
when I created a rocket blast that was transparent at the bottom and nearly
opaque at the top. When viewed from the side, it looked great, but when the
camera moved below looking up at the blast, it disappeared. So, for a laser
glow in Imagine, you could create a transparency map that goes from clear
to opaque to clear and wrap that on the side of your cylinder but the effect
will be view dependant. It still looks magnitudes better than a hard edged
tranlucent cylinder.
|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
|      `       '        Mark Thompson                 CONCURRENT COMPUTER  |
| --==* RADIANT *==--   [e-mail]        Principal Graphics   |
|      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   |
|     Productions       [phone] [phone]   & General Nuisance   |
|                                                                          |
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

##

Subject: Re: Flaming Lasers and Rundgren Video
Date: Mon, 22 Apr 91 08:27:56 -0700
From: [e-mail] (Ed Chadez)

Mark, et. al.

regarding lasers (and other translucent objects): what if you have two
cylinders, one inside the other?  The outside one would have your
transparency map (or texture or whatever) and the inside cylinder would be
a bright (or better yet, "light" object)?  The inner one would be much
smaller than the outer one, and just might produce the effect we'd like.

-Ed.

--
--//---------------------------------------------------------------------------
\X/ [e-mail]/Edward Chadez                    CARL Systems(303)861-5319
-------------------------------------------------------------------------------

##

Subject: Imagine Compendium (Free!) on abcfd...
Date: Mon, 22 Apr 91 11:37:32 EDT
From: Sandy Antunes <[e-mail]>

Hello folks!

    The latest version of the FREE Imagine Compendium is done.  This is a
collection of many of the postings of this list, put in some sort of order and
edited down a lot.  It's still getting close to book length-- around 30 pages
of material, in total.  I've changed the organization slightly over the
previous version, hopefully for the better.

    It is posted (of course) to abcfd20.larc.nasa.gov (128.155.23.64)
in the /incoming/amiga/3d/Imagine/TEXT directory under the name
ImagineComp.91b.Z.  Yes, the .Z indicated it is compressed using the unix
compress command.  Otherwise, it'd be too big.  It includes everything from
the 91a version, so I deleted the 91a version from that directory.  Overall,
with the upgrade of Imagine to 1.1 and such, the compendium has doubled in
size and hopefully doubled in usefulness.

    Included are: More then anyone wants to know on Brush Wraps (including
Steve's long treatise), the infamous "Art of Glass", the info and reviews
of Surface Master (yea!) and the Centaur tape (boo!), info on TTDDD usage,
three projects, info on dumping to videotape, a lot of Attributes, and even
more good stuff.  More packed with good stuff then an egg is with meat.  :)
I've included the index seperately in the same directory, for those who want
to know what it contains without downloading the entire thing.

    As before, if anyone cannot ftp it (or can't uncompress it),
feel free to write me and I'll email you a copy, as long as you're sure
your mailer won't choke on a 1.5 meg file... or maybe I'll get smart and
break it into pieces.  Either way, write me if you have problems.

                        sandy
-------------------------------------------------------------------
[e-mail]     Sandy Antunes, El Loco d'Waupelani
  it's 2am... do you know what time it is?      -------------------

##

Subject: Re: Flaming Lasers and Rundgren Video
Date: Mon, 22 Apr 91 12:11:36 EDT
From: Mark Thompson <[e-mail]>

> regarding lasers (and other translucent objects): what if you have two
> cylinders, one inside the other?  The outside one would have your
> transparency map (or texture or whatever) and the inside cylinder would be
> a bright (or better yet, "light" object)?

Yes, this is exactly how I have implimented them in the past. The real
trick though is still getting a good glow effect on the outer cylinder.
I have done similar things to create glowing flames (luminescent object
inside a translucent one). That was how I did the candle flame in my
'Office' pic that was on ab20 and the jet exhaust in the 'F15' image.
|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
|      `       '        Mark Thompson                 CONCURRENT COMPUTER  |
| --==* RADIANT *==--   [e-mail]        Principal Graphics   |
|      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   |
|     Productions       [phone] [phone]   & General Nuisance   |
|                                                                          |
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

##

Subject: I need a object lession! ;-)
Date: 22 Apr 91 13:19:00 EST
From: "SYSTEM MANAGER" <[e-mail]>

'ello fellow 3ders...

I am trying to create a Starship Enterprise object.  At first this
seemed simple since this object contains a disk and three cylnders.
However, for me, this is proving to be a much harder project than
I previously imagined.  (grin).

I thought the first thing I would do is use a primative disk object
for the saucer section.  I thought that I could use the magnetic
functions to 'puff up' the center of the disk on both sides of the
cylnder.  However, this does not seem to be a workable solution.
as there are no points on the disk to be effected by the
magnetic effect.  How does one effect the way a primative looks?

So I decide that I will come back to the saucer section, and so
I decide to build the nacelles for the engines.  I used the
cylnder primative.  Of course I have discovered that I can't really
do much to this object except to scale and rotate it.

Primatives aren't that primative aren't they?

Since it appears that I can't really use the primatives to build my
model, I ask how do you build something like a saucer section or
the nacells?  If I try to just to use the add lines and create an
object by hand, I get one perspective to look sorta right, but not
the other two.

I would sure appreciate a good tutorial on how to make objects that
are not regular shaped.  Doing pawns, glasses, etc. is understandable
since these are regular objects.

Also if you creating a complex object do you build little pieces of
the object and put it together later?  Or do you attempt to build the
entire object in the detail editor at the same time?  For example,
should I just have the saucer section loaded while I am working on it?
Or all the parts?

On another topic:  When using the primative disk, I rotated it and then
decided that I wanted to make the disk thicker.  So I used the extrude
function.  Extrude only seems to work in the X direction, is this true?

Am I expecting too much of myself to be able to create objects and
ultimately pictures only after spending about 40-50 hours playing,
doing tutorials, etc.?

 -mark=

 +--------+   ==================================================
 | \/     |   Mark D. Manes   "Mr. AmigaVision,  The 32 bit guy"
 | /\  \/ |   [e-mail]
 |     /  |   [phone]    "Make up your own mind! - AMIGA"
 +--------+   ==================================================




##

Subject: Acceleration Problems
Date: Mon, 22 Apr 91 23:07:14 PDT
From: [e-mail]

Hello folks. Hopefully someone out there can help me. What I am trying to
do is accelerate a rotation. What I really want to do is have a sphere
begin still then start spinning, with acceleration of course. Unfortunately
we have only been given the option of acceleration/deceleration along a
path, not for tweening or effects. So how do I do this. One option is to
break up the alignment bar frame by frame, figuring out how much it should
have changed from one frame to the next along an acceleration curve.
Personally, I'm not fond of that idea. I don't feel like doing the calculations
and it would take a lot of work to make entries for every frame.

I did figure out a second option, which has led me to discover a strange
bug. My idea goes like this: If I want this sphere to accelerate in it's
spinning and I can only accelerate along a path what do I do? I decided to
make a closed path that encircles the sphere at a common distance all the
way around. Attach a track (axis) to the path, aligning Y to the path, and
have that track move around the path, with acceleration. Lastly, tell the
sphere (in the center of the path) to align to the track. The axis of the
sphere SHOULD follow the track around the path, also following it's
acceleration. Right? Kind of.

There are so many different bugs happening with this situation I don't
know where to start. Also, the problems don't seem to be consistent.
First of all, if I look at the set-up on a frame by frame basis,
(i.e. look at frame 1, next frame, next frame) I can see that the axis
of the sphere is doing exactly what I want it to do. The axis of the
sphere rotates to follow the track. However, if I put the camera
directly above the sphere (so I can watch the true rotation) and do a
"make" to watch the wireframe animation, it doesn't do that. It begins
spinning fast, slows down and then spins the opposite direction.
Not what it looked like it was doing on the frame by frame basis.

There is an entirely different problem as well. To be perfectly honest
I decided to substitute a smaller sphere for the track in my initial
tests. This way I would be able to watch the small sphere go around
and see if the large sphere was following it. The bad news is, the
sphere following the path, the small one, doesn't do what it is
supposed to either. I'm sorry, I can't give you specifics on the
numbers I used, I tried so many different combinations. But I can
tell you that I always had it follow path for the entire length of
the animation (60 frames). Sometimes the (small) sphere would start
fast, slow down in the middle and go fast again; sometimes it would
start fast then go very slow at the end. Sometimes it would start
going backwards slowly, then take off fast in the correct direction.
During all of these variations the only thing I changed was the
acceleration, starting speed, deceleration and ending speed. I pride
myself on being meticulous during debugging, but I could not find any
consistency in the variations that occured. In all cases, it would never
make it all the way around the closed path. Again, if I looked at the
set-up on a frame by frame basis, the (small sphere) would ALWAYS make
it the all the way around the path with what appeared to be a regular
speed. This is true no matter what settings I had made in the action
editor. In all situations the animation did something different when
the wireframe was made then how appeared on a frame by frame basis.

I can upload the scenes to some site if someone wants to take a look
at it. I would appreciate it if anyone would try this out so I know
I'm not crazy, the results are completely bizarre. Does anyone have any
suggestions on any other way to achieve my accelerating rotation?
I do apoligize for the length of this posting.

By the way, I'm running Imagine 1.1 FP version on a 25Mhz 3000 with
10MB of memory.

=======================================================================
Sean Schur                        USENET: [e-mail]
Assistant Director Amiga/Media Lab        Compuserve: 70731,1102
Character Animation Department            Plink: OSS259
California Institute of the Arts
=======================================================================


##

Subject: Rotation blues
Date: Tue, 23 Apr 91 09:07:20 EDT
From: [e-mail]

Sean, your method of accelerating a sphere rotation sounds very
reasonable- its the way I would do it myself.

The symptoms of your malady- OK for slow speeds, then can slow down
and even reverse itself, sounds exactly like a strobe effect. If the
sphere rotates 340 degrees per frame, your eye will interpret this as
a rotation of -20 degrees per frame. For slow speeds, everything is
OK, but high speeds would cause this angle aliasing. This effect is
sometimes called the "wagon-wheel" effect since in old westerns the
wagon wheels would look like they're going at weird speeds.

This might or might not be the real problem- you've been trying to
debug this project, so you would have probably noticed it. It's
something to check, anyway.

If this is not the problem, the way to go about fixing it is in steps.
IGNORE the spinning sphere. Get the track to work. Replacing the track
with a real object (a sphere) was an excellent idea. How do you have
the path set up? A closed path that coils around and around like a
tight spring? Or did you use one loop and place N position bars in the
action editor?  If you placed N position bars, remember acceleration
and velocity won't "carry on" to the next bar. Also, the closed path
starts and ends at the exact same point. A 4 frame circle will NOT go
0, 90, 180, 270 degrees, it will go 0,120,240,360 degrees.

If you're using the coiled spring (making it must have been fun), I
don't really know how it would be going wrong.

Remember, get the track working. Ignore the sphere, it's just going to
confuse you. If the track works, then the sphere should be easy.

-Steve

---------------------------------------------------------------------------
Steve Worley                                        [e-mail]
---------------------------------------------------------------------------


##

Date: Tue, 23 Apr 91 13:45:10 mdt
From: Steven Lee Webb <[e-mail]>

In response to:

>=======================================================================
>Sean Schur                                     USENET: [e-mail]
>Assistant Director Amiga/Media Lab             Compuserve: 70731,1102
>Character Animation Department                 Plink: OSS259
>California Institute of the Arts
>=======================================================================

Hello fellow frusterated animator!

I've posted the same gripe about the implementation of acceleration in imagine
and I'v got a 'messy' fix for it...  What you want to figure out is the rate
of acceleration.  This can be easily calculated by the use of a simple sine
wave.  Basically, the chart for an acceleration path is similar to the one
roughly graphed below.

          Y                    a constant speed
          |                ........................
          |             ..                          ..
          |           .                               .
   speed  |          .   <-- acceleration              .  <-- deceleration
          |         .                                   .
          |      ..                                       ..
          |  ...                                             ...
     -----+-------------------------------------------------------X
          |    frame #

If you want to accelerate to a rotation speed of say 20 UPF (units/frame)
over a series of 23 frames, this is what you would do.

Total number of acceleration frames: 23
                            Max UPF: 20

First, divide 90 by the number of frames that you wish to accelerate along.
                               90/23=3.91304
                     This is your incremental factor.
      Frame#        sine factor        Percentage inc.    Rotation increment
------------------------------------------------------------------------------
        1               0                   0                   0
        2               3.913               .0682               1.3648
        3               7.826               .1361               2.7233
        4               11.73               .2034               4.0691
        5               15.65               .2697               5.3959
        6               19.56               .3348               6.6975
        7               23.47               .3984               7.9680
        8               27.39               .4600               9.2013
        9               31.30               .5195               10.391
        10              35.21               .5766               11.533
        11              39.13               .6310               12.621
        12              43.04               .6825               13.651
        13              46.95               .7308               14.616
        14              50.86               .7757               15.514
        15              54.78               .8169               16.339
        16              58.69               .8544               17.088
        17              62.60               .8878               17.577
        18              66.52               .9172               18.344
        19              70.43               .9422               18.845
        20              74.34               .9629               19.258
        21              78.26               .9790               19.581
        22              82.17               .9906               19.813
        23              86.08               .9976               19.953

               (I told you that this was going to be messy)

The equations for the above are not as difficult as they look.

Sine factor     = incremental factor * (framenum-1)
Percentage inc. = sin (Sine factor)
Rotation inc.   = Max UPF * Percentage inc.

Then simply plug those numbers into your alignment variables for your object
in the action editor.  Notice that you will end up with lots of single dots
in your action editor for the alignment, instead of having one steady line,
but this gets the job done.  Good thing Newton invented these functions, huh?
     _________
    /  ______/\
   /  /\_____\/    ---------------------------------------
  /_____   /\            It's clever, but is it art?
 _\____/  / /      ---------------------------------------
/________/ /
\________\/teven Webb  Reply to: [e-mail]
           A Computer Science Major @ Colorado State University
            I have an amiga 500 with 5MEGS RAM, and a 50MEG HD.
                   [why are you looking at me so funny?]
                     (Yes, I HAVE voided my warranty!)


##

Subject: Re: Acceleration Problems
Date: Tue, 23 Apr 91 13:07:56 PDT
From: "Jim Lange" <[e-mail]>

In-Reply-To: WRPYR:[e-mail]'s message of 04-22-91 23:07

In response to Sean Schur's rotating sphere problem:

This sounds like a good candidate for a videotape log sent to Impulse (as they
recommend in their newsletter).

Here is another way to accelerate a rotation.  Create a small circular path
(can be an open path, but you'll need more than two axes), then scale it so it
is very small, then attach your sphere to this path and set the alignment bar
to 'align to path'.  The idea is that the sphere will not move perceptibly from
it's fixed position, but it will rotate as it follows the path.  You can then
apply whatever acceleration you wish.  I am using this technique for a swinging
pendulum effect.

Actually the path doesn't have to be tiny to eliminate non-rotational motion,
just offset the sphere axis to match the radius of the curve.  Thus if the
curve radius is 1 unit and counter clockwise, then the sphere axis should be
moved 1 unit on the X axis.  The Y axis will then align with the path and the
true center of the sphere will remain at the center of the path.

I have been seeing many strange effects (bugs) relating to paths also.  I have
been experimenting with acceleration to simulate gravity and harmonic
motion (I will post a summary when I finish working the kinks out) and have
found that if I try to rotate or scale a path during an animation, the objects
that are attached to the path sometimes move all over the place very wildly!
However, I can see this in the stage editor also as I move from frame to frame.
The path actually moves away from its axis--very strange.


-------------------------------------------------------------------------------
Jim Lange                [e-mail]
Oracle Corporation            {uunet,apple,hplabs}!oracle!jlange
-------------------------------------------------------------------------------


##

Subject: Re: Imagine 2.0
Date: Tue, 23 Apr 91 11:31:59 PDT
From: "Jim Lange" <[e-mail]>

In-Reply-To: WRPYR:[e-mail]'s message of 04-21-91 15:09

In response to Steve Worley's call for an Arexx interface to the Project
editor--I agree that this is where it should be added as a minimum.

In the meantime, I was considering trying ScriptIt, a PD (shareware?) scripting
program that can be used to control non-Arexx programs by manipulating their
menus and gadgets.  I currently don't have the program, but it looks promising.
ScriptIt does have an Arexx port so could be integrated easily with other Arexx
controllable programs as suggested in Steve's original message.  Has anyone
used this program?  (I originally read about ScriptIt in Amiga Transactor over
a year ago.)

-------------------------------------------------------------------------------
Jim Lange                [e-mail]
Oracle Corporation            {uunet,apple,hplabs}!oracle!jlange
-------------------------------------------------------------------------------


##

Subject: Re: I need a object lession! ;-)
Date: Tue, 23 Apr 91 13:58:23 PDT
From: "Jim Lange" <[e-mail]>

Mark,

Learning all the techniques and tricks in any 3D modeler takes time and
patience (certainly more time that I have), so don't be discouraged.

For the saucer section of your Enterprise object, I would suggest using spin.
Your profile would look something like this:

            __
              -------.
                      \_________
                                \
               _________________/
            __/
            ^
            |
            spin with axis here

That way you have much more control of the contours and can determine the
number of polygons if you need to tweek it later with magnetism (or other
tools).

If you need to increase the number of polygons in an object (such as your disk
primative), select the object, switch to pick faces mode, choose select all
(all faces will turn blue), then select fracture.  All triangles will be
subdivided into smaller triangles.  If you need to subdivide a smaller area use
lasso to select just the faces you need instead of 'select all'.

I'm not sure what the nacelles are (I don't follow Star Trek that closely), but
don't underestimate the power of the Forms Editor.  The tutorial is very week,
but it pays to spend some time to learn what it can do.  You could also use it
to model your saucer section and have much better control over the final form
than with spin.  Just set symetry to the option that coordinates the front and
right views (can't remember the name--90 degrees, right angles, something like
that).  The best way to think about the Forms editor is that the top view
defines the "shape" of the cross-section from top to bottom, while the other
views describe the profile when viewed from the front and side.

As far as assembling the ship, yes, the best plan is to create each part
separately, then connect them in a group.


-------------------------------------------------------------------------------
Jim Lange                [e-mail]
Oracle Corporation            {uunet,apple,hplabs}!oracle!jlange
-------------------------------------------------------------------------------

##

Subject: Xit/Scripit
Date: Wed, 24 Apr 91 13:04 PDT
From: [e-mail] (Scott Busse)

Jim, I did some work with Scripit and Turbo Silver a while back, and although
somewhat tedious to set up(like all programming :), I had almost total succes
with what I was trying to do. I basically mapped the turbo screens layout, and
the gadget window layouts, so that I could simulate the intuition events for
everything. It worked, all except for the ability to make the Amiga-V or find
requester go away. I could make it work if I started my script there (at the
Amiga-V) but if I started on the main screen, it would hang every time. I
eventually gave up and used the Amiga-N event (next object) to step through and
find the object I wanted to work on, but then I had to know the current order
of the objects in the cel. I talked to Mike Halvorson about this just a short
while ago, and he said it was necessary to pad the gadget buffer with spaces to
the max size in order to make it go away.
   Sorry if I'm springing too much detail too soon, but if you're going to
attempt a Scripit port, then it may help at some point to know this. I do plan
to get back to work on such a port (Imagine/Scripit) as soon as I have time, so
lets keep in touch.
   Of course anyone else out there who is interested is welcome/invited to join
in! We could make up for a big AREXX hole that Impulse is leaving in Imagine...
--
* Scott Busse email:           O    O   O_     _      ___ .....
* CIS 73040,2114              |||  /|\  /\   O/\_     /         O    )=|
* [e-mail]    l   | |   |\    / \   /\                _\
* [e-mail]                  Live Long and Animate... \


##

Subject: ab20 archives
Date: Wed, 24 Apr 91 16:27:00 EDT
From: [e-mail]

The ab20 archives have been reorganized yet one more time (yesterday).


I cannot find the Imagine files that were there- I looked for quite
a while.

No worry, though.  I believe ALL of the text, objects, brushmaps, etc.
can still be found on hubcap.clemson.edu in /pub/amiga/incoming/3D/IMAGINE.
Our friend, Doug's, FTP site. Thanks, Doug.

-Steve

---------------------------------------------------------------------------
Steve Worley                                        [e-mail]
---------------------------------------------------------------------------

##

Subject: Re: ab20 archives
Date: Wed, 24 Apr 91 16:48:05 EDT
From: [e-mail] (Doug)

> The ab20 archives have been reorganized yet one more time (yesterday).

Yeah, I went there to find the imagine.comp.91b thing

>
>
> I cannot find the Imagine files that were there- I looked for quite
> a while.

I checked FILES.Z and all.. Perhaps they are being shuffled now in a
private directory (Pretty sure thats it).

>
> No worry, though.  I believe ALL of the text, objects, brushmaps, etc.
> can still be found on hubcap.clemson.edu in /pub/amiga/incoming/3D/IMAGINE.
> Our friend, Doug's, FTP site. Thanks, Doug.

Aw, gee.  Alas, I am just a moderator (thats the only way clemson would
give the AUG a pub directory... with bible belt and nude gifs and all -
true!).  Anyway, I was afraid of this happening before I got a chance to
download stuff, hence the "up to dateness" of hub.

>
> -Steve

I have been meaning to ask you, why haven't you (or anyone) uploaded
any pictures? (at least, I haven't found any). I am really interested
in seeing others work (I though that by uploading that simple experiment
others would do so but NOOOOOOOOOOOOOOOOOOOOOOOOO)

PS.  One more week and.... FINALS ARE OVER weeeeeeeeeeeeeeeeeeeeeeeee



##

Date: Thu, 25 Apr 91 01:20:35 -0400
From: [e-mail]

The followin is what I posted to a neighborhood BBS concering
the 040 board tha we played with at MITE AVISTA.

Well, today was pretty eventful. I have some
info pertaining to the Production 91 show at
place Bonaventure this summer. It will be as
big as last year. Maison du Logiciel will be
there with a rather large booth (40 x 20) and
will bw showing off lots of neat A/V equipment
(and Amigas of course). AmiLink will be there
(the editing software which runs in parallel
with the toaster). MITE AVISTA will be
associating with SONY and Citi electronique.
We will most probably have 2 workstations in
the Citi area. We will be demonstrating what
multimedia is all about using AmigaVision,
laserdiscs, the toaster and associated
software/hardware. TLA Multimedia (my
company) will most probably be giving
seminar(s) on the mezzanine for the
technically minded (probably AmigaVision).



Also, many users of the lab (including me)
had the chance to put an actual 68040 board
thru its paces today. It was a prototype board (I
have never seen so many little wires tightly
knit onto a circuit board!) from RCS
Management. Here are some points about
what we saw

- For the A2000. RCS says an A3000 board
will be available from them this summer. The
market (50 - 100 000 A3000s) is not big
enough yet.

- Takes over machine (does not run in parallel
with 68000 (which *is* possible by the way)).

- 32 MB of SIMMS can be added directly to the
board. Future add-on will allow for 64 MB.
(***THIS IS IN AN A2000, DON'T
FORGET***) Our test was on a 16MB version (
I never thiught i'd see the day when the a2000
would have more than 9MB RAM)

.- We couldn't formally benchmark the
machine. Although we ran Imagine 1.1 and
Lightwave (at the same time BTW) and the
results we excellent. The shaded perspective
updates were at least twice as fast as our
A3000. Lightwave showed even more
impressive results rendering the Utah Teapot
in Hires (with background) 24bit in less than 1
minute.

Again, this was not a formal evaluation of the
board but just a "wow, it exists, it's fast".

Interesting to note that 2 people on the design
team are from Concordia University. RCS will
be receiving their production boards within 2
weeks and will be shipping then. Suresh (the
president) is a *very* nice person. It seems
that AWolrd screwed up their add in their
magazine (they put the wrong one in...).

PRICE: He said for a 040-4MB RAM version:
$3 400 American

TO NOTE: SURESH said that his 040 board
for the A3000 will run concurrently with the
CPU on the motherboard. That's right boys
and girls, *parallel processing*. Why waste
the 7 MIPS coming from the 68030? He said
that it wasn't worth it for the 68000...

TO sum up: it works. It's fast (how fast we're
not sure, but much faster than A3000). Will be
shipping in less than a month. Nice people.
Kinda makes me want to ditch the old 3000
and get a spanking new A2000. Naw, I'll wait
for the 050 for my 3000... Ciao dudes...



                                            LePew




##

Subject: objects flyin' everywhere!
Date: Thu, 25 Apr 91 10:43:01 EDT
From: [e-mail]

Relay from S. Menzies:

===============================

I've been reading alot inn this group about objects flying
off course in the middle of an animation. I too, have had a
lot of this and while I don't have an inteligent explanation nor
a "fix all", I do have something that may help:

    I have had many instances where I have wanted an object
or a grouped object to either make a simple rotation ( w/rotation F/X),
or have interactively rotated the object and used the keyframe
capability on the alignment "bar".  Often (not always) the object
actually "flys" away from it's own axis rather than rotating with it.
I have been able to stop this "flying away" by making a small incretment
of only .0001 of a unit in the position bar (action editor) on one of
the x,y,z axis (sometimes 2 axis!?) from the original position of the
object. This works! (I don't know why, but it does:)

This brings me to another point. Imagine's animation capabilities are
so bad (even if their wasn't any bugs) that I think we should begin
some serious discussion on what makes a good animation package and hope
they (Impulse) catch wind of it. It's my personal opinion that Imagine
is capable of little more that robotic movement, loops and fly thru's
and this is animation in it's most primative state. Why? Well, for
many, they know why and for other's I try a simple explanation:

    All movement in Imagine is linear. This means that when you
perform any kind of transformation (translation, scaling, rotation)
or morph (interpolations, attributes) from point A to point B, you
kiss the object goodbye at "A" and wait for it's arrival at "B" (if
it doesn't fall apart on the way:). You have no *control* over HOW
it gets there. In otherwords, how expressively this motion task is
performed by the actor (charcter) or how the actor "changes over time".
    Take for example Imagine's morphing "on a channel" or it's
new 1.1 keyframing capability. What Imagine does is divide the transformtion
equally over the chosen # of frames. Imagine gives the same amplitude to
the motion or movement in the first frame as to the middle frame as to
the last frame. You may try to discretly add more keyframes "around" the
extremities of the motions (cycle editor or stage editor) to try and
soften or curve the angularness at the extremes but an analogy or
visual equivelance of this would be adding more polygons to the structure
of an object to create a smoother curvature but stll render it in FACETED
mode. Can you imagine what your objects would look like (and the # of polys)
if we didn't have phong shading?? Well thats what happening to our
animations.
    We do have velocity control though. Unfortunately they couldn't
have made it any more limited. It cannot be used except with a path and
it is almost impossible to coordinate with other transformations. Talk
about foot slip!

So what do we need?? I have a few suggestions.
1> keyframing at the minimum *must* create an internal spline.
    (I believe Lightwave does this for translations atleast)
2> Throw out the Action editor. It's unfriendly :
    a> visually, colored bars represent nothing in animation(ie amplitude);
    b> it's completely useless when atempting to coordinate an object
           at the top of the chart with one 25 down (you can't see them
           at the same time)

3> key framing *must* be correctable on the same screen it's performed
   on.
4> Easier and faster path construction. Path construction should not
   be a separate mode that momentarily erases the sceen on a zoom. Spline
   translation paths should be automatically built when keyframeing an
   objects translation. (visual and be easily editable).
5> all transformations (everything from lights to attribute morphing)
   should have its amplitude or relationship of movement and time,
   represented by a editable function curve (ala JourneyMan and almost
   all highend software). Lightwave also has some function curves, called
   envelopes, I think. Not for everything though (lights, morphing and
   I'm not sure what else but they plan to have it for everthing).
   Lightwaves function curves are very primative though, small and
   awkward to work with and only one at a time.
6> If (5) happened, you could toss out the cycle editor also as there
   most likely could be away to extrapolate a cycle from the function
   curve.

Any Comments? I hope so...it is important. I've noticed 2 common threads
in this group. A lot of success with object construction/rendering and
some very bad experiences animating. I agree with you. I've spent
several years working with 3D on the Amiga and I stand very disappointed
with the animation tools. Is Impulse listening? I wonder? After all,
they don't seem to care enough to send us news of an update...boy, is
that gonna hurt them!

cya -steve
--
Stephen Menzies
Computer Animation Professor,
Department of Cinema,
Concordia University
Email: [e-mail]


=====================================================

-Steve

---------------------------------------------------------------------------
Steve Worley                                        [e-mail]
---------------------------------------------------------------------------


##

Subject: Re:  Acceleration Problems
Date: Fri, 26 Apr 91 14:57:30 -0400
From: [e-mail] (Stuart L Williams)

This sounds like a case of TEMPORAL Aliasing!!!  This is the old wagon-wheel
syndrome.  Remember in the old western, when the wagons would take off, wheels
spinning, and the film would show them accellerating and decelerating.  This
is caused by the sampling rate of the film, in the wagon-wheel case, of 24 fps
catching motion that went beyond 1 revolution in a 24th of a second.  Got it?

What I think is happening is that Imagine doesn't motion blur the object as it
renders them.  What you have is a snapshot of what is happening, an instant not
a duration.  Thusly what you want to see isn't happening in the animation.

Impulse could correct this by including motion-blurring into the software.
There are several ways to do this, but all of them add considerably to the rendering times, as the engine needs to figure out where the object is over time,
and to create a blurred image of that object.

This is the only real way to capture rotational motion that exceeds one
revolution per frame.

On a side note, here are some further suggestions for improvements to Imagine.

    *Real time feedback, switchable on/off, for the wireframe cam view
    *Motion-blurring for all moving objects
    *Improved acceleration handling
    *Import/Export file formats to/from CAD programs directly
    *Water, Water Everywhere

These are just a few thoughts, so please, no flames.  I think that this is a
great piece of software, and it is making lots of heads turn to our fav machine.

PS..    I think that they should add depth of field effects as well.


##

Subject: Pic on hubcap
Date: Fri, 26 Apr 91 16:08:50 EDT
From: [e-mail]

Someone asked me to upload a pic showing my repeating sidewalk. I
uploaded my test image [really a project in it's own right]. It
is a street scene with a friend of mine painted on a wall. It's
an overscan HAM image.

The pic is at hubcap.clemson.edu (130.127.8.1)  in
/pub/amiga/incoming/3D/IMAGINE/PICTURES and is called punk.lzh. This
directory also contains Stephen Menzies' "Exploded Garden" and
Doug Dyer's "Metal Dreams".

My "Ocean Sunset" has to be rerendered... I noticed half way through
the 40 frame anim  that I misordered my wave pictures, so the
waves are going backwards. Arggh! :-)

Upload your pix too!

-Steve

---------------------------------------------------------------------------
Steve Worley                                        [e-mail]
---------------------------------------------------------------------------


##

Subject: 24-bit display
Date: Fri, 26 Apr 91 19:58:16 EDT
From: [e-mail] (Alan Price)

This is my first entry to the mailing list (or anywhere for that matter)
but I have been receiving it for a while now and have found many enlightening
bits of info. The Imagine Compendium that Sandy compiled is an incredible \
resource! (Thankyou, Steve for telling me how to use DigiView on my A3000,
and the part on how to load grouped objects into the Cycle-editor in 1.0,
ala "Pose" in 1.1, was a revelation!)
    Anyway, the question I would like to get some feedback on is about
the 24-bit display boards available. In particular, a comparison between
DCTV and Colorburst. It seems that DCTV is quickly becoming the most popular,
but I'm wondering if that's truly justified. OK, DCTV's big advantange is
that it digitizes. But let's say I don't want that (I'll settle for DigiView
21-bit images.) What I'm most interested in is getting animations to play
real-time out to video in 24-bit (Imagine animations, ofcourse). 3 boards
contend to do this: DCTV, HAM-E, and Colorburst. I've cancelled out HAM-E
because of it's lower max. resolution and limit to 256K colors.
    At the moment I lean towards Colorburst because it is a RGB display,
and if I'm viewing still frames, painting, or touching up, I would MUCH
rather be looking at an RGB display than a mushy composite. If I understand
correctly, Colorburst output will go into any regular Amiga genlock/encoder,
so you've got both worlds. DCTV is composite only, and they 'say' that they
will have an RGB encoder coming out 'soon', but I have a feeling that t
NTSC to RGB decoder is going to be PRETTY expensive. (Plus I've also heard
some pretty rude rumors concerning the origin of DCTV's hardware and the
inability of the company to support it with much technical info. - something
about retired military techs making quick bucks and bad contracts - anyway,
it was actually an Amiga store sales clerk, and sometimes you just can't
those guys.)
    Anyway, if Colorburst can display it's full resolution (768x560) in
an RGB display and supports real-time animation in the same way as DCTV
(which is 736x480), why should'nt Colorburst be the best choice?
    Can an owner of Colorburst please give some advice? How about a few
of the DCTV owners? How about a well-to-do owner of both?



##

Subject: Slice
Date: 26 Apr 91 23:17 -0500
From: Colin Stobbe <[e-mail]>

Hello,
     I just got the Imagine Compendium 91b today, but noticed that there
isn't much about SLICE, one of my favorite features, so I decided to tell
all of you what I do with it.
     One of the objects I've been working on for the past while, requires
quite a number of 'decals' (I'm actually modelling a plastic model I have).
I however have had little success with brush maps in the past, however (it
was before I got on this echo).  When I got Imagine, with it's slice feature,
my problems were solved.  Now what I do, is draw the decal in a paint
package, import it into Imagine with the CONVERT IFF/ILBM feature in the
detail editor and extrude it.  What I've now got is basically a punch, which
I position partially inside the object I want to put the 'decal' on.  I next
SLICE them together.
     What I'm left with is usually four objects.  I've got the 'punch' in
two pieces, which I dicard, the origional object with a hole the shape of
the 'decal', and a piece of the origional object in the shape of the 'decal'.
     I've now avoided using a brushmap, but more importantly, I can change
the attributes of the 'decal'.  Insted of having just a flat color, I can
change the reflectivity, roughness, etc., and put textures and even
brushmaps on the 'decal', making it look quite a bit more interesting.
     The only problem with SLICE is that it takes memory, lots of it.  When
I only had 3MB, I ran out all the time, and even at 10MB I've run out of
memory.

          Colin Stobbe
          [e-mail]
               From the University of Manitoba, Canada

##

Subject: AlienChessboard picture
Date: Sat, 27 Apr 91 02:16:33 -0400
From: Udo K Schuermann <[e-mail]>

I've just uploaded to hubcap.clemson.edu what I consider one of my
better creations with Imagine: a raytraced picture of a rather strange
chessboard.
    The picture is 352x470 HAM and 125K big (the .lzh is 113K) and
took 2.75 hours on my 32 MHz 2000/68030.  It's the result of playing
with different orientations and colors of the wood pattern.  I was
hoping to get something not too far from a marble pattern for the
chessboard, but quite obviously I failed miserably.  The result is a
chessboard that looks quite interesting, although I'd hate to engage
in a match using that one :-)
    Let me know what you think!

Steve:  your sidewalk does look nice!  Funky picture, too!  I like it.

 ._.  Udo Schuermann      "Does he talk?  DOES HE TALK?" -- "Of course I talk.
 ( )  [e-mail]   I'm the Minister for Internal Affairs." -- M.Python

##

Subject: Has Anyone Made Fonts?
Date: Sat, 27 Apr 91 11:35:39 PDT
From: [e-mail]

I need to do some lettering work. I was wondering if anyone out there had
completed any font sets for Imagine/Turbo Silver. I could do it myself,
and will probably end up doing that, but I thought I'd try to save some
time and see if anyone else out there had done some already. Making
fonts can be time consuming, even with IFF/ILBM convert.

Also, I heard somewhere that there is a program that is supposed to come
with TTDDD to convert fonts. But it didn't come with the version of
TTDDD that I downloaded off a Fish disk. I think the program is called
TSTex or something like that.

=======================================================================
Sean Schur                        USENET: [e-mail]
Assistant Director Amiga/Media Lab        Compuserve: 70731,1102
Character Animation Department            Plink: OSS259
California Institute of the Arts
=======================================================================

##

Subject: Re: Has Anyone Made Fonts?
Date: Sat, 27 Apr 91 15:24:55 -0400
From: Udo K Schuermann <[e-mail]>

> I need to do some lettering work. I was wondering if anyone out there had
> completed any font sets for Imagine/Turbo Silver.

Since Imagine has this nice ILBM outline converter (which has some
accuracy problems with small brushes) I've never bothered to create
font objects.  I have 65 different fonts in my fonts: directory, in
lots of sizes (2.8 Megs all) -- using DPaint, it takes only a few
minutes to create myself a custom object, not much more for a beveled
one.
    What I'm basically saying is that it's perhaps better to draw
on your resources in fonts: and on DPaint as you need them, rather
than to largely duplicate what is readily available.
    No?

 ._.  Udo Schuermann      "Does he talk?  DOES HE TALK?" -- "Of course I talk.
 ( )  [e-mail]   I'm the Minister for Internal Affairs." -- M.Python

##

Subject: Fonts
Date: Sat, 27 Apr 91 20:27:28 EDT
From: [e-mail] (Rob Borsari)

I have had good results using ProPDraw to create the word on screen then using
snap to capture them into imagine.  Since the fonts are Scaleable the objects
are very smooth.    -R-
[e-mail]       Rob Borsari

##

Subject: Re: objects flyin' everywhere!
Date: Sun, 28 Apr 1991 22:51:36 GMT
From: [e-mail] (Stephen Menzies)

[e-mail] writes:


>Relay from S. Menzies:

>===============================
>[alot of stuff deleted from my original post]

>So what do we need?? I have a few suggestions.
>1> keyframing at the minimum *must* create an internal spline.
>    (I believe Lightwave does this for translations atleast)

To clarify, I mean that when an object's translation is keyframed
the software should at the minimum give you a choice as to whether
this motion ,thru 3 or more keyframe positions, should be smoothed
via an internal spline or angular (shortest route from oe point to
another).
    For simple keyframe translations through 2 or more positions,
adding a "hinge" object (I find the best position for the hinge is
often (too) hard to find. For many cases in camera translations, the
hinge works well when placed near the tracked object) may smooth the
keyframing. Needless to say though, considering the time spent and
poor results...you might as well have built a spline path and skipped
the keyframing.

>cya -steve
>--
>Stephen Menzies
>Computer Animation Professor,
>Department of Cinema,
>Concordia University
>Email: [e-mail]


>=====================================================

>-Steve

>---------------------------------------------------------------------------
>Steve Worley                                        [e-mail]
>---------------------------------------------------------------------------

cya -stephen
--
Stephen Menzies
Email: [e-mail]

##

Subject: Re: 24-bit display
Date: Mon, 29 Apr 91 09:15:13 EDT
From: Mark Thompson <[e-mail]>

>     Anyway, if Colorburst can display it's full resolution (768x560) in
> an RGB display and supports real-time animation in the same way as DCTV
> (which is 736x480), why should'nt Colorburst be the best choice?

The key question is can Colorburst support real-time animation. Unless it
comes with some sort of software to do some fairly dramatic animation/image
compression, the answer is no. At least not to the extent that DCTV does.
DCTV attains its speed by not working with 24bits, but a highly compressed
NTSC composite representation of 24bit data. Hence, the faster animation
rates and lower image quality. The designer of Colorburst says it can do
20 frame/s 24bit animation but only a very small amount of image change
(the equivalent of about a 100x100 brush).

>     Can an owner of Colorburst please give some advice? How about a few
> of the DCTV owners?

There are no Colorburst owners in the US other than a few registered
developers. I own DCTV and a Toaster. DCTV image quality is fairly poor
for hi-res RGB work, but acceptable for video. If I didn't already have a
Toaster, I probably would wait to see Colorburst operate before purchasing
DCTV.
|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
|      `       '        Mark Thompson                 CONCURRENT COMPUTER  |
| --==* RADIANT *==--   [e-mail]        Principal Graphics   |
|      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   |
|     Productions       [phone] [phone]   & General Nuisance   |
|                                                                          |
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

##

Subject: Re: objects flyin' everywhere!
Date: Mon, 29 Apr 91 09:43:08 EDT
From: Mark Thompson <[e-mail]>

> So what do we need?? I have a few suggestions.
> 1> keyframing at the minimum *must* create an internal spline.
>     (I believe Lightwave does this for translations atleast)

Lightwave allows any keyframe to be either linear or spline interpolated
and the spline interpolation can be done for translation, rotation, stretching,
and scaling of objects and camera/light motion. Also, all other keyframe-able
attributes such as polygon size, object morph, object dissolve, etc. can
be spline interpolated.
|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
|      `       '        Mark Thompson                 CONCURRENT COMPUTER  |
| --==* RADIANT *==--   [e-mail]        Principal Graphics   |
|      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   |
|     Productions       [phone] [phone]   & General Nuisance   |
|                                                                          |
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

##

Subject: Has Anyone Made Fonts?
Date: Mon, 29 Apr 91 08:26:14 PDT
From: [e-mail] (Glenn M. Lewis ~)

>>>>> On Sat, 27 Apr 91 11:35:39 PDT, [e-mail] said:

Sean> I need to do some lettering work. I was wondering if anyone out
Sean> there had completed any font sets for Imagine/Turbo Silver. I
Sean> could do it myself, and will probably end up doing that, but I
Sean> thought I'd try to save some time and see if anyone else out there
Sean> had done some already. Making fonts can be time consuming, even
Sean> with IFF/ILBM convert.

Sean> Also, I heard somewhere that there is a program that is supposed
Sean> to come with TTDDD to convert fonts. But it didn't come with the
Sean> version of TTDDD that I downloaded off a Fish disk. I think the
Sean> program is called TSTex or something like that.

    Yes, I am the author of TTDDD and TSTeX.  TSTeX takes TeX fonts
and converts them into 3-D objects that can be loaded into Imagine or
Turbo Silver after piping them through WriteTDDD (which comes with th
TTDDD package).
    Registered owners of TTDDD get TSTeX, SQuad (a superquadric
generator) and the source to ReadTDDD and WriteTDDD.

    To understand TSTeX, it is important to understand the algorithm
used.  Basically, TSTeX read the PK format of the TeX font, which is a
font that has already been prepared at a certain point size for a device
that has a given resolution.  The "pixels" that appear on the device are
translated into boxes in the 3-D object.  Horizontal scan-line
optimization is performed so that you will not end up with scads of
individual boxes that are touching each other.  Performing horizontal
and vertical optimization would take much more processing time, and I
didn't feel it was worth it.  Basically, that would have possibly cut
down on the final number of boxes, but not by much.

    The point of all of this is that the larger the font from TeX
you use, at the highest "DPI" (dots per inch) resolution gives you the
nicest font within Imagine or Turbo Silver.  Of course, those letters
also take up the most amount of memory.

    Therefore, there is a trade-off between object size and the
resulting "jagginess" of the font.  This is, of course, expected.

    Please feel free to send me e-mail if you have more questions
about TSTeX.

                            -- Glenn Lewis

[e-mail] | These are my own opinions... not Intel's

##

Date: Mon, 29 Apr 91 22:39:40 -0500
From: [e-mail]

I agree that Imagine's animation facilities are primitive compared to its other
powerful modules.  I have seen and used (to an extent) 3D Studio for MS-DOG
computers.  It's keyframer is incredible.  It should be though, it was written
by Dan Silva, who of course wrote the Dpaint trilogy (Yes, we did lose him to
IBM programming... thanks a lot Electronic Arts!).

Everything in the keyframer is splined!  EVERYTHING.  I worked on a project
with an IBM buddy that used LOTS of harmonic motion, and all we had to do was
set up the motion spline, the ease in/out an go.  Can you imagine trying to
make a convincing bouncing ball in Imagine (one that slows to a stop at the
apex of it's bounce)... it would be >rough< !

I believe that Imagine's rendering is much better than 3Dstudio, and the modeler
has more features (if not as friendly), but animation is sorry compared to it.

I agree that We should scrap the Action editor and cycle editor, and make a
keyframe editor combining the best of both: Keyframes can be made on F/X,
alignment, morphing, etc... WITH easing!!  Object cycles are a neat idea, but
it is only a step in the right direction.

The way 3D Studio ( which costs nearly as much as two Toasters, by the way )
does it is to have a dialog box that represents each frame (like Action Ed).  Each
object has the following channels (also like Action Ed) move, morph, align, s
size/scale (maybe more).  If a change is made at a frame, then a dot apears in
the dialog in the appropriate channel.  If you double click on a dot, you bring
up a requestor that controls the dot's function: eases, bias, spline magnitute,
etc. (This is what IMAGINE is missing!)  Every dot has this ability.  A cycle
could be copied over frames by copying a range of keyframes one after another
( a tool is provided to do this easily).  Dots do there thing from the status
of the previous dot ( in Imagine, dot to dot would be a bar ).

Dot's can be made in the dialog, or in the tri-view.  Any change on a frame will


Imagine has the dot thing sortof, which is fine, but it is missing the
transition definition (eases bias etc) and the friendly creation of keys in
the tri view... We are part way there with Imagine.  Imagine could be better
than 3Dstudio with hinge, FX, etc, as well as morphing etc. IF it cought up
on transition control.

Imagine 2.0???  :^)



##

Subject: Re: Has Anyone Made Fonts?
Date: Tue, 30 Apr 91 12:54:56 +0200
From: Marek Rzewuski <[e-mail]>

How much does it cost to get a registered copy of TTDDD?
and what shell I do to get one?

What do you convert the TDDD files to? I know that this is some
kind of humar-readable-ascii-file, but did you invent this fileformat
by your selv, or did you use someone elses? If you did, what is the
name of the fileformat, and where can I find some info about it?




##

Subject: Re: 24-bit display
Date: Tue, 30 Apr 91 09:47:22 EDT
From: [e-mail] (Bob Lindabury - SysAdm)

Mark Thompson <rutgers!westford.ccur.com!mark> writes:

> The key question is can Colorburst support real-time animation. Unless it
> comes with some sort of software to do some fairly dramatic animation/image
> compression, the answer is no. At least not to the extent that DCTV does.
> DCTV attains its speed by not working with 24bits, but a highly compressed
> NTSC composite representation of 24bit data. Hence, the faster animation
> rates and lower image quality. The designer of Colorburst says it can do
> 20 frame/s 24bit animation but only a very small amount of image change
> (the equivalent of about a 100x100 brush).

 From what I heard you can have lower resolution and lower bit-plane
images/buffers which could easily be animated in real time.  I would
be *very* happy with 8-bit hi-res images animated in real time.  I
think it IS possible to animate something like 512x482 8-bit images
in real time.

-- Bob

 The Graphics BBS  908/469-0049  "It's better than a sharp stick in the eye!"
 ============================================================================
  InterNet: [e-mail]                | Raven Enterprises
      UUCP: ...rutgers!bobsbox!graphics!bobl      | 25 Raven Avenue
    BitNet: bobl%graphics.rent.com@pucc           | Piscataway, NJ 08854
    Home #: 908/560-7353                          | 908/271-8878

##

Subject: Re: Has Anyone Made Fonts?
Date: Tue, 30 Apr 91 08:22:34 PDT
From: [e-mail] (Glenn M. Lewis ~)

>>>>> On Tue, 30 Apr 91 12:54:56 +0200, Marek Rzewuski <[e-mail]> said:

[Please forgive me, Imagineers, if this is not of general interest...  I
thought that the last paragraph might be interesting to some people - GML]

Marek> How much does it cost to get a registered copy of TTDDD?

Ten U.S. dollars.

Marek> and what shall I do to get one?

Send $10.00 to me with your name and address, and I'll send you the
registered copy of TTDDD.  In the meantime, ftp to ab20.larc.nasa.gov,
and get the file called "TTDDD.zoo".  (Search "FILES.Z" for its
location... things have been moving around lately on that site as the
administrators clean up the archives.)  That will give you a lot more
information about TTDDD (Textual Three Dimensional Data Description).
My address is in the archive.

Marek> What do you convert the TDDD files to? I know that this is some
Marek> kind of human-readable-ascii-file, but did you invent this
Marek> file format by yourself, or did you use someone else's?

    Yes, it is an ASCII file that can be edited with a text editor
or created algorithmically.  Although the files tend to be large for
complex objects, they are quite malleable by people and programs.  Yes,
I invented the format myself.  The TTDDD parser ("WriteTDDD") is quite
versatile and is not terribly restrictive.  To read a TTDDD file
generated by "ReadTDDD", *your* parser does not have to be as complex
because ReadTDDD always outputs TTDDD in the same format.

Marek> If you did, what is the name of the fileformat, and where can I
Marek> find some info about it?

    Take a look at the Freely-Distributable version of TTDDD.zoo on
ab20.larc.nasa.gov, and you will get a lot of information about the
format and what the possibilities are.

                            -- Glenn Lewis

[e-mail] | These are my own opinions... not Intel's

##

Subject: Re: none
Date: Tue, 30 Apr 1991 11:21:53 GMT
From: [e-mail] (Stephen Menzies)

[e-mail] writes:


>I agree that Imagine's animation facilities are primitive compared to its other
>powerful modules.  I have seen and used (to an extent) 3D Studio for MS-DOG
>computers.  It's keyframer is incredible.  It should be though, it was written
>by Dan Silva, who of course wrote the Dpaint trilogy (Yes, we did lose him to
>IBM programming... thanks a lot Electronic Arts!).

[stuff deleted]
> Can you imagine trying to
>make a convincing bouncing ball in Imagine (one that slows to a stop at the
>apex of it's bounce)... it would be >rough< !

    A very good example. I have tried this several times and I still
can't believe how difficult it is to do such a simple thing.

[stuff deleted]
>I agree that We should scrap the Action editor and cycle editor, and make a
>keyframe editor combining the best of both: Keyframes can be made on F/X,
>alignment, morphing, etc... WITH easing!!  Object cycles are a neat idea, but
>it is only a step in the right direction.

I looks like Impulse completely forgot about keyframing and plugged
it in very quickly and very incompletely in 1.1  What really got me here
was that there was no way to call up your prev/next keys from the
stage editor. For anyone who has done alot of interactive keyframing
and then has to finetune or remove/change a key , I hope they have
a photographic memory.

>The way 3D Studio ( which costs nearly as much as two Toasters, by the way )
>does it is to have a dialog box that represents each frame (like Action Ed).  Each
>object has the following channels (also like Action Ed) move, morph, align, s
>size/scale (maybe more).  If a change is made at a frame, then a dot apears in
>the dialog in the appropriate channel.  If you double click on a dot, you bring
>up a requestor that controls the dot's function: eases, bias, spline magnitute,
>etc. (This is what IMAGINE is missing!)  Every dot has this ability.  A cycle
>could be copied over frames by copying a range of keyframes one after another
>( a tool is provided to do this easily).  Dots do there thing from the status
>of the previous dot ( in Imagine, dot to dot would be a bar ).

    Are you presented with a requester that requires numerical entry
or are you presented with a graph (function curve)? If it's just numerical
data that's entered, I think this can be a problem compounded by the fact
that the data is hidden when the requester is quitted. For example, just
looking at Imagine's bars actually tell you nothing. Trying to coordinate
between one bar and another bar is just too difficult. Trying to coordinate
between tens of these bars is next too impossible. I can see that ActionEd
is "animation" program  it's just that I don't think bars or dots (hidden
information) is the way.
    I would rather see this information displayed in graph form with
interactive editing of the graphs function curve. It should be possible
to call up the curve for the bouncing balls translation and overlap it
with the balls rotation and then overlap it with the balls scaling etc.
You should be able to change the curve by adding contol points (keys)
at will and turning them to change the amplitude. No numbers!
    You should also be able to zoom in/out on the graph so that
you can see all the frames or just frame x to frame y.
    It's also important that the neccessary information that you
need to acomplish the immediate needs is
*not* several screens deep. It must all be on the same screen.
    If this causes problems on limited machines then maybe Impulse
should make a lesser upgradable version and a professional version
with Xmemory and accelerator.

>Imagine 2.0???  :^)

Hope so...are they listening though?
--
Stephen Menzies
Email: [e-mail]


##

Subject: Re: objects flyin' everywhere!
Date: Mon, 29 Apr 1991 20:36:38 GMT
From: [e-mail] (Stephen Menzies)

Mark Thompson <[e-mail]> writes:

>> So what do we need?? I have a few suggestions.
>> 1> keyframing at the minimum *must* create an internal spline.
>>     (I believe Lightwave does this for translations atleast)

>Lightwave allows any keyframe to be either linear or spline interpolated
>and the spline interpolation can be done for translation, rotation, stretching,
>and scaling of objects and camera/light motion. Also, all other keyframe-able
>attributes such as polygon size, object morph, object dissolve, etc. can
>be spline interpolated.
>|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
>|      `       '        Mark Thompson                 CONCURRENT COMPUTER  |
>| --==* RADIANT *==--   [e-mail]        Principal Graphics   |
>|      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   |
>|     Productions       [phone] [phone]   & General Nuisance   |
>|                                                                          |
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Even better than I had thought. Clearly, Impulse has fallen well behind
in the animation department. It's hard to consider any software  that
can't spline interpolate motion and transformations ,today, an animation
software. This is something Impulse should have solved before they went
after F/X's and cycle editors etc.
    Mark, can you tell me which spline interpolated transformations
provide you with an interactive spline for finetuneing in Lightwave?
I recall  this bein the case for lights and morphing. I also remember the
display box (envelope?) be too small and somewhat primative but atleast
they were there. Can you zoom in and out on the graph? Can you interpolate
between linear and spline? Any info would be appreciated...

stephen menzies
--
Stephen Menzies
Email: [e-mail]


##

Subject: IFF24 & raw formats
Date: Tue, 30 Apr 91 17:53:37 EDT
From: [e-mail]

Something I discovered after writing my own IFF24 read/write routines:
The Sculpt 4D picture format is a RAW format. No header, no compression.
I wanted to have a raw format so I could output to a Kodak dye-sublimation
printer, and ended up writing the conversion in C! The Art Department
will cheerfully read and write the Sculpt files.

Just a note for those people trying to convert and port color pix.

-Steve

PS- The Kodak printer is state-of-the art! Over 2K by 2K resolution at
300 dpi, 32 bit color (extra 8 bits for black overlay). Output is
printed on photographic paper. No halftoning- true shading. $10 per
print, though. Ouch!


---------------------------------------------------------------------------
Steve Worley                                        [e-mail]
---------------------------------------------------------------------------


##

Subject: Re: objects flyin' everywhere!
Date: Tue, 30 Apr 91 18:30:14 EDT
From: Mark Thompson <[e-mail]>

Stephen Menzies writes:
>     Mark, can you tell me which spline interpolated transformations
> provide you with an interactive spline for finetuneing in Lightwave?
> I recall  this bein the case for lights and morphing. I also remember the
> display box (envelope?) be too small and somewhat primative but atleast
> they were there. Can you zoom in and out on the graph? Can you interpolate
> between linear and spline? Any info would be appreciated...

Here ya go Stephen...

Envelopes as they are refered to in Lightwave currently exist for:
object metamorph
object dissolve (fade a specific object in or out)
object polygon size (used for explosion effects)
light ambient intensity
light intensity (for each light other than ambient)
camera zoom factor
camera saturation (controls color saturation)
foreground dissolve (allowing a fade from one scene to the next)

The intent for future releases is to greatly expand this list. Along with
a graph of the value over time is a set of controls for using the envelope.
The controls available for each envelope include:
Create Key
Delete key
End Behavior (what to do when final envelope frame is reached)
    Reset (return to defaults when evelope is finished)
    Stop (use the final envelope value for the rest of thge animation)
    Repeat (cycle the envelope)
Current Frame
    Number
    Envelope value
    Linear/Spline (interpolation method from previous keyframe to this one)

I don't believe you can zoom in on the graph, although since most of my
animations consist of dozens of simpler scenes, I haven't had a need to.
They are somewhat primative. I would like to be able to grab the curve
with the mouse and freely manipulate it in various ways but currently
that can only be done by manually setting values at each keyframe. Each
keyframe to keyframe span can either be spline or linear interpolated
(but you cannot interpolate from linear to spline in the same key span).

The other keyframeable attributes (such as object translation, rotation, etc)
have controls nearly identical to the envelopes but without the visual graph
(perhaps because they already can be visualized in the scene layout window).
Additional controls added include position, direction, scale, parent, and
target (the last two are for scene hierarchy and camera/light targetting
respectively).  I might note that unlike Imagine, things like keyframed
rotation are very simple. If you want something to spin 10 times over a
period of five seconds, set its heading to 0 degrees in frame 1 and 3600
degrees in frame 150 (you are done). Also, any created motion can be saved
and used with other objects, lights, etc.

Hope this helps.
%~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~%
%      `       '        Mark Thompson                 CONCURRENT COMPUTER  %
% --==* RADIANT *==--   [e-mail]        Principal Graphics   %
%      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   %
%     Productions       [phone] [phone]   & General Nuisance   %
%                                                                          %
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

##

Subject: Re: objects flyin' everywhere!
Date: Wed, 1 May 1991 07:54:55 GMT
From: [e-mail] (Stephen Menzies)

Mark Thompson <[e-mail]> writes:

>Stephen Menzies writes:
>> [stuff deleted]
>> between linear and spline? Any info would be appreciated...

>Here ya go Stephen...
>[lots of info deleted]

>The other keyframeable attributes (such as object translation, rotation, etc)
>have controls nearly identical to the envelopes but without the visual graph
>(perhaps because they already can be visualized in the scene layout window).
>Additional controls added include position, direction, scale, parent, and
>target (the last two are for scene hierarchy and camera/light targetting
>respectively).  I might note that unlike Imagine, things like keyframed
>rotation are very simple. If you want something to spin 10 times over a
>period of five seconds, set its heading to 0 degrees in frame 1 and 3600
>degrees in frame 150 (you are done). Also, any created motion can be saved
 >and used with other objects, lights, etc.

Thanks for all the info, Mark. With this list I'll know what I'm looking
for (whats there and whats not) the next time I get up to a dealers.
Sounds like they (Hastings) knows what he's doing when it comes to the
animation modules. I would have to see it though and "feel" how it all
works together but it obviously leaves Imagine in the dust.

>Hope this helps.

sure will !

>%~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~%
>%      `       '        Mark Thompson                 CONCURRENT COMPUTER  %
>% --==* RADIANT *==--   [e-mail]        Principal Graphics   %
>%      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   %
>%     Productions       [phone] [phone]   & General Nuisance   %
>%                                                                          %
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Stephen Menzies
--
Stephen Menzies
Email: [e-mail]

##

Subject: New Detail commands
Date: Wed, 01 May 91 12:39:09 EDT
From: [e-mail]

Ed Chadez wanted to know what some of the new 1.1 Detail Editor
commands did.

Fracture will take any selected faces and split each into four new,
"fractured" faces by adding three new points, one on the bisector of
each line segment. Faces that share an edge with a fractured face will
split in two, so they keep the same edges in common. Fracture is
useful for increasing "face resolution" for better Phong shading,
coloring, and attribute assignment.

Split will take all of the selected points and "split" them apart from
an object to form a new object. Thus, you could sever the wing of a
plane to form a wing and a fusalage.  Previously, you'd have to use
slice, or delete the appropriate points on two copies of the object.
Any edges and faces connecting the selected, "split", points and the
non-selected points will be deleted. A new axis is created with the
same size, position, and orientation as the orignal axis.

Taut takes and makes a set of connected line segments line up in a
straight line. For example, if you "add lines", you'll get a string of
connected points in a row. If you "select points", select all of them
in order, then use "taut", they will string themselves out in a
straight line between the first selected point and the last. This
feature migght be useful in creating pointed paths, or making outlines
for skin or extrude.

-Steve

---------------------------------------------------------------------------
Steve Worley                                        [e-mail]
---------------------------------------------------------------------------




##

Subject: Re: IFF24 & raw formats
Date: Thu, 02 May 91 11:44:50 EDT
From: Mark Thompson <[e-mail]>

Steve Worley writes:
> PS- The Kodak printer is state-of-the art! Over 2K by 2K resolution at
> 300 dpi, 32 bit color (extra 8 bits for black overlay). Output is
> printed on photographic paper. No halftoning- true shading. $10 per
> print, though. Ouch!

I needed some slides done from 24bit IFF files and found only one place that
did pixel interpolation going from 768x482 up to full slide resolution
of a few thousand lines. The company is Computer Creations II and they
charged a whopping $50 a slide.

#######                           ###
#     #  #    #   ####   #    #   ###
#     #  #    #  #    #  #    #   ###
#     #  #    #  #       ######    #
#     #  #    #  #       #    #
#     #  #    #  #    #  #    #   ###
#######   ####    ####   #    #   ###

|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
|      `       '        Mark Thompson                 CONCURRENT COMPUTER  |
| --==* RADIANT *==--   [e-mail]        Principal Graphics   |
|      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   |
|     Productions       [phone] [phone]   & General Nuisance   |
|                                                                          |
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

##

Subject: Imagine/MMR Benchmark!
Date: Fri, 3 May 91 13:11:29 PDT
From: [e-mail] (John T. Grieggs)

        Benchmarking the elusive Mega-Midget Racer!

I recently acquired a Mega-Midget Racer from Computer Systems Associates.  My
goal was to speed up Imagine - nice program, but glacially slow.  I initially
bought just the CPU board, with the 33 MHz CPU.  This sped things up, but
not all that much.  How much?  Well, to answer that question, I did some
benchmarks.  The first set of timings below compare the 33 MHz Mega-Midget
Racer, using various cache settings, to a stock Amiga 500.

Scanline rendering a 320x200 ILBM-12 image of a wooden sphere with one light
source and some non-zero Zenith and Horizon settings:

    CPU          68000    68030    68030    68030    68030
    Cache Settings         I/N D/N  I/Y D/N  I/N D/Y  I/Y D/Y

    Timings         15:29    14:31     8:21    14:03     7:51
      in Seconds       929      871      501      843      471

    % improvement       0.0%     6.2%    46.1%     9.3%    49.3%
    68000 equivalence    1      1.07     1.85     1.10     1.97

Not bad, but not what I was hoping for.  The maximum improvement in my
application seems to be just under 50% - not quite twice as fast.  Not
bad, but a far cry from what I was after.  So, based on information
gleaned from the guys at CSA (whom I can't say enough nice things about!),
I ordered a 33 MHz 68882 and 512K of SRAM.

The chips arrived, and I installed them myself.  No problem - I'm not
afraid of hardware, and the documentation is perfectly adequate.  Be sure
to read it all the way through, tho - the diagram showing the proper
orientation for the 68882 is way in the back!

Based on my prior benchmarks, I did all further tests with both caches
enabled.  My system seems a bit more stable as a rule with the data cache
turned off, tho, so I usually run with it that way now.  But for the
benchmarks, both on.  The SRAM was installed with 1 wait state (the fastest
it would run at).  So here's the same wooden sphere with the new hardware:

    CPU          68000    68030    68030    68030
    Cache Settings         I/Y D/Y  I/Y D/Y  I/Y D/Y
    Imagine version           Int.     F.P.     F.P.
    SROM                Y         Y        N
    SRAM                Y         Y          Y

    Timings         15:29     5:31     3:24     2:03
      in Seconds       929      331      204      123

    % improvement       0.0%    64.4%    78.0%    86.8%
    68000 equivalence    1      2.81     4.55     7.55

Note that the SRAM-only number is significantly faster than the others - this
is because most of the rendering is done in the 512K of SRAM.  What I did for
this test was remove all of my neat stuff (including the OS) from SRAM memory,
forcing all 512K to be free.  However, the code doing the rendering is still
executing in "slow" RAM - 4 Mb on a Trumpcard.  It seemed obvious that more
32-bit RAM would speed things up further by allowing the Imagine code and the
OS to execute more quickly as well. This was sufficient reason for me to
procrastinate on writing up this benchmark for a couple of months while
waiting for a check from the government. :)  When it came, I bought myself a
fully populated, 8 Mb board from CSA.  This put me up to just about the
fastest Amiga configuration I have heard of: 33 MHz 68030, 33 MHz 68882,
512K SRAM, 8 Meg 32-bit DRAM.  Ahhhhh...

Once it was in, I ran the wooden sphere test once more, with these results:

    CPU          68000    68030
    Cache Settings         I/Y D/Y
    Imagine version           F.P.
    SROM                Y
    SRAM                Y
    32-bit DRAM            Y

    Timings         15:29     1:33
      in Seconds       929       93

    % improvement       0.0%     90.0%
    68000 equivalence    1      9.99

Now, that's more like it!  Since the actual time involved in the test was
getting so small towards the end there, I did a couple more tests using
a second, more elaborate test scenario.  This scenario is a trace of three
spheres.  Sphere one is color-mapped with a picture of Einstein (which I
believe came from the net).  Sphere two is reflect-mapped with the same
picture.  Sphere three is color-mapped AND reflect-mapped with the same
picture.  This image also used a star-field.  As might be expected, this
gave Imagine a bit more of a workout!

Ray-tracing a 320x400 RGBN-12 image of 3 spheres with brush maps, with one
light source and a star-field:

    CPU          68000    68030    68030    68030
    Cache Settings         I/Y D/Y  I/Y D/Y  I/Y D/Y
    Imagine version           F.P.     F.P.     F.P.
    SROM                Y        N        Y
    SRAM                Y         Y        Y
    32-bit DRAM            N        N        Y

    Timings           1:39:57    19:26    11:54     9:39
      in Seconds      5997     1166      714      579

    % improvement       0.0%    80.6%    88.1%    90.3%
    68000 equivalence    1      5.14     8.40    10.36

About the tests:

Timing was done using a stopwatch.  I would have much preferred to do it
with AREXX or something similar, but Imagine doesn't let you do that.  Each
test was a single frame animation.  All times are from the instant I hit
the "No" button on the Subproject screen (this is the final prompt before
it takes off) to the time the "Cancel" requestor disappears.  If I were to
restart this from scratch, I would probably do it a little differently, but
I thought keeping it consistent was more important than re-starting it every
time I learned something new about Imagine!  Empirically, the time used to
do the animation phase looked like a tiny fraction of the time taken to
render.  I believe it can be treated as a fixed overhead (i.e., ignored).

% improvement was derived by subtracting the time in seconds for the entire
process from the time in seconds for the same process on a stock 68000, then
dividing by the latter number.

68000 equivalence was derived by dividing the time in seconds for the entire
process on a stock 68000 by the time in seconds on the configuration under
scrutiny - I like to think of this as the number of 68000s you would need to
stack on top of each other to do the same job in the same time.  :)

I included the raw numbers as well so that the curious and bored could come
up with their own statistics.

Further benchmarking:

I am pretty well satisfied with my results, but I am sure there are others
out there with souped-up systems who would like to compare my numbers to
theirs.  In particular, I would be interested in hearing how my MMR stacks
up against the new 68040 accelerators starting to appear.  Also, there is
one more small increment I could do to upgrade it - I could add a 50 MHz
68882, but would first like to know how much it would help.

I may live to regret this, but here - I'll volunteer.  I will be willing to
make my scenarios and support available to anyone with a turbo system that
wants to try and get some comparable numbers.  You need to have your own
copy of Imagine and a fast system - send me mail.  I would like to do this
in a semi-controlled manner, so that I can keep the results together for
easy reference by the Net community.

Summary:

The MMR makes my system a real pleasure to use!  I highly recommend it for
anyone who can afford it - excellent yuppie toy!  With the OS in 32-bit RAM,
just about everything enjoys a noticeable speed-up.  Window re-draws, object
manipulation, and rendering all blaze. Once you have seen a re-draw of a
complex object in the detail editor on my system, you will never want to go
back to anything less!

Bottom Line:

Imagine renders around 10 times faster on a 33 MHz 38030 with 33 MHz 38882,
512k of SRAM, and 8 Mb of 32-bit DRAM than on a stock Amiga 500.

John T. Grieggs (Telos @ Jet Propulsion Laboratory)
4800 Oak Grove Drive, Pasadena, Ca. 91109 M/S 301-320T    [phone]
Uucp: {cit-vax,elroy,chas2}!jpl-devvax!grieggs
Arpa: ...jpl-devvax![e-mail]

##

Subject: Re: Imagine/MMR Benchmark!
Date: Fri, 3 May 91 21:22:55 -0500
From: Donald Richard Tillery Jr <[e-mail]>

Congratulations on your Mega Midget Racer purchase(s).  I'd like to
compare my system to yours if you don't mind.  I have a GVP A3001 28Mhz
68030/882 w/4Megs of 32 bit Dynamic (Nibble Mode Simms) RAM (0 wait
states).  I'll give you the results on your Imagine files as soon as you
get them to me and I get them rendered (finals are coming up you understand).

I am supposed to be on the list to beta test the GVP 68040 card when it
reaches that stage, but that looks like about the end of the summer.  They
say the 3000 version should be out in a couple of months (they aren't happy
with the 17 MIPS they are getting right now and are working on improving it),
and the 2000 version will follow on its heels (they say :-).  Assuming their
non-disclosure agreement permits it, I'll get the results of tests with that
board to you then (but you'll have to give my your address or phone number
because about that time I will lose my access to UseNet and be moving to
San Diego).

Please forward the files to me and I'll get back to you with the results.

Thanks.

Rick Tillery ([e-mail])


##

Subject: Re:  Imagine/MMR Benchmark!
Date: Fri, 3 May 91 19:52:29 PDT
From: [e-mail] (Dave Coughran)

I noticed the same sort of speedup when I expanded the fast memory of my
3000 to 4megs.  I had rendered a chrome sphere floating over a shiney
floor that took about 2 hours with the stock 2 megs of memory.  After
installing the new memory the rendering time dropped to something like
15 minutes.  I assume this is because Imagine was running in chip ram
in the 3000 because the fast ram was full of Kickstart, Workbench, and
Imagine.

The extra memory space is nice, but the improved performance make
it money well spent for anybody with a 2 meg 3000.


David Coughran            [e-mail]
                ...!uunet!lll-winken!koko!drc


##

Subject: ImConEdit
Date: Sun, 05 May 91 16:00:57 EDT
From: [e-mail]

Thank you, Stephen, for uploading ImConEdit! And especially thanks to the
author, Sheldon Arnst!

I had the same annoyance- continually editing the config file to set
the antialias level (and # of reflections, sometimes).

For those who haven't seen it, ImConEdit is a very simple program that
will set your config parameters without having to use a text editor. Its
nothing complex, but it doesn't have to be. Its a lot easier to push
a slider than run emacs and change numbers.

Stephen- one improvement for Sheldon- when you are editing a parameter
like "EDLE" the program should put a one sentence description in the
slider gadget, saying something like "Edge Level- 0=High Antialiasing"
This would help people identify some of the more abstract parameters since
we lose the comment explanation thats in the config file. Perhaps the
description could even BE the comment for that line.

Again, cudos for Sheldon Arnst.

-Steve

---------------------------------------------------------------------------
Steve Worley                                        [e-mail]
---------------------------------------------------------------------------



##

Subject: Resizing and rotating PART of an obj.???
Date: Mon, 6 May 91 1:41:51 MET DST
From: [e-mail]

Hi!

So I finally got my Imagine package.. :)

My question:  How do I resize or rotate a small part of an object?
I know I can move them with "drag points" but I cant resize them or
rotate them. (Small part = A few points or a few faces)



  --   Henrik


Internet : [e-mail]

Maybe I should create on of those flashy sig's you all have... :)



##

Subject: Deleteing points and faces? and need the upgrade to 1.1
Date: Sun, 05 May 91 19:01:56 -0700
From: [e-mail]

I seem to have a problem with my 1.0 version right now (I need to
get the upgrade; Do I have to call up Impulse or what?)

I want to delete points off an object, as well as its connecting faces
obviously, but there doesn't seem to be a function to do that.  When I try,
the program wants to delete the entire object itself!!
Can anyone give me some pointers on this?

--Dave"Radboy Go!"

##

Subject: Re: ImConEdit
Date: Mon, 6 May 1991 08:09:18 GMT
From: [e-mail] (Stephen Menzies)

[e-mail] writes:

>Thank you, Stephen, for uploading ImConEdit! And especially thanks to the
>author, Sheldon Arnst!

Just a small correction so we don't confuse anyone. The actual name of
this utility is "ImConEd" and the version is 1.1

>[stuff deleted]
>Stephen- one improvement for Sheldon- when you are editing a parameter
>like "EDLE" the program should put a one sentence description in the
>slider gadget, saying something like "Edge Level- 0=High Antialiasing"
>This would help people identify some of the more abstract parameters since
>we lose the comment explanation thats in the config file. Perhaps the
>description could even BE the comment for that line.

I quite agree with you steve, and I know Sheldon does too (he's mentioned
that to me before). I suspect see this addessed in a future version. Btw,
I've seen an incomplete version of the Function Key Editor and it should
make the (interactive) editing of the FKeys a snap.

>Again, cudos for Sheldon Arnst.

I'll pass your comments along...

>-Steve

>---------------------------------------------------------------------------
>Steve Worley                                        [e-mail]
>---------------------------------------------------------------------------

--stephen
--
Stephen Menzies
Email: [e-mail]

##

Subject: Ground
Date: Mon, 6 May 91 15:33:34 CDT
From: [e-mail] (Randy Carris)

Steve,
I posted a message on the Imagine mail list, but I don't know if it was recieved or simply no
cared to answer.  Could you do me a favor and upload a screen-shot of  the proper brush
position to get a repeating brush-map to appear on the ground.  The only way I've been
able to succeed is to map a brush on a plane.  But a plane gets cut off at the world "edge"
and doesn't continue to the horizon like I want.  I have the world size set to
8000x8000x8000.  Any tips or a screen shot sent to ab20 would be greatly welcomed.

Thanks,
Randy Carris

##

Subject: ground brush maps
Date: Tue, 7 May 91 10:59:00 -0700
From: [e-mail] (Ed Chadez)

I know I've asked this before, so please excuse the bandwidth.

How do you manipulate the brush axis (if necessary) to wrap a brush to the
ground object.  Has anyone used a 'ground' as something other than an
infinate "floor" (ie, sky?, wall?), and how would the brush axis be
manipulated for those objects.

I'm sure this subject has come up before, but I can't seem to find the
reply in my mail-box.  Oh well. :-)

Thanks in advance!

-Ed Chadez

btw--anyone want to buy an amiga500??

--
--//---------------------------------------------------------------------------
\X/ [e-mail]/Edward Chadez                    CARL Systems(303)861-5319
-------------------------------------------------------------------------------

##

Subject: Ground brush mapping
Date: Tue, 07 May 91 21:16:26 EDT
From: [e-mail]

Ask and ye shall recieve. There is now a file on hubcap.clemson.edu in
/pub/amiga/incoming/IMAGINE/MISC called flatmappix.lzh which has
screen grabs for both ground and flat plane brush map axis placement.
There is also a long brush map tutorial on the same site in the
IMAGINE/TEXT directory, in the 4th archive file of this list.

-Steve

---------------------------------------------------------------------------
Steve Worley                                        [e-mail]
---------------------------------------------------------------------------


##

Subject: Lights - What Does Y Axis Do?
Date: Tue, 07 May 91 22:50:01 PDT
From: [e-mail]

I have a question that isn't covered in the manual (amazing isn't it).
It says that for Cylindrical and Conical light sources the size parameter
for X and Y axis affect the light (but not the Z). Obviously the X axis
is the radius of the light beam, but how would length (Y) affect the
light? Also, does the light (cylindrical) spread out over distance, or
does it stay at it's set radius forever?

Just curious.

=======================================================================
Sean Schur                        USENET: [e-mail]
Assistant Director Amiga/Media Lab        Compuserve: 70731,1102
Character Animation Department            Plink: OSS259
California Institute of the Arts
=======================================================================

##

Subject: ground brush maps
Date: Wed, 08 May 91 09:25:19 EDT
From: [e-mail] (Bob Lindabury - SysAdm)

rutgers!carl.org!echadez (Ed Chadez) writes:

> I know I've asked this before, so please excuse the bandwidth.
>
> How do you manipulate the brush axis (if necessary) to wrap a brush to the
> ground object.  Has anyone used a 'ground' as something other than an
> infinate "floor" (ie, sky?, wall?), and how would the brush axis be
> manipulated for those objects.

Hmm..it seems to me that the ground would work ok for a sky but for a
wall, I don't think so.  I think you would be better off using just a
flat plane with brush/texture maps as walls rather than the ground
object.  I have yet to try it as a sky.  I have mapped the wood
texture on the ground however.  You manuipulate the brush axis just
as you would a normal object I imagine. <grin>

I guess part of the reason there are so many questions is because of
the docs (or lack thereof).  I think the name reflects the docs more
than the program. <grin>

While we are on the brush subject...

I can't for the life of me get the filter brush map to work.  The
docs don't really explain it and I can't seem to get any transparency
at all.   As I understand it, you should be able to take a
multi-colored brush, map it on an object and select filter and be
able to see through the various colors.  Am I mistaken?

What I did is taken a plane and made it into a rectangle.  I then
extruded it about by 20.  I then took a brush I made in Dpaint which
was created as follows:

Enter dpaint in 640x400 4 color mode.  Use the default colors, clear
the background to the blue color.  I then drew a filled rectangle in
black and then copied it in white.  I then set the white slider all
the way up to it's highest white settings (15,15,15) and captured the
two rectangles (white & black) as a brush.  Saved the brush.

I then mapped the brush onto my rectangle..a bit smaller than the
rectangle so that it has 2 smaller rectangles in the middle of it.  I
picked filter brush.  I made the Y axis of the brush go all the way
through the rectangle object.  I then placed an object behind the
rectangle so that I could see something behind the rectangle if the
filter worked.

Now, I don't know how the filter works..is white a full filter or is
black a full filter or doesn't it work that way.  Either way, I
didn't get any of the rectangle to be transparent with the filter
map.  I placed a light on the object behind the rectangle as well.

The rectangle that is mapped with the filter map is just a plain
default setting with a brown kinda color.  No other attributes are
set.  What's wrong with this picture.  It basically looks exactly the
same as a color map.  I get a brown rectangle with the outside of the
brush color (blue) with two smaller rectangles (white & black) on the
face of the rectangle.  Nothing is transparent or filtered from what
I can tell.

I did this in both scanline and trace mode with no difference in the
image.

Quite frustrating I assure you.  Have I missed something?  Am I
turning into a major putz?  Why is it simple things don't seem to
happen?  When can we expect some "real" documentation.  I am
currently using version 1.1.

-- Bob

 The Graphics BBS  908/469-0049  "It's better than a sharp stick in the eye!"
 ============================================================================
  InterNet: [e-mail]                | Raven Enterprises
      UUCP: ...rutgers!bobsbox!graphics!bobl      | 25 Raven Avenue
    BitNet: bobl%graphics.rent.com@pucc           | Piscataway, NJ 08854
    Home #: 908/560-7353                          | 908/271-8878

##

Subject: Filter Brushes
Date: Wed, 8 May 91 13:39:59 -0400
From: Udo K Schuermann <[e-mail]>

[e-mail]:
> I can't for the life of me get the filter brush map to work.  The
> docs don't really explain it and I can't seem to get any transparency
> at all.   As I understand it, you should be able to take a
> multi-colored brush, map it on an object and select filter and be
> able to see through the various colors.  Am I mistaken?

Bob, you've got the correct idea, regarding filter brushes:

1. The color of any pixel indicates how much light it lets pass
   through (just like Imagine's general filter slider).  Black will
   pass nothing, blue passes only blue light, white passes all, etc.
   The RGB guns of the color determine how much of a particular
   "wavelength" is let through.

2. I've found a relatively simple formula for getting all brushmaps to
   work flawlessly.  I can consistently map an iff onto an object (at
   least the flat ones):

    brushes "live" only in the POSITIVE ( >0 ) X,Y,Z of the brush
    axes.  Always, ALWAYS, move/scale the axes so that the object
    lies in the positive area ... not 0, but ABOVE 0.  Get it?
        This is different from Textures, which live in both
    the positive and negative coordinates.

Bob, your problem may be that the Z-axis is flat on the object and the
result is all black color or perhaps some random noise pattern, which
may be mostly black on black.

Hope this helps!

 ._.  Udo Schuermann       "Did you ever wonder why we had to run for shelter
 ( )  [e-mail]    with the promise of the brave new world unfurled
                     beneath the clear blue sky?" -- Pink Floyd

##

Subject: Transparency Brush Maps
Date: Wed, 08 May 91 15:19:46 EDT
From: [e-mail]

Bob, your description of the setup of a filter brush SHOULD work. I
don't see anthing fundamentally wrong with it.

Some answers: Filter brushes let light through where there is color on the
map- black is opaque, light blue lets a little blue through, white
is  transparent.

I do not think you are turning into a major putz, but there are a couple
things you can try. Are you sure the brush AXIS covers the entire rectangle,
or do you have the brush BOUNDING BOX covering the rectangle? If its the
latter, only half the thickness of the rectangle would be transparent, and
it would look a lot like a color map, similar to what you're getting.

Also, try using a flat plane and see what happens. I've done a couple of
filter maps on solids and they worked out, but I've done many, many more
on planes, and they are pretty reliable.

An interesting problem that shouldn't cause a filter map to fail, but
is useful anyway- if you use a Dpaint 0-15 pallette range, your filter
maps will work, but you cannot get pure whites [Completely transparent]!
This is because 15*16 != 255. If you want a perfect transparency, you
have to use The Art Department (or something similar) to change the brush
to a 24 bit brush. Then it works beautifully. Otherwise the pure whites
are only about 94.1% transparent. Trivia I should have put in my brush
mapping article.

Bob, try using a flat plane, tell me what happens via E-mail. We'll get
it to work. Somehow. :)

-Steve



---------------------------------------------------------------------------
Steve Worley                                        [e-mail]
---------------------------------------------------------------------------


##

Subject: Transparency Brush Maps
Date: Thu, 09 May 91 09:09:20 EDT
From: [e-mail] (Bob Lindabury - SysAdm)

rutgers!athena.mit.edu!spworley writes:

> Bob, your description of the setup of a filter brush SHOULD work. I
> don't see anthing fundamentally wrong with it.

Neither did I.  That's why I was so puzzled.

> Some answers: Filter brushes let light through where there is color on the
> map- black is opaque, light blue lets a little blue through, white
> is  transparent.

Well, I received a note from Rick Rodriguez who stated the the
manuals were written before the filter maps were a feature of Imagine
and he said that the docs were all wrong.  He also stated that the
filter, altitude and reflection maps all work on grey scales with the
highest level giving you the most transparency/reflection and visa
versa.  So white = full transparency or reflection.

> I do not think you are turning into a major putz, but there are a couple
> things you can try. Are you sure the brush AXIS covers the entire rectangle,
> or do you have the brush BOUNDING BOX covering the rectangle? If its the
> latter, only half the thickness of the rectangle would be transparent, and
> it would look a lot like a color map, similar to what you're getting.

Yes, I will start with a flat plane and then move to solids.
However, I did make my Y axis pass *all* the way through the extruded
rectangle.  It only seemed logical.  I may have left the brushs Y
position at 0 as is the objects front plane.  From what I understand,
this can cause problems and that you should move your brush position
to be just a little in front of your object.  Is this a fact?

> An interesting problem that shouldn't cause a filter map to fail, but
> is useful anyway- if you use a Dpaint 0-15 pallette range, your filter
> maps will work, but you cannot get pure whites [Completely transparent]!
> This is because 15*16 != 255. If you want a perfect transparency, you
> have to use The Art Department (or something similar) to change the brush
> to a 24 bit brush. Then it works beautifully. Otherwise the pure whites
> are only about 94.1% transparent. Trivia I should have put in my brush
> mapping article.

Thanks for the additional info.

> Bob, try using a flat plane, tell me what happens via E-mail. We'll get
> it to work. Somehow. :)

Will do!

> -Steve

-- Bob

BTW, just got my 2630 yesterday.  WHAT A DIFFERENCE!!!  I now feel I
can experiment with Imagine more because a scene that took 5 minutes
and 30 seconds before now takes me 52 seconds to render!!  And
wireframe previews are at least bearable to both generate and view in
the actual speed they are suppose to run.  I'm in heaven...

 The Graphics BBS  908/469-0049  "It's better than a sharp stick in the eye!"
 ============================================================================
  InterNet: [e-mail]                | Raven Enterprises
      UUCP: ...rutgers!bobsbox!graphics!bobl      | 25 Raven Avenue
    BitNet: bobl%graphics.rent.com@pucc           | Piscataway, NJ 08854
    Home #: 908/560-7353                          | 908/271-8878

##

Subject: Strange problem...
Date: Thu, 9 May 91 13:37:33 EST
From: [e-mail] (Kevin Goroway)

I think I have seen this one before, but here goes...

I am positioning the axis/box of a brush map, interactively.
I get it to where I want it (I'm actually scaling...) and hit the space
bar.  Select edit again, and the box is back where it was, as if I had
hit esc previously...
This happens whether I edit in local or world mode.  Anyone ever see this
problem? Is there a work around?

It is quit annoying to have a brushmap slightly off, and not being able to
move it...

Thanks
+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=
| Worcester Polytechnic Institute   | "It happens sometimes, people just     |
| [e-mail]                  |   explode, natural causes."-Repo Man   |
+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=
+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=