Amiga Tools 3/Amiga Tools 3.iso/grafik/raytracing/imagine/tips/arc08.txt

Text File | 1993-03-12 | 169KB | 3,925 lines

IMAGINE archive: collected off of [e-mail]

ARCHIVE VIII
Jul. 14 '91 - Jul. 25 '91

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

note: each message seperated by a '##'

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

Subject: Tuts
Date: Sun, 14 Jul 91 15:26:51 EDT
From: [e-mail]

I either lost my Imagine manuals, or left them in Massachusetts. Bummer.

I've been busy at work, and only just got my 3000 up and running. From
all of your input, I've decided on the following Tutorial plan.

1) Detail Part II (Advanced fun features)
2) Stage Part I (Intro)
3) Stage Part II (Advanced)
4) The Art of Object Creation
5) Stage III (Planning and executing a non-lame animation)
6) Sneaky tricks (Tribute to Glenn Lewis)

And more later, like Cycle, auxilliary programs and hardware, like The
Art Department and DCTV, hi-end options like single frame recording,
Attributes, Imagine 2.0 features and mini-tutorial explanations (no, I don't
know what they are, but eventually... ),  and whatever else I think of.


Since I just got my computer up, expect the nect Detail Tutorial in
about 2 weeks.

-Steve

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

##

Subject: more trace problems?
Date: Mon, 15 Jul 91 14:35:44 PDT
From: [e-mail] (Aaron Tucker)

Hello all. Well, does 1.1 automatically scale the objects and world size up to
max (~32k I believe)? I added a size bar to the global and changed the world
size to 32000 on X Y and Z. When I scanline, there is no problem. But when I
go into trace mode...I only get a small window in the middle of the screen.
The image is not compressed, it is a window to my scene.

Thoughts? Fixes? Shrugs? Has this been hashed before?


Juan
[e-mail]

##

Subject: World sizes
Date: Mon, 15 Jul 91 20:04:03 EDT
From: [e-mail]

Juan asks about world sizes.

Here's my thoughts on them. World size is an awfully confusing option
in Imagine, which I think shouldn't even exist. Basically, what
happens is that when in TRACE mode, there is an algorithm that sorts
objects (and parts of objects) in a rough depth order so that figuring
out what obscures what is a bit faster. The world size tells Imagine
how big of a volume to expect objects in (what volume to sort).
Unfortunately, if you have objects outside this volume, they won't be
sorted and probably won't even be rendered. Thus, keeping your objects
INSIDE the world is important.

However, if you keep your objects contained within your world by
making EVERYTHING really tiny, your objects will take up only a small
volume of the world. When Imagine starts "sorting the world" it has to
wade through a lot of empty space to "get to" your objects.
Unfortunately, all of the extra empty volume takes time to sort
though, so the rendering slows down. Sometimes this slowdown is
DRAMATIC, like 10 times slower in trace.

[Forgive the handwaving explanations of an octree. I know you're
groaning, Mark.]

Thus, there is a certain scale you want to make all of your objects.
You do NOT want to make your objects larger than the world, or they
won't render. You also do NOT want to make your objects too small, or
you'll wait forever to render.

The way you can manually change world size is by adding a size bar to
the GLOBALS actor in the Action Editor. Unfortunately, a lot of people
use this to get around the sizing problem. Unfortunately, because a
lot of people don't know exactly what it does and MISusing global
world size is a very frustrating experience. You're welcome to play
with this "solution", but I'll give my FOOLPROOF method of sizing.

Steve's Foolproof World Size:

The DEFAULT world size is a cube that is 2048 units wide, high, and
deep, centered around 0 0 0. [X, Y, and Z coordinates range from -1024
to +1024] As long as all of your objects are within this range, they
will render, and you do NOT have to add a size bar to the Global
actor. This is the big trick- use the default! If you start mucking
with size bars, you're gonna confuse yourself. I do.

"But what if an object accidentally goes outside the volume? Don't I
have to increase the world size then?"

Well, no. You can always move and rescale your scene to fit inside the
world volume. Remember you might also have a scene that is too SMALL
so it takes forever to render, and this situation is almost at bad as
having cut-off objects.

What you do (and what _I_ always do) is design my scene at about the
right scale (ie, my scene's objects stay within a volume about 2048
units on a side.) I do NOT worry about going outside the volume or
trying to keep things big. I make sure that I don't do something
boneheaded like make cars 1 unit wide or 10000 units wide. The best
way to do this is to set a grid size of about 50. If I zoom in too
much, the grid dissapears, reminding me I might have the scale too
small. If I zoom way out, the grid blurs into a near-solid patters,
remining me I'm on a too-large scale. Remember, I'm not really paying
attention to the real size, I'm just making sure that I'm somewhere
within an order of magnitude of it.

Finally, when my scene is done, I use a trick to make everything fit
into the world nice and neatly. First, I set the grid size to 1023.
This makes the world size easy to see, its the volume between the first
grid lines. I select an object near the center of my scene, then I
multi-pick ALL of the objects (by holding down the shift key and
clicking on the objects one at a time.) This INCLUDES the camera and
the lights. Nothing should be left out: no objects drawn in white
should be on the scene.

Then (I) hit "m" for move. ALL of the objects are moved at the same
time, and their relative positions are not changed. Move the objects
so that they are centered about the 0 0 0 world origin. "Centered"
means the the far leftmost object is as far away from the origin as
the rightmost object, the frontmost is as far in front of the origin
as the backmost object is in back of the origin, etc.

Now, hit "s" to scale the objects. ALL of the objects will shrink or
expand around one point (the object you selected first). You want to
size the objects so that they JUST fit into the world volume that
you've defined by the grid lines. Usually just one dimension is the
limiting factor; size the objects to that the largest dimension fits.
(DON'T individually size the X,Y,Z dimensions!) You might want to use
"m" again if the shrinking/expanding scene moves off of center. Your
final configuration should have all of the objects centered inside
the world with the objects taking up most of the volume of the world
(in at least one direction).

The beauty of this method is twofold: first, you don't have to muck
with world size, but everything will render at optimum speed. Second,
you don't have to pay attention to world size while you create your
scene. The resizing will NOT, repeat will NOT, change your camera
views! The camera is resized with your objects, so everything stays in
proper perspective (literally). (The technical explanation is that
cameras measure angles, not distances.)

Thats it. I don't worrry about world sizes or slow renders now.
Frankly, I never understood how to tweek world size for perfect
renders, so this method saved me a lot of grief.

Hope this helps you Imagineers.

-Steve

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

PS sorry about the change of the active person from being "I" to "you".
I'm too lazy to fix it.

##

Subject: Changing world sizes, what is the problem?
Date: Tue, 16 Jul 91 10:56:00 +0200
From: [e-mail] (Helge Egelund Rasmussen)

Steve Worley <[e-mail]> write:
> The way you can manually change world size is by adding a size bar to
> the GLOBALS actor in the Action Editor. Unfortunately, a lot of people
> use this to get around the sizing problem. Unfortunately, because a
> lot of people don't know exactly what it does and MISusing global
> world size is a very frustrating experience. You're welcome to play
> with this "solution", but I'll give my FOOLPROOF method of sizing.

I've been (mis- :-) using the globals actor size without any problems until
now. I thought that this size defined the volume which was subdivided by the
octree algorithm. The problem with small objects would then be that the
algorithm only subdivided a given number of times, so that small objects would
fit entirely into a single 'octree box'. Because of this the octree algorithm
wouldn't give any performance enhancement for small objects.
On the other hand all objects must be in some 'octree box' to be rendered,so
an object bigger than the world size wouldn't render.
My guess then, is that the world size is used to optimize the 'compute ray
hit' algorithm, but doesn't affect the final picture (except for clipping).

Because of this I normally just create a scene (ignoring absolute sizes), and
then set the world size so that all that should be visible will be inside the
'world'. This make it easier to add a new object without having to scale it etc.
afterwards.

What are the problems with this method??

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: Re: World sizes
Date: Tue, 16 Jul 91 09:22:11 EDT
From: Mark Thompson <[e-mail]>

Steve Worley writes:
> However, if you keep your objects contained within your world by
> making EVERYTHING really tiny, your objects will take up only a small
> volume of the world. When Imagine starts "sorting the world" it has to
> wade through a lot of empty space to "get to" your objects.
> Unfortunately, all of the extra empty volume takes time to sort
> though, so the rendering slows down. Sometimes this slowdown is
> DRAMATIC, like 10 times slower in trace.

> [Forgive the handwaving explanations of an octree. I know you're
> groaning, Mark.]

Hahaha, I like the hand waving Steve. The thing is, I have a tough time
believing that the problem is algorithm related rather than programmer
implementation related. The intent of adaptive spatial subdivision (for which
octrees are a data structure for) is to break the world up into volumes
of interest that vary in size and density depending on the objects in the
environment. In other words, confine calculations only to the areas of
interest. The fact that making stuff too small massively increases trace
time runs counter to the adaptive spatial subdivision philosophy. There is
no reason that finding and sorting through empty volume should cause a 10x
slowdown considering that the subdivision is a very small calculation
compared to the actual tracing. I cannot claim to be a ray-tracing expert,
but it does not jive with me.
|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
|      `       '        Mark Thompson                 CONCURRENT COMPUTER  |
| --==* RADIANT *==--   [e-mail]        Principal Graphics   |
|      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   |
|     Productions       [phone] [phone]   & General Nuisance   |
|                                                                          |
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

##

Subject: Re: World sizes
Date: Tue, 16 Jul 91 10:46:24 -0400
From: Udo K Schuermann <[e-mail]>

Steve Worley writes:
> The resizing will NOT, repeat will NOT, change your camera views!

But it will NOT change scale-related values in textures either.  THAT
is a problem!  Try scaling an object with a nice wood-texture:  The
rings will become disproportional to the object :-(

For this reason I tend to do what Helge Rasmussen does:  scale the
world to fit my objects/scene instead the other way around.

A good solution would be to make size values relative to the object
they're applied to, rather than absolute (i.e. a percentage value).
That, of course, would have to come from Impulse.

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

##

Subject: Ship.lzh
Date: Tue, 16 Jul 91 11:53:23 EDT
From: [e-mail] (Rob Borsari)

I finally got my ftp straight and ship.lzh is posted to hubcap.  There is
a readme with it.  If anyone has any objects that would go well with mine (
boats aircraft missles helicopters) and is interested in a project let me know.
I built it and now I want to do somthing with it.  I am also interested in
converting it to lightwave.  The object file is aprox. 400k.  Has anyone
had louck with converting large objects?  Comments and Questions welcome.

-R-
[e-mail]   Rob Borsari  "You built it, You find somthing to do with
it"

##

Subject: Re: World size
Date: Tue, 16 Jul 91 18:33:43 +0200
From: [e-mail] (Helge Egelund Rasmussen)

Mark Thompson <[e-mail]> writes:

> The fact that making stuff too small massively increases trace
> time runs counter to the adaptive spatial subdivision philosophy. There is
> no reason that finding and sorting through empty volume should cause a 10x
> slowdown considering that the subdivision is a very small calculation
> compared to the actual tracing.

I agree. Normally you would just continue subdividing until the 'boxes'
in the areas occupying objects are small enough.
Impulse might however have a limit on the max number of subdivisions.
In this case there would be a limit on the size of the smallest box.
The reason for this might be that they don't do adaptive subdivison, but
                        --------
only subdivison (ie. subdivide everything, not only space occupied by
objects).

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 Sizes
Date: Tue, 16 Jul 91 12:57:46 EDT
From: [e-mail]

Helge Rasmussen writes "I've been (mis- :-) using the globals actor
size without any problems until now." He uses the world size bar without
any problems, so why use my wacky method?

First there isn't anything WRONG with adding a world size bar. The problem
is that what it does and how to use it is very confusing, especially
to the casual or new user. If you can use the world size bar and
know what you are doing, great! Note that you still should CENTER your
scene, though you don't need to scale it.

My method is slightly easier because you don't have to deal with the
world size at all. One disadvantage, as Helge points out, is that newly
added objects might have to be rescaled to fit in the scene.

As I said, I still don't understand world sizes. Sometimes when you set
the XYZ values to DIFFERENT numbers (like 1024, 512, and 34653) the
objects are still clipped, even though the volume defined by the rectangular
solid (not a cube anymore!) contains your objects. Also, I'm unclear wether
a size of 1024 means a span of -512 to 512 or a span of -1024 to 1024. Like
I said, I just ignore the problem, and it goes away. :-)

-Steve

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

##

Subject: World sizes again
Date: Tue, 16 Jul 91 13:20:01 EDT
From: [e-mail]

Mark Thompson writes:

Hahaha, I like the hand waving Steve. The thing is, I have a tough time
believing that the problem is algorithm related rather than programmer
implementation related. The intent of adaptive spatial subdivision (for which
octrees are a data structure for) is to break the world up into volumes
of interest that vary in size and density depending on the objects in the
environment. In other words, confine calculations only to the areas of
interest. The fact that making stuff too small massively increases trace
time runs counter to the adaptive spatial subdivision philosophy. There is
no reason that finding and sorting through empty volume should cause a 10x
slowdown considering that the subdivision is a very small calculation
compared to the actual tracing. I cannot claim to be a ray-tracing expert,
but it does not jive with me.

I respond:

I agree completely. The first step of the programmer should be to "normalize"
the octree space by first figuring out the gross scale of the scene (the
span of the world in each dimension) and then scaling and shifting the
base octree to match this volume. Instead, Imagine seems to use a fixed
octree that has an initial volume that is hardcoded (by default, and overridden
by the animator). If the inital octree volume were self-selecting, the
renders would always be optimum and nothing would ever get clipped.
Definately a poor implementation. Looks like Zack was cutting corners. :-)

-Steve

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

##

Subject: Morphing
Date: Tue, 16 Jul 91 18:22:22 EDT
From: [e-mail]

I was thinking about morphing in Imagine. Everybody is talking about
Terminator II morphing, and (Udo S., I think ??) was playing around
with it.

First, Imagine supports morphing. Very nicely, I might add. The big
problem is it cannot morph between objects with different numbers of
points.  Even if the # of points is the same, if the topology isn't
similar, the morph is going to turn into a big mess with the objects
turning inside out and scrambling themselves in an effort to match the
shape of the other object.

The reason that Imagine has the same # of point criteria is that
making a general purpose morph algorithm is VERY VERY VERY VERY
difficult, if not impossible. What Imagine does is take and make a new
object by moving each point an interpolated amount between its
position in object #1 and #2. Thus, the object slowly changes its
shape from one to the other. The crossfades of brushmaps and the
interpolation of texture and attribute parameters add some powerful
bells and whistles.

There can be problems: what if the beginning and ending objects don't
correspond well? They might have the same # of points, but each point
corresponds with an essentially random point on the other object. In
this case, you get a muddled, confused morph. I remember someone
months ago talking about morphing a torus to a sphere, and it turned
out awfully.

The best places to use morph are where you morph an object into
another version of itself. You could take a Coke can object, then
modify it with Magnetism to have a big dent in the side. A morph from
one version to the other would be very smooth: the dent would grow cleanly,
since the correspondence is nearly perfect. You could morph a bird's
wing to a curved version so that the wings will bend as they flap as
opposed to move like they were on a flat hinge.

The big question is WHY can't Imagine just morph from one object to
another?  We're not asking for Terminator II effects, but something a
bit more powerful than what we have now.

The answer is that is is EXTREMELY hard to perform these sort of
morphs.  The problem is that it is difficult to tell the computer WHAT
goes WHERE.  If you were morphing a 3D object of my hand, into, say, a
sword, how would Imagine know where to move the surface? With linear
point interpolation, it just uses the start and end position of each
point. With completely seperate objects, it has NO idea what to
correlate with what. We might think, "Hmm, the top of the middle
finger should become the tip of the sword, the thumb should merge with
the palm as it flattens out into a blade..."  and so on. BUT IMAGINE
DOESN'T KNOW WHAT A MIDDLE FINGER IS!

The way the hi-enders do it is smooth, but takes a lot of work. They
take the two models, and MANUALLY move, say, the middle finger tip to
the blade tip. Then move the thumb to the center of the blade. For
many, many, points, they manually correspond them with the final morph
positions.  When they have many points defined like this, they ask the
computer to interpolate the rest; often they have to correlate extra
points because interpolation doesn't hack it in a certain region. NOW
the computer knows how move one model to form the other. It can then
"tween" the starting and ending positions to make a smooth morph. Of
course, sometimes there is in-between stages that are also defined,
since a linear morph often won't cut it. (ie the modeller might design
a stage that is half hand, half sword, and do two morphs sequentially,
using the half-way point like a keyframe).

This kinda stuff is VERY high end, but you can see its effectiveness in
Terminator II.

I just wanted to talk about this a little bit, because some people were
wondering if Imagine's morph could do something similar.

-Steve

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

##

Subject: Attribute assignment idea
Date: Tue, 16 Jul 91 19:03:21 EDT
From: [e-mail]

I just thought of a truly terrific idea for Imagine 2.0, which would
make a VERY powerful tool and be easy to program.

Here's the problem:

a) You want to have an object like a dented Coke can. Fine, you build your
can, add a dent, wrap an IFF of the Coke logo on it, and render. What
happens? You get a dented Coke can, but THE LOGO IS ALL SCREWED UP
by the dent! This is because the object is no longer a nice cylinder;
its a complex shape that is wrapped differently.

b) you want to animate a piece of paper folding itself into a paper airplane.
Cool, huh? Well, what if you want to add notebook paper lines that fold
as well? oog, kinda tough.

c) You make an anim of an exploding wooden ball. Unfortunatly, as the
 little confetti pieces expand, the wood texture on each piece changes as
the pieces fly through the object's texture space.


The need for even better control of textures and brushmaps, especially
for strangely shaped or morphing objects, is obvious. I would love to
be able to have a "newspaper" object with an IFF brush map of a real
newspaper, but have it folded and wrinked, with dog-eared corners,
just like a real paper. Unfortunately, brush mapping doesn't follow
the surface this way.

Here's my idea: You are given the option to designate SURROGATE
objects to define the attributes, textures, and brushmaps. These
surrogate objects must have the same number of points and the same
topology as the "real" object. Most likely the surrogate would be the
virgin copy of your object before you dent, wrinkle, fold, and explode
it.

What happens is that when your dented object is rendered, the
attributes defined by the surrogate object are used, NOT the
attributes of the dented version. A ray striking a certain face on the
dented can would get its attributes from the corresponding face on the
REFERENCE object. The exact position on the struck face can be
interpolated on the reference object (simple 3 point linear
interpolation) and the color, reflectivity, and transparency can be
determined. This, if a ray strikes the center of a dent, Imagine looks
at the same face on the undented reference object. The brush map(s) and
texture(s) that color the point on the REFERENCE object are applied to
the real object. Voila! The dented area is now properly colored.

The great part is that this can be done with very little work for
Imagine; the overhead is the addition of a 3 point interpolation per
ray that strikes the object. In comparison to the computations that
compute texture color and brushmap location, this extra time is VERY
small. The code required to implement this lookup is also pretty
straightforward; I could see a snag if you wanted to include altitude
mapping, though. (In render quality, not ability to implement).

Do people see how useful this could be? You could animate a ball
shattering, and the coloring from the checker texture you used on the
ball would be intact. A morphing object would make the a brushmap logo
convolute with the shifting surface, as opposed to stay "floating" on
top. If Impulse wanted, you might even have the option of using a
reference object then add EXTRA textures or brushmaps the AREN'T
referenced, so you could add effects that aren't twisted to follow the
"real" objects contours.

This is an idea that suddenly sprang to me. It's completely original,
and I'm kinda proud of it. Anyone have any improvements/suggestions to
it?  I want Impulse to use it, and I already have ideas about some
neat effects that I could do.

The reference object idea could also be used in Lightwave or 3D Pro,
too; its not tied to Imagine specifically at all.

Anyone have any other neat effects that you could do if you had this
ability? Denting, folding, exploding, and other physical morphs are
the obvious situations where it might come in handy.

-Steve

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

##

Subject: Re: World sizes
Date: Tue, 16 Jul 91 19:44:37 EDT
From: [e-mail] (Bob Lindabury - SysAdm)

Udo K Schuermann <rutgers!wam.umd.edu!walrus> writes:

> Steve Worley writes:
> > The resizing will NOT, repeat will NOT, change your camera views!
>
> But it will NOT change scale-related values in textures either.  THAT
> is a problem!  Try scaling an object with a nice wood-texture:  The
> rings will become disproportional to the object :-(

This is a very good point.  I recently (well, not that recently)
posted a starburst object to this list along with an explaination on
how to use it.  This "tutorial" suggested that you rotate the
starburst object as you sized it up and down in the scene.  Well,
shame on me!  I tried just that *after* I posted the object and
tutorial and I found that the radial filter texture I had applied to
the disk portion of the object did NOT size proportionally with the
disk!  ARGH!

This shouldn't happen should it?  Is this a problem with Imagine or
am I being a putz expecting the texture to proportionally size with
an object?

Is there another way to achieve the desired effect?

-- 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: Textures/FX
Date: Tue, 16 Jul 91 21:04 MST
From: "Vax Headroom (Dave E Martin)" <[e-mail]>

Since textures and effects are loadable, what are the chances of
impulse releasing the format so people could make their own?

Are they loadable code routines or formulas or what?

I would like a texture similar to the one used in DKBTrace to get clouds.

I've tried camoflauge, but it doesn't quite give the same effect.

##

Subject: Morphing and the like
Date: Tue, 16 Jul 91 23:42:25 PDT
From: Daryl T. Bartley <[e-mail]>

 Okay, a quick question on the hand-morphing-into-a-sword thing...would it be
 possible (or practical) to make more than one intermediate object? Say the
 final 'sword' (or whatever) has half as mnay points as the hand (I know, it
 wouldn't, but just for sake of argument)...you make an intermediat object, same# of points as the hand, and like 'line up' some of the points so it is a flat surface...then, make a second object just like that, but delete the excess points that are lined up, hopefully ending up with the same face-point count as the final sword (Okay, so the sword might have to be a little overcomplicated), then
 replace the 2nd object (hand-sword) with the 3rd...if done right, the switch won't show up, and it can morph on into the sword.

 Now, if anyone can figure out what the heck I just said (including ME!), then
 would it work?

 P.S. Oh, and is there any chance when the 'sneak peek' pics of the DCTV-T2 animare let out, that there might be some lowly HAM versions? (i.e. that I can see?)

 A t D h V a A n N k C s E,

Daryl Bartley
[e-mail]

##

Subject: "projection maps," are they possible?
Date: Wed, 17 Jul 91 00:16:08 CDT
From: mit-eddie!harvard!scubed!pro-party.cts.com!seanc (Sean Cunningham)

Someone on Studio Amiga (817-467-3658) was having trouble getting a disco
ball to produce the little light reflections on the walls in the room he
designed.  Because Imagine doesn't do backward-beam raytracing, this kind
of effect is impossible w/o doing some kind of mapping.

Someone else posted that the effect could be simulated by mapping an IFF of
some spots onto a light.  How is this done???

I would have asked the person who suggested this myself, but due to a
rather _LARGE_ phonebill I won't be calling there for quite a while :)

HELP!

Sean
                                        /\
 RealWorld: Sean Cunningham            /  \     "Doing our business is what
      INET: [e-mail]   VISION                Amigas are for."
     Voice: [phone]             \  /
                                 //     \/      "Hasta lavista, baby."
    KEEP THE COMPETITION UNDER \X/   GRAPHICS         the Terminator, CSM101

##

Subject: Re: Attribute assignment idea
Date: Wed, 17 Jul 91 10:01:09 EDT
From: Mark Thompson <[e-mail]>

Sean Schur writes:
> Sorry Steve, this idea is not "completely original". This is a feature
> that is commonly implemented in high-end rendering programs.

Yes, wrapped textures as apposed to projected textures have been around for
some time. They aren't to common on lower end rendering programs because of
the complexity of actually using them as well as the additional difficulty of
implementing them well.  Last years Siggraph had a paper "Performance-Driven
Facial Animation" in which an image of a face was assiged control points and
"wrapped" onto a model of a head. The control points could then be animated to
give the face various expressions. The technique was used in the film "The
Audition" shown at the Film and Video show in which a bulldog's facial image
was wrapped onto a character's head and animated.

I might add that although LightWave currently only uses projected textures,
objects that are scaled in Layout (the animation module) automatically scale
the texure.
|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
|      `       '        Mark Thompson                 CONCURRENT COMPUTER  |
| --==* RADIANT *==--   [e-mail]        Principal Graphics   |
|      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   |
|     Productions       [phone] [phone]   & General Nuisance   |
|                                                                          |
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

##

Subject: Re: Attribute assignment idea
Date: Wed, 17 Jul 91 05:56:15 PDT
From: [e-mail]

>
> I just thought of a truly terrific idea for Imagine 2.0, which would
> make a VERY powerful tool and be easy to program.

  <stuff deleted>

> Here's my idea: You are given the option to designate SURROGATE
> objects to define the attributes, textures, and brushmaps. These
> surrogate objects must have the same number of points and the same
> topology as the "real" object. Most likely the surrogate would be the
> virgin copy of your object before you dent, wrinkle, fold, and explode
> it.

  <stuff deleted>

> What happens is that when your dented object is rendered, the
> attributes defined by the surrogate object are used, NOT the
> attributes of the dented version. A ray striking a certain face on the
> This is an idea that suddenly sprang to me. It's completely original,
> and I'm kinda proud of it. Anyone have any improvements/suggestions to
> it?  I want Impulse to use it, and I already have ideas about some
> neat effects that I could do.

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

Sorry Steve, this idea is not "completely original". This is a feature
that is commonly implemented in high-end rendering programs. As a matter
of fact, for those of us using high-end and middle-end (or whatever you
want to call Imagine, I wouldn't say low-end) programs, this question
is inevitably one of the first to pop up. Does it do projection mapping
or true image mapping? Because this is exactly what Imagine or Lightwave
or whatever other Amiga renderer you use does (except Journeyman). It
does not actually "wrap" an image onto an object, it "projects" it
into the general area of the object.

For a specific example of what I am talking about (I'm sure many of you
have tried this already) wrap some brush onto an object, say a sphere,
in Imagine. Then render a simple animation having the sphere explode
and watch carefully as the pieces move outward. You can literally see
the pieces move through the image. The image stays put within the
parameters of the brush axis you defined. You can even see when
the fragments leave the area that the brush is projected into because
they revert to their defined base color.

It is a wonderful idea, and especially would be wonderful if Imagine
would implement it. That would make Imagine unique in it's class.
But unfortunately it is not original. I am almost certain that Journeyman
works properly in this way. You build your object then stamp a brush
where you want it to go on a quick-rendered version of the object.
That brush wraps onto the object at that point and sticks. I can't
say absolutely that they have true brush wrapping, I haven't played
enough with distorting objects yet to tell. Maybe I'll play with it
any let you know if anyone is interested.

=======================================================================
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: "projection maps," are they possible?
Date: Wed, 17 Jul 91 11:35:21 EDT
From: [e-mail]

--------

mit-eddie!harvard!scubed!pro-party.cts.com!seanc (Sean Cunningham)
asks how to make a disco ball.

You're right, Imagine does NOT do backwards ray tracing other than
shadow casting. (You cannot shine a light on a reflective surface and
have it illuminate something with the reflected light). Backwards
raytracing is VERY expensive (computationally) to do, and just isn't
feasable even with a 100MHz 68040.

However, you could get a disco ball effect in less direct ways...

One way to do this would be to make an object that was a sphere with
one or more brush maps on it. You build a brushmap that is mostly
pure black with small rectangular white holes on it. You could scatter them
randomly or make a nice grid. Wrap the brushmap as a TRANSPARENCY map
on the sphere. Now, in the Stage editor, put a light inside the
sphere. Voila! you have little beams of light coming out of
the sphere. You can slowly animate the spophere rotating to make the
spots move around.

This is only an APPROXIMATION to a disco ball, though, because all
of the light is radial to the ball. With a real ball, the light from
(say) a spotlight isn't reflected radially at all. Also, the reflected
light should really be coming from only one side (with only one spotlight)
whereas this cheesy model is uniform.

But at least it's something that would work to a good approximation.

-Steve

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

##

Subject: Re: attribute assignment idea
Date: Wed, 17 Jul 91 12:06:05 EDT
From: [e-mail]

Sean Shur writes "Sorry, Steve, but your idea isn't completely original."

-----


It is for me!  :-)

Yes, it's obvious that hi-end systems must use a method of defining
complex texture/image spaces that follow the object's complex shape
or motion.

Mark Thompson says that one method uses control points to define
the placement. Actually, my idea uses the same method, except EVERY
point is a control point.

It's obvious that the idea of making an alternative texture/image space
for complex attribute assignment HAS been done before. What IS original
is the method I'm suggesting to define it. It has limitations, but is
particularly easy to use (for the animator) and simple for the programmer
to implement. (Then again, I'm not doing the coding!)

Talking about hi-end features, the author of 3D Pro is bragging that
3D Pro 2.0 will have radiosity. Yow! Radiosity is a very complex method
of lighting scenes that is particularly realistic; areas like room corners
and alcoves are more dimly lit than the middle of the room. It is a VERY
subtle feature that makes startlingly lifelike renderings. The trick is
that it is so expensive (hard to code well, and a LOT of extra compuation in
the render) that an Amiga (or Apple or NeXT or IBM) implementation that
would actually be useful seems nearly impossible. Of course, what do I
know?  One other feature of radiosity is that if you have a scene with
the radiosity lighting computed, you can render different views without
re-computing radisity. This makes room fly-throughs very popular..


-Steve

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

##

Subject: Imagine to Lightwave
Date: Wed, 17 Jul 91 11:17:24 CDT
From: [e-mail] (Randy Carris)

I'm looking for a way to cpnvert Imagine objects to Lightwave objects.  I know
that with Interchange you can convert Turbo to Sculpt and load them that way,
but do you loose anything.  I obviously don't want the shape distorted etc.
Can you keep the attributes or do they get lost or screwed up?  Is there any
dirrect way to do these conversions?  I like the object creation abilities I
get in Imagine (when I ever get better control over them), but I con output to
tape a lot easier with our Toaster.

Thanks,
Randy Carris
[e-mail]

P.S. ,
I got my plane tickets to Las Vegas for July 28-Aug. 2  I'm a student volunteer
at SIGGRAPH this year, who all is going to be there.  Anyone else on here going
as a volunteer?

##

Subject: Imagine's capabilities (was Re: "projection maps," are they possible?)
Date: Wed, 17 Jul 91 13:09:50 EDT
From: [e-mail] (Brian Reynolds)

    I'm a little confused about Imagine's capabilities.  I don't
currently own an Amiga or Imagine.  I have used several PD raytracers
(currently I use rayshade 4.0 beta), and I have read about half of
Andrew Glassner's "An Introduction to Ray Tracing."  I don't
understand the statement that Imagine does not do backwards
raytracing.  Glassner's book defines all practical ray tracers as
backwards ray tracers.  A ray is traced backwards from the eye point,
through the image plane, and into the scene, where it may hit an
object, and be reflected towards another object, empty space, or a
light source.  The color of a pixel is determined by combining the
color of any objects hit by the ray and the color of light hitting the
objects, adjusting for distance and the number of bounces.  It seems
to me that any ray tracer using this approach should get illumination
from reflected light for free since an object that receives light from
a relection off a mirror type object would be almost (assuming
imperfect mirrors) as bright as an object that is directly illuminated
(at the same ray path distance), and brighter than an object that
receives light after a reflection off a non-mirror object.  (Of course
now I'll have to test rayshade and see if it works.)  A forward ray
tracer is defined as one that traces a ray from the light sources
outward, possibly hitting an object, reflecting around, and eventually
reaching the eye point.  Because the vast majority of rays leaving a
light source will never do anything interesting, forward ray tracers
aren't considered practical.

    Can someone clear up exactly what Imagine does when it traces
a scene?

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

##

Subject: Altitude Maps
Date: Wed, 17 Jul 1991 00:28 CDT
From: "Ryan Brooks, 040 Wanna-be" <[e-mail]>

Is there anyway to do real altitude mapping with imagine?  I'd like to
take a gray scale image, map it onto a plane, and have the plane actually
change shape, not just appear to have depth, as Imagine appears to do.
Any help would be greatly appreciated.  Ryan Brooks
[e-mail]

##

Subject: Re: attribute assignment idea
Date: Wed, 17 Jul 91 12:46:50 EDT
From: Mark Thompson <[e-mail]>

Steve Worley writes:
> 3D Pro 2.0 will have radiosity.

I hope its a two pass algorithm to handle the specular component :-)

> it is so expensive (hard to code well, and a LOT of extra compuation in
> the render) that an Amiga (or Apple or NeXT or IBM) implementation that
> would actually be useful seems nearly impossible.

Yeah, I don't think we will be seeing too many radiosity animations of
any length.

> One other feature of radiosity is that if you have a scene with
> the radiosity lighting computed, you can render different views without
> re-computing radisity. This makes room fly-throughs very popular..

Unfortunately, nothing in the room can move except the camera. But for
architectual/interior design applications, this limitation doesn't usually
matter.
|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
|      `       '        Mark Thompson                 CONCURRENT COMPUTER  |
| --==* RADIANT *==--   [e-mail]        Principal Graphics   |
|      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   |
|     Productions       [phone] [phone]   & General Nuisance   |
|                                                                          |
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

##

Subject: Re: World sizes
Date: Wed, 17 Jul 91 12:09:29 -0400
From: Udo K Schuermann <[e-mail]>

>> Steve Worley wrote:
>>> The resizing will NOT, repeat will NOT, change your camera views!

> I wrote:
>> But it will NOT change scale-related values in textures either.  THAT
>> is a problem!  Try scaling an object with a nice wood-texture:  The
>> rings will become disproportional to the object :-(

Bob writes:
> [Description of scaling the StarBurst object removed]
> This shouldn't happen should it?  Is this a problem with Imagine or
> am I being a putz expecting the texture to proportionally size with
> an object?

Not your fault, Bob.  Impulse didn't think things through quite as
much as they should have with respect that that.  Ideally, sizing an
object should affect textures equally.  The problem lies in the fact
that Imagine cannot know which of the texture parameters are
size-related and which aren't.  When the object is scaled, the
parameters are unaffected.
    If Impulse redefines size-related parameters to where they
represent parent-object relative size, the problem would go away and,
in my view, would require very little code to implement.
    Ex:  set the size of the rings in a wood-texture to 0.05,
meaning 5% of the object's axis size (perhaps avg of xyz?).  If the
object is scaled up, the texture will be scaled by default.

> Is there another way to achieve the desired effect?

Other than manually altering all the involved texture parameters, no.

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

##

Subject: gnicartyaR sdrawkcaB
Date: Wed, 17 Jul 91 12:42:30 PDT
From: Daryl T. Bartley <[e-mail]>

 You mean, it can't reflect light off of one object onto another? I could have
 sworn I saw some 're-cast' light in one trace I did (And yes, to prove how much of a nutter I am, it took 26 hours on my stock 2000), from some glass in front of a mirror...could this just have been the light going through the glass and hitting the floor, or was it coming off of the mirror and back through the glass?

 Darn, if it can't reflect light, that shoots down another 1000 project ideas. ::-)


Daryl Bartley
[e-mail]

##

Subject: Re: Imagine to Lightwave
Date: Wed, 17 Jul 91 15:58:10 EDT
From: Mark Thompson <[e-mail]>

Randy Carris writes:
> I'm looking for a way to convert Imagine objects to Lightwave objects.  I
> know that with Interchange you can convert Turbo to Sculpt and load them
> that way, but do you loose anything.  I obviously don't want the shape
> distorted etc.  Can you keep the attributes or do they get lost or
> screwed up?  Is there any dirrect way to do these conversions?

Well Syndesis is currently beta-testing conversion modules for both LightWave
and Imagine. I am testing the LightWave module. Unfortunately, InterChange
currently converts all objects from any format into its own internal
simplified format. While this does not alter the shape of the object, it
seriously limits the variety of surface attributes that may be passed.
Syndesis has proposed expanding their surface attributes support and you can
bet your butt that I was screaming all for it. As it stands now, it is
limited to a color definition, smooth/flat, shiny, and dull. I don't even
think it handles transparency.

Also of note, the upcoming LightWave 2.0 release will include a complete
object import/export facility done by John Foust of Syndesis.

> P.S. ,
> I got my plane tickets to Las Vegas for July 28-Aug 2. I'm a student
> volunteer at SIGGRAPH this year, who all is going to be there.

I'll be there. Currently I am attempting to quickly organize a small
gathering of Amiga graphics enthusiasts at the show.
|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
|      `       '        Mark Thompson                 CONCURRENT COMPUTER  |
| --==* RADIANT *==--   [e-mail]        Principal Graphics   |
|      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   |
|     Productions       [phone] [phone]   & General Nuisance   |
|                                                                          |
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

##

Subject: Bug in 1.1
Date: Wed, 17 Jul 91 23:14:30 EET DST
From: Juha Kallioinen <[e-mail]>

I think I'v discovered a bug under Imagine 1.1

When you're in the Forms editor and by accident or intentionally make the
object look like a line in the TOP-view the result is a guru. Seems like
the forms editor cannot handle zero-width; that's acceptable but it should
pop up a requester, not a guru.

The system I'm running is A500 w 3 megs, 1.3 kick, 68010 (sigh)
So, there it is.

--
o-----------------------o----o--Q: How many IBM cpu's does it take to do---o
| [e-mail]/  __ |     a logical right shift?                  |
| Juha Kallioinen     /_ /// |  A: 33.  1 to hold the bits and 32 to push  |
o--------------------o \XX/  o-----the register.---------------------------o

##

Subject: Help! I'm morphing!
Date: Wed, 17 Jul 91 16:23:50 EDT
From: [e-mail] (Alan Price)

Hey, here's something which has probably been covered, but I missed it
if it has.

In all the time I've been using Imagine, I finally got around to seriously
experimenting with morphing attributes and textures. I set up a series of
spheres and morphed through four different attribute/texture settings over
a length of 60 frames. The results varied:

Going from one wood texture to another worked very nice, both the attributes
(color and specularity) AND the woodgrain texture transformed smoothly.

Going from wood texture to angular did not work. The attributes morphed, but
the new colors of the angular texture seetings did not appear until the
transformation was entirely complete. (The wood grain did slowly disappear,
however.)

Going from angular to disturbed did not work too well either. The colors
in the angular texture slowly gave way to the color of the disturbed sphere,
but the disturbed texture did not appear until the end of the morph. All of
a sudden POP! it's disturbed.

From this short test I am assuming that Imagine morphs object attributes
(color, specularity, hardness, rough, shiny, reflect) smoothly and effectively.
But does not do well morphing TO a certain texture. (Actually since wood was
successful, it seems appropriate that bricks, and grids may work well, too.
But does dots? or others?) Disturbed, and angular did not morph. Or am I
doing something wrong?

Another question:

I clicked on the []Genlock sky option in the globals attributes and nothing
changed. I wanted the gradient fill that I assigned in the Horizon and +&-
Zenith settings to appear in my reflective objects, but to have a black sky.
The sky rendered with the gradient fill instead. What am I doing wrong?

An observation: There has been interesting mail about both morphing and
re-sizing objects, etc. On the subject of re-sizing the texture settings
along with the scaling of an object, there seems to be mostly negative
opinions about it. Well here's one advantage:

I have a plane with a brick texture applied to it. It's called 'brick wall',
of course. I load it in the stage editor repeatedly, creating many walls
for an alley-like background for an animation I'm working on. I can easliy
resize this plane to build 'buildings' and the bricks always stay the same
size. It would be a pain to have to make all thes different wall objects in
this particular case. I'm sure there are other situations where you use the
same object multiple times, rescale it, but want to keep the same texture
size. A wood-pile, trees, or chips hewn from the same marble, are some.
I'm sure there are more instances where it would be better to have the
texture resize along with the object, but if this change were made in 2.0,
perhaps there would be some way of making it an option rather than a fixed
way of scaling.

AP.

##

Subject: Help! I'm morphing!
Date: 17 Jul 91 16:14 -0500
From: "Jeff A. Bell" <[e-mail]>

> Another question:
>
> I clicked on the []Genlock sky option in the globals attributes and nothing
> changed. I wanted the gradient fill that I assigned in the Horizon and +&-
> Zenith settings to appear in my reflective objects, but to have a black sky.
> The sky rendered with the gradient fill instead. What am I doing wrong?

Probably nothing.  I had exactly the same question a while back.  As it turns
out, the 'Genlock sky' option will not work if you're rendering in 24bit
(which was my 'problem').  Like you, I wanted the horizon gradient to show
up in objects that I was rendering, but I didn't want the sky to show up..
My solution was to give up on ray-tracing (takes too long to ray-trace an
animation on a stock 68000 machine.  Ugh), and go to scanline, using many
brush maps and a reflection map to make up for the lack of true reflection.
(After having seen your 'SpiralStairs' pic, I'm sure you realize all this.
Sorry for babbling.  Nice job on that pic, BTW.  I liked the way the light
and reflections played of the hardwood floor.  The stairs looked AWFUL
tedious to model).

Later Alan,

Jeff
--
[e-mail]

##

Subject: Displacement mapping
Date: Wed, 17 Jul 91 17:23:02 EDT
From: [e-mail]

Ryan Brooks asks:

Is there anyway to do real altitude mapping with imagine?  I'd like to
take a gray scale image, map it onto a plane, and have the plane actually
change shape, not just appear to have depth, as Imagine appears to do.
Any help would be greatly appreciated.  Ryan Brooks

---

Imagine does not support displacement maps (where the surface is actually
changed). However, if you just want to map altitudes onto a plane, there
is a very practical alternative: TTDDD by Glenn Lewis will allow you
to define objects by scripts or programs. A plane is particularly
easy to set up as an object, though you'll have to figure out how to
read an IFF file and output the intensity as a string of numbers.

The steps I would take would be to write a quick C program that
takes a filename and an x and y dimension as input. It reads the
filename in as an array of bytes representing intensity. It outputs
a TTDDD file containg an XY mesh with the individual Z components
given by the input file. With this root program completed, I would
make a AREXX script to call The Art Department Professional, load
your IFF, convert it to gray scale, and output in raw format (Sculpt).
Then the AREXX script would call my program, which outputs a TTDDD
file, then the script would run WriteTDD. Boom, a displacement
map like you asked for. The AREXX script would be about 8 lines long,
the C code maybe 20.

If you really want such a program, I could write one for you pretty
quickly.

The final answer is "No, Imagine can't use displacement maps. But WE can!"

-Steve


Geez, a lot of good posts the past few days. We're up to about 980...
I think my "next week" prediction was pessimistic.

Oh, the prize for the 1000th post is a good idea. The 1000th LEGIT
post (by my standards) will be a set of floppies containing a bunch of
useful Imagine files, such as objects, brushmaps, tutorials, etc.
Most of this will be from the hubcap FTP site, but some of it is stuff
I haven't gotten around to posting yet. Why this prize? I'm copying a
set of disks like this for Impulse and Rick Rodriguez and an extra
copy is easy to make. It will be worth a few bucks to see how you guys
try to come up with legit ideas to talk about. Note that the posts
have to be legit: a few lines saying "Yes, I agree with you
completely, Fred!" is an obvious filler. :-) Also, note that 1000 is
defined by the order in which I get them, and some people might have a
different order because of mail delays.


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

##

Subject: Re: Morphing
Date: Wed, 17 Jul 91 14:32:11 CDT
From: mit-eddie!harvard!scubed!pro-party.cts.com!seanc (Sean Cunningham)

I'm doing some extensive experimentation right now trying to get a similar
effect as seen in T2.

So far, I've been doing puddles that grow, stretch, and "crawl" using
digitized reflection maps.

I've just about determined that I'll have to continue work in another
package to progress much further.  It's not that any other package does
better morphing (they all do point morphing), it's that the smoothing
sometimes turns off on faces around the "lip" of my puddle, giving a
faceted look.  Also, I need to be able to bend points, where the DETAIL
editor will only let me drag points.

An interim solution may be to do all of my distortion modelling in
Modeler3D and then import the object into Imagine via Interchange.

Sean
                                        /\
 RealWorld: Sean Cunningham            /  \     "Doing our business is what
      INET: [e-mail]   VISION                Amigas are for."
     Voice: [phone]             \  /
                                 //     \/      "Hasta lavista, baby."
    KEEP THE COMPETITION UNDER \X/   GRAPHICS         the Terminator, CSM101

##

Subject: Colorburst
Date: Wed, 17 Jul 1991 22:04:04 -0500
From: Donald Richard Tillery Jr <[e-mail]>

ATTENTION 24 BIT ENTHUSIASTS!!!!!!!!!!!!!!!!

Finally, after much ado, I have in my posession a _COLORBURST_!!

It is a developer model, so it is officially only FCC Class A approved,
but I haven't noticed any interference with my TV, so I assume it will
pass Class B whenever the FCC gets on the ball.

Right now I'm using VI on top of a 24 bit image as a backdrop (the
first time I've ever been pleased with VI :-), and it looks GREAT!
(Some of the letters are a bit hard to read here and there :-)

I'll post a full review (I don't usually hold any punches, so you can
be sure of a hard nosed review) in a few days (probably early next week)
to comp.sys.amiga.reviews, and messages telling everyone to read it
to comp.sys.amiga.graphics, and probably alt.binaries.pictures.d or
any other groups I can find with the words "graphics" or "pictures."
(ANY SUGGESTIONS?)

Anyway, this is at least ONE MAST product that IS NOT vaporware!

Rick Tillery ([e-mail])

##

Subject: SIGGRAPH '91 (was Re: Imagine to Lightwave )
Date: Wed, 17 Jul 91 20:44:40 EDT
From: [e-mail] (Brian Reynolds)

    I'm going to be at SIGGRAPH '91 also.  A couple of months ago
a get together of imagine mail list members was suggested.  Anyone
want to organize it?  (I'll do it if there's enough interest in a
meeting, but no interest in organizing it. :) I'll be arriving
Saturday and plan on being at registration as early as possible
Sunday.  If anyone's interested I can try and figure out how to
arrange a BOF (birds of a feather) meeting after I'm done at
registration.  I don't know what kind of facilities SIGGRAPH is
offering for BOFs this year (it varies from year to year).  See you
there.

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

Steve does this count towards the 1000th message? :)

##

Subject: Re: "projection maps," are they possible?
Date: Wed, 17 Jul 91 23:04:22 CDT
From: mit-eddie!harvard!scubed!pro-party.cts.com!seanc (Sean Cunningham)

(Steve Worley explains how to do the disco ball effect...)

I kinda thought that might be a way to do it, but my attempts at filter
mapping have all been duds.

My only question is: will a SPHERICAL light set at 255 all be bright enough
to cause visible spots on the surfaces in a scene?

To be honest, I'm not really interested in doing a disco ball.  That was
someone else.  What I want to do is get the effect of light reflected off
the surface of water.  Like when you're under water in a pool and see all
the little criss-crossings on the bottom.

Anyone out there have a subscription to MILIMETER magazine???  That picture
of the Sea Cow created with METABOLS for that HDTV special was hot wasn't
it?!  He was using a projection map to do the reflections on the Sea Cow's
back.  It looked increadible.

Sean
                                        /\
 RealWorld: Sean Cunningham            /  \     "Doing our business is what
      INET: [e-mail]   VISION                Amigas are for."
     Voice: [phone]             \  /
                                 //     \/      "Hasta lavista, baby."
    KEEP THE COMPETITION UNDER \X/   GRAPHICS         the Terminator, CSM101

##

Subject: Radiosity
Date: Wed, 17 Jul 91 23:14:37 CDT
From: mit-eddie!harvard!scubed!pro-party.cts.com!seanc (Sean Cunningham)

I've seen some fairly realistic renderings created with radiosity (like the
factory scene in one of HP's workstation brochures).  But from what I've
been reading lately, it's not really the solution for realistic
renderings...at least not by itself.

Radiosity isn't supposed to be able to do shiny or reflective surfaces,
it's more tuned to matte finishes.  But I've seen some _very_ impressive
renderings that used a combination of radiosity and raytracing.  Beautiful
shadows and shading with raytracing's true reflections.

Did Bill Coldwell give any approx. times for computing the lighting model
to use for a scene (heh, listen to me...sounds like I really know what I'm
talking about, eh :)   That first part of the render process where it
figures out the way the objects are lit, and how their surface gives off
light is supposed to take quite a long time...but further images of the
same scene are supposed to be fairly quick, hence your "walkthrough"
reference.

It'd be neat wouldn't it.  Sit there and wait for the lighting to be set up
for as long as you'd normally wait for a whole animation to be traced...but
then be able to change views and only have to wait for the screen to update
(probubly not _this_ fast...but I can dream, right?)

Sean
                                        /\
 RealWorld: Sean Cunningham            /  \     "Doing our business is what
      INET: [e-mail]   VISION                Amigas are for."
     Voice: [phone]             \  /
                                 //     \/      "Hasta lavista, baby."
    KEEP THE COMPETITION UNDER \X/   GRAPHICS         the Terminator, CSM101

##

Subject: How to precisely add objects (and all that)?
Date: Thu, 18 Jul 91 15:46:05 EET DST
From: Juha Kallioinen <[e-mail]>

I'm making a house which has many rooms and thus many walls, what is the
best way of adding walls to the house so that they leave no cracks to the
corners ?

Now I always have to zoom as much I can to move a new wall precisely to a
corner. But it is pretty tedious to zoom around many many times and then
back again.

Stupid question probably, I'm sure there's a simple solution to this, but
I can't figure it out, so heeelp.

(I'm thinkin of making a house-walk-through)

--
o-----------------------o----o---------------------------------------------o
| [e-mail]/  __ | Diplomacy is the art of saying "nice doggy" |
| Juha Kallioinen     /_ /// |       until you can find a rock.            |
o--------------------o \XX/  o---------------------------------------------o

##

Subject: Re: Radiosity
Date: Thu, 18 Jul 91 10:08:15 EDT
From: Mark Thompson <[e-mail]>

> Radiosity isn't supposed to be able to do shiny or reflective surfaces,
> it's more tuned to matte finishes.  But I've seen some _very_ impressive
> renderings that used a combination of radiosity and raytracing.  Beautiful
> shadows and shading with raytracing's true reflections.

Yes, this is the two-pass radiosity algorithms I have refered to in a couple
of recent posts. Some of the images I have seen using the two pass method
are just ridiculously real looking. Wow.
|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
|      `       '        Mark Thompson                 CONCURRENT COMPUTER  |
| --==* RADIANT *==--   [e-mail]        Principal Graphics   |
|      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   |
|     Productions       [phone] [phone]   & General Nuisance   |
|                                                                          |
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

##

Subject: Disco ball light strength
Date: Thu, 18 Jul 91 11:49:46 EDT
From: [e-mail]

Sean Cunningham asks:

My only question is: will a SPHERICAL light set at 255 all be bright enough
to cause visible spots on the surfaces in a scene?

-----

No, I bet that a light of 255 won't work too well.


So crank it up to 2000! For a long time I thought lights were limited to
255: I was wrong. (Its logical that lights can be as bright as they want).
Values above 500 or so are pretty severe: they cast strong shadows,
like a very sunny day. Above 3000 or so and it looks like your world is
lit by nuclear weapons.

-Steve


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

##

Subject: No-seam house walls
Date: Thu, 18 Jul 91 12:05:37 EDT
From: [e-mail]

Two methods. First, you can design your walls to be nice dimensions
like 100 or 1024 instead of 383.38. Place your axis at the corner
of each wall. Then, to get seamless joins, use "snap to grid" in
the project editor which will instantly adjust your walls to a perfect
fit. [As long os your wall lengths are muliples of the grid line spacing.]

The other cheesy option is to make your walls too big. Then INTERSECT them.
You get a mess BEHIND the wall, but if you don't look there, you'll never
see it.


      wall 1           |
     ------------------+--
                       |
                       | wall 2
                       |
                       |
                       |
      camera X

-Steve

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

##

Subject: Re: How to precisely add objects (and all that)?
Date: Thu, 18 Jul 91 11:45:40 CDT
From: Wayne Haufler 283-4160 <[e-mail]>

Hello fellow imagineers,

Juha Kallioinen writes:
> (I'm thinkin of making a house-walk-through)

That reminds me:  Whatever happened to The Rooms Project?
Is all involved busily but quietly working on their pieces?
Or has it gone by the wayside?  I wonder if I get ALL the messages.

Perhaps someone (Jake?) could fill in the newcomers.  At least it sounds
like Juha is new and would like in on it, if it is not too late.  :)

Unfortunately, MY Amiga 2000 has been out of commission for two months
now!  :( That's EIGHT WEEKS (mostly weekends) of unfulfilled dreams,
including the work on the Rooms Project.  I am itching to get back into
it.  Sorry guys, I had to gripe to somebody.

> I'm making a house which has many rooms and thus many walls, what is the
> best way of adding walls to the house so that they leave no cracks to the
> corners ?

This problem might be solved with an improvement I would like to see
for Imagine (2.0?).  This would involve a way in the Stage Editor
and/or Detail Editor for specifying an object to move, a direction to move,
(or move it interactively with the mouse), and one or more objects to avoid.
The object could then be moved until it collides with one of the 'to avoid'
objects.  Or Imagine could automatically move the object in the specified
x, y, or z direction until it collides.  Of course, the legal directions
would probably have to be limited to along one of the three global axis.
This would require a collision detection algorithm, but it shouldn't be
too difficult to implement, since it is only active on demand, and only
for a specified set of objects.  Am I right? :/

This would allow one, say, to accurately and easily place a vase on a
table, or butt two walls together without cracks (though exact corners
may be difficult), or assemble an object precisely from existing
pieces.

An alternative approach that may be easier to implement and more general,
would be to allow the user to specify one point on the 'to move' object
and another point on a reference object, and then Imagine could move the
object such that the two points coincide.  This may not be suitable for
many situations, like butting two walls together in a 'T' where there
are no points on the reference object to coincide with.

I hope these descriptions are understandable.  I don't claim to be familiar
with collision detection algorithms, so I may be all wet. :^)
Does anybody have any comments, embellishments, or improvements to these
ideas?

Ah!  Maybe a more immediate solution might involve TTDDD (?).  But I think
TTDDD only works with one object at a time and not positions of objects
within a scene, but I really don't know for sure.

                Wayne

===============================================================================
          ____  Wayne A. Haufler     ([e-mail])
\\   /\\   /\\   //    McDonnell Douglas Space Systems Company - Houston Div.
 \\ /--\\ // \\ //--[Christian/Lutheran/SW Engineer/Amigan/Tenor/Violist/Single]
  \//   \//   \//___   "Exploring the Use of Computer Graphics and Animations"
        //                   "To Serve and Support Christian Endeavors"
       //
            Standard Disclaimers Apply.
===============================================================================

##

Subject: Re: How to precisely add objects (and all that)?
Date: Thu, 18 Jul 91 10:38:33 PDT
From: [e-mail] (Glenn M. Lewis ~)

>>>>> On Thu, 18 Jul 91 11:45:40 CDT, Wayne Haufler 283-4160 <[e-mail]> said:

Wayne> Ah!  Maybe a more immediate solution might involve TTDDD (?).
Wayne> But I think TTDDD only works with one object at a time and not
Wayne> positions of objects within a scene, but I really don't know for
Wayne> sure.

Wayne>                 Wayne

    I just thought I would clear up a little bit of confusion here.
I wrote TTDDD (Textual Three Dimension Data Description... I thought up
the first `T', Impulse thought up the "TDDD"), and it includes two
utilities.  One translates TDDD into TTDDD files (ReadTDDD) and the
other goes the other direction (WriteTDDD).  The description contains
all information about all the objects in the file, including grouping
information.  Position, points, edges, faces, colors, etc. are all in
there.  Steve Worley and I have been brainstorming about morphing and
dynamic constraint programs that we could create with TTDDD as a basis,
but you can also simply load the results of ReadTDDD into a text editor,
change any information you desire (like position or individual points,
etc.) and then pump that back through WriteTDDD to get your modified
object.
    If you haven't taken a look at the TTDDD package, you can grab
it off of ab20.larc.nasa.gov [128.155.23.64].  Take a look at FILES.Z
for its exact location.  It is shareware.  Registration is $10, and
entitles you to the source code to ReadTDDD and WriteTDDD, and two
utilities, SQuad (super quadric object generator) and TSTeX (TeX font
conversion into 3-D objects).

                        -- Glenn

##

Subject: Attributes
Date: Thu, 18 Jul 1991 12:11 MDT
From: [e-mail]

Greetings fellow Imagineers. (Isn't that term taken by Disney or something?)

When you save your attributes it only saves the color etc. but it doesn't save
the set up texture or brush maps along with the set up.  My question is, how
can I get it to do this?  I would like to have some standard attributes set up
complete with textures/brushmaps that I use frequently that I can just load.
I would rather not have to type in all my values for my textures each and every
time I use it for objects that are textures the same.

I also have a question about the Firecracker 24 from Impulse.  What kind of
monitor does it require?  Does it have to be able to sync down to 15mhz?  I
have a 3000 and a seiko CM1440 (nice picture by the way) but it doesn't sync
down to 15.  Also, I would assume that in using the FC 24 you have to bypass
the "flicker fixer" in the 3000, if not, how does it work?  Also, (just one
more question I promise)  Can you display 24 bit anims on the FC or do you
have to save it somehow to tape one frame at a time and then view it?  Does
anyone have a FC and can tell me basically what the setup is.  I have read
a number of reviews on how great it is, but nothing on the system requirements.

Thanks alot.

John Zollinger
[e-mail]

"IBM? Is that Amiga compatible?"

##

Subject: How to precisely add objects
Date: Thu, 18 Jul 91 13:32:53 -0400
From: Udo K Schuermann <[e-mail]>

Juha Kallioinen writes:
> I'm making a house which has many rooms and thus many walls, what is the
> best way of adding walls to the house so that they leave no cracks to the
> corners ?

Just overlap the walls enough to make sure there are no cracks:
                   |    |
        -----------+-.  |
                   | |  |
        -----------+-'  |
                   |    |
No need to line the edges up precisely, unless you need to follow
exact measurements.

> (I'm thinkin of making a house-walk-through)

If you're interested in joining such a project, Alan Price (with a
little help from me) is soon going to revive the 'House Project.'
New and revised.  True fun for The Creative Mind.  Give it a few
days or so ...

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

##

Subject: Re: No-seam house walls
Date: Thu, 18 Jul 91 15:40:46 EDT
From: Mark Thompson <[e-mail]>

I am surprised no one has mentioned the method I would have used. Create a
2D outline of the floorplan and extrude it up. Then simply pop a ceiling
and floor on it. The floorplan could be created either in Imagine or even
a paint program and then auto-traced.  This means you will have to bust up
a few polygons to add the windows and doors but that is a minor task if
you have a complex floorplan. I'm not sure how difficult it would be to add
the windows/doors in Imagine's modeler, but it would be a breeze in
LightWave. 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: How to precisely add objects (and all that)?
Date: Thu, 18 Jul 91 12:48:03 PDT
From: [e-mail] (R. L. Zarling)

>I'm making a house which has many rooms and thus many walls, what is the
>best way of adding walls to the house so that they leave no cracks to the
>corners ?

I have also struggled with this problem.  For me, it was even worse because
I kept changing the size of rooms which meant that all the adjoining rooms
needed to be adjusted again.

I finally decided that the best solution was to use a script-based drawing
model.  You enter coordinates as numbers, and thus get perfectly accurate
positioning.  The script files can be modified using an editor, which makes
the job of doing changes a lot easier.

TTDDD would be one approach that would allow you to do this.  Personally,
I used PageRender 3D, which has a real good scripting language.  The resulting
objects can be converted to Imagine format using interchange.

As a final alternative, you can of course enter coordinates directly in
imagine in the transform requester.  Only problem with this is that you'd
have to do it over again manually if you made changes.

--Ray
[e-mail]

##

Subject: Re: How to precisely add objects (and all that)?
Date: Thu, 18 Jul 91 14:15:23 PDT
From: [e-mail] (R. L. Zarling)

>> TTDDD would be one approach that would allow you to do this.  Personally,
>> I used PageRender 3D, which has a real good scripting language.  The resulting
>> objects can be converted to Imagine format using interchange.
>>
>
>What is interchange, is it PD, and if it is, where can one get it ?

Interchange is a commercial program by Syndesis (I think).  It transforms
3d objects in one format (like Sculpt, or Turbo, or PageRender) to another.
As new formats become available, they make new modules available to handle
them.  They also have modules to do some useful transformations like snap
to grid or remove duplicate points, and to simply report statistics on how
many points, polgons, etc. an object has.  I have found it very useful in
a number of ways.

I have heard rumors that a module is in the works for Lightwave, and an
updated one for Imagine.

--Ray
[e-mail]

##

Subject: FC24 Paint
Date: Thu, 18 Jul 91 20:20:47 EDT
From: [e-mail] (John J Humpal)

A friend of mine just bought a Firecracker and he loves it.  He'd like to
work with the beta version of the 24-bit paint program, but without any docs,
he's more or less lost.  Does anybody have any hints or tips I could pass
along on using this paint program?  Or would you recommend that he purchase
some other paint program?

Thanks in advance.

John J. Humpal -- [e-mail] -- short .sig, std. disclaimer

##

Subject: Revised Rooms Project
Date: Thu, 18 Jul 91 23:28:52 -0400
From: Udo K Schuermann <[e-mail]>

[Posted (with permission) for Alan Price]


    Alan Price's Improved Updated Revised ROOMS PROJECT proposal
    ------------------------------------------------------------

About two months ago, the idea for a communal project was proposed. This is an
open invitation to everyone on this list, as well as others who may not be, to
participate in that project.

This idea started when I posted an image on Hubcap.clemson.edu. The image was
a trace of a living room. By coincidence, another image had just been posted
by Udo Schuermann, called 'CastleRoom'. This prompted the idea that everyone
could make a room and post it and then we could announce a sort of 'open house
Imagine tour'!

Of course, even just a slideshow of everyone's rooms rendered as still pics
would be a pretty nice showcase for Imagine. But how about the idea of
everyone contributing their object files for their room to be connected
together with all the other rooms to create a kind of 'gallery' of rooms! Then
an animation could be rendered where the camera flies through all of the
rooms. This could be output to video tape and distributed to all kinds of
people and places as a showcase for both Imagine and the creativity and
ingenuity of its users!

"So," you ask, "Why hasn't anything happened if it first came up two months
ago?" Well, a lot did happen. There seemed to be a lot of excitement about the
idea. One person even said that this was the first time that they had ever
heard of such a thing happening on the networks. Being new to this, that
surprised me, but if we can pull it off, and it hasn't been done before, then
we'll be making history as well!

"Then why did it die off?" Well, that's more difficult to answer, but I can
say that I'm going to try awful hard to get it going again. The suggestions
for how we go about the process quickly turned to the idea of creating a fully
completed house into which we would place everyone's contributions. This would
first begin with the drafting of a floorplan, then people would claim rooms.
Then the subject turned to: "Who's going to make the floorplan?"

Well, I guess I don't have to tell you that floorplan was never made.

I thought about tackling the task, and I already had a list of rooms that some
people wanted to make to get me started. But I just couldn't do it. How can
you when you can't even begin to imagine what wild things people may come up
with to enter in the project?


So here's my new, improved, updated, revised ROOMS PROJECT proposal (IMHO):

I would like to reverse the process in that first everyone interested enough
to make a room (or rooms!) should go ahead and make them. Then after all the
rooms are contributed, a floorplan will be designed to best accommodate all
the entries. Walls may be altered slightly for the best fit and to maintain
the sense of the rooms as being part of a larger structure (i.e., house,
mansion, etc.) Every effort will be made to maintain the integrity of the
original rooms design.

Okay, so now to try to cover all the bases on how to go about this. Many
standards on scale and file management are going to have to be laid out.

Those standards follow in this order:

    1) Starting up again
    2) Rendering process (read 'Plea')
    3) Object ownership (a formality)
    4) Scale
    5) Ceilings & Walls
    6) Doors and Windows
    7) File paths
    8) Submitting the objects
    9) Submitting the Staging file
   10) Submitting ILBM files for brush wraps
   11) Deadlines

Save this to a file for future reference.



1) STARTING UP AGAIN

If you plan to make a room (or rooms) for this project. Please send e-mail
to the following address with 'ROOMS' as the subject.

                     [e-mail]

                OR   [e-mail]


You can write a brief description of the room(s) you plan to make or you can
save it for a surprise,if you wish. In any case, this is for taking an initial
head-count of people interested in the collaboration. If you have already sent
such a note, PLEASE do so again.

Tell other friends who aren't on the list about the project. If they are
interested in participating, include a note about that, too.

A good estimate for the cost of producing the tape would be to sell the tape
at about $10.00 each VHS tape. Please mention whether you plan to purchase the
tape at this price.

DON'T FORGET to post your ideas, suggestions, etc., about the group project to
the Imagine Mailing list.



2) RENDERING PROCESS (read 'PLEA')

As I said the first time, "Everyone could post their room objects and then
whoever has the guts could download them all and render a fly-through
animation."
Well, I've got some guts, but not enough that it wouldn't help if others could
throw in some of theirs.
I've got an A3000 and access to an A2500 with a Toaster and a BCD5000 frame
controller and a BVU-950 SP 3/4" VTR. (Oh yeah, and an A500 w/1 meg. Carefull!
Don't stir up that dust! Hey anyone wanna a good deal on a...:) Sounds good, I
know, but if we wind up getting even just ten rooms...  well, let's do some
modest calculatin':

IF we fly the camera through each room and want the 'trip' through each room
to last about, say, 7 seconds, and we wanted to go for 30 f.p.s., that would
be 210 frames per room and now let's just come up with an average trace time
of about 40 minutes per frame, (Oh yeah? What's yours? :) that's 8400 minutes,
or 140 hours, or 5.833 days. So mutiply that by 10 rooms.

SO you can certainly understand that if I get no offers of assistance in
rendering you can expect a posting at some point announcing that the animation
will be rendered in SCANLINE mode. (Other options will be considered before
having to resort to scanline, such as 15 f.p.s., resolve depth, flying a
little faster, etc.)

So if you have storage space and/or a fast machine and tend to go to sleep at
night, please consider joining in on this courageous venture! If you are
willing, send me a note saying so and we'll work out the details.

Oh yeah, 24bit is the plan.



3) OBJECT OWNERSHIP

This is just a simple statement concerning the room that you submit:

Any object, ILBM, or related files submitted for this project will remain
soley the property of the person submitting the files. These files will not be
redistributed in any way except in the form of the rendered animation on video
media. The only exception may be in the case of transferring required files to
a party participating in the rendering of the animation.  This second party
will adhere to the same ownership policy stated above.  After rendering is
complete, all copies of files will be deleted.

You will, ofcourse, receive full credit for your room in the animation.  We'll
scroll credits at the end of the animation. You can also have a contact
address included if you wish. (Actually, listing the location that each
participant is from would be really cool for elaborating on the fact that this
is a collaboration over the net!)


4) SCALE

This is the part that is initially most important. In order for everyone's
rooms to match up. The scale sizes must be the same. Ofcourse, we can re-size
rooms to fit, but that would be a pain and you already know the potential
problems with procedural textures, etc. So it would be best to start out with
the following rules:

                        40 UNITS = 1 FOOT

If your'e thinking, "Your'e going to expand larger than Imagine's default
world size pretty darn quick when you put all these rooms together." Your'e
right - partially. But if you consider that it would be pretty crazy (or at
least EXTREMELY memory intensive) to try to load all the rooms, objects, and
ILBM wrap files at once, you can see where it would be smart to only load two
or three rooms at a time. Since there are walls, you could never tell they
were missing. As the camera makes its way from one room to the next, the last
one is dumped and the next one coming up is loaded, depending on how far you
can see through the present doorways. (This is where closed doors that swing
open will come in very handy.)

CONSIDER the average size of a room to be 16 FEET SQUARE.  THIS IS BY NO MEANS
A REQUIRED DIMENSION, JUST A STARTING POINT.


5) CEILINGS & WALLS

The average ceiling is 8 feet high. Base your designs on this. If you want to
design a vaulted ceiling or some such variation, go right ahead, we'll find a
way to work it in. Just keep the eight feet in mind, even if your ceiling
rises above that height in middle areas, try to make it meet the walls at the
eight foot mark. This is unless, ofcourse, you are making an atrium or
something that requires TWO floors of height. A real room that has a ceiling
two stories high is more than 16 feet high depending upon the structure of the
floor in neighboring rooms. This does not matter in Imagine however, because
the floors will not collapse if you put too much furniture in the room. :) So
lets have a very modest floor thickness of 20 units (6").

                 FLOOR THICKNESS = 20 UNITS (6")

You don't have to worry about this as far as your actual room object is
concerned. Your floor and ceiling can be only one plane (0 thickness) when
submitted. When the rooms are all put together, one's that are on top of
another will just be raised 20 units more.

             WALL THICKNESS = 20 UNITS (6")

Yup, the same goes for the walls. Pretty uniform, huh?  Again, the inside
walls of your room need only be one plane thick, they'll just have 20 units of
space put between them and the neighboring rooms. BUT here's the trick:

The DOORWAYS show the thickness of the walls. SO you have to make a 'lip'
around your doorway. Decipher this illustration (View is from above):


        inside room     doorway
           __________              ___________
                     |            |            } 20 units
        outside room


This is easily done by selecting the four points that make up your doorway,
isolating them as an object and extruding them 20 units, then joining and
merging the object back to your rooms walls. Viola!

When the doorways of two rooms are butted together, the points of one or the
others 'lip' will deleted.


6) DOORS & WINDOWS

DOORWAYS ARE 26O UNITS HIGH (6'6")

DOORWAYS ARE 110 UNITS WIDE (33")


Sure, lots of doors are different heights and most exterior doors are 36" wide
while most interior doors are 32", but I went with 33" so that the UNITS
measurement would be a round number. However, it is very important that you
adhere to this dimenson. This more than any other.

Door objects should have their axis placed along the 'hinged' edge of the
door. This way it can be easily pivoted open by using 'Align'.

If you like, at your request, I will gladly send you a blank door that you
could embellish with doorknob, moulding, etc., and could even use to slice the
doorway into the wall.

NOTE - if your are a detail perfectionist and realize that there is a gap all
around the door between it and the doorframe, shrink the door slightly rather
than enlarge the doorframe.

Your room should have two doorways. (It would be preferable that the camera
did not have to back out of your room :) Doors are not necessary. If you do
make doors consider the doors as seen from the INSIDE of the room. I may wind
up "splicing" the inside of your door to the inside door of someone else's
room - thus a door becoming a true collaboration!  IF you want more than two
doors, OK, but only two will be used to pass through (Closets or whatever.)

Since no restrictions have been made on the actual size of rooms, it's a
difficult matter to specify where along the wall you should put the doors.
Start by placing the door somewhere along the wall using 1/4 as a rule.  That
is, either place the door smack in the middle of the wall or 1/4 of the length
from the left or right side. Sure this is very unlikely to match up with many
other rooms right away, but I may wind up having to drag doorways around a
bit. If this becomes a problem with the way you have the contents of the room
or things on the wall, I'll get in touch with you.

WINDOWS -

Stained glass, skylights, portals, ANYTHING GOES! The only thing that won't
work is windows on all four walls. Actually even two walls are a problem
unless your room happens to be a corner of the building.  Again, let your
creativity flow, and if we run into a problem later, we'll work it out
together. (Windows will not be blocked up thoughtlessly.)

If you want something else to appear outside the window besides a gradient
blue sky and green ground (the global setting to be applied) such as an ILBM,
wrap it to a plane placed just outside the window.

If you want a beam of sunlight casting shadows through your window(s) you can
specify placement and the lights attributes in a staging file. (See
'Submitting the Staging file' section 9)



7) FILE PATHS

IMPORTANT to save later frustration:

A universal system of file management has to set up for this to work smoothly.
Since I am suggesting that you submit a staging file along with your objects
and ILBMs, this sections consists of two file path specifications:

OBJECT FILES -

a) Any object which references an Imagine texture (i.e., wood, grid, etc.)
   MUST reference that texture in a directory called 'Im_Textures'. Example:

            Im_Textures/Checks

b) Any object which references a ILBM brush wrap ('brush') MUST reference
   that ILBM in a directory called 'WRAPS'. Example:

            WRAPS/MyPicture.ilbm


STAGING FILE -


a) Any object loaded into the staging file (as seen in the action editor)
   MUST use a directory called 'OBJECTS' as its path. Example:

            OBJECTS/Chair.obj


As I receive submissions, all objects can be copied to a general directory
called OBJECTS, and all ILBMS to WRAPS. Make sure that you save at these
directory levels only and that nothing preceeds them (i.e.-
DH0:Imagine/WRAPS/) I am NOT going to deal with reassigning paths or using
c:assign all the time. If your files start asking for unknown devices and
directories you'll hear from me. It's a pain, I know, I did it to someone else
once. :(




8) SUBMITTING THE OBJECTS

Shouldn't be any problems here. Be sure to follow the brush wrap path noted
above.

Try to save your objects with unique names that won't wind up accidentally
overwriting another, or someone else's overwriting yours. SO instead of
'Chair.obj', try naming it with your initials in it or something ('APchair')

It would be best if you save your objects at the coordinates that they are
supposed to be located in the room you have made. An easy way of doing this is
by using "Snapshot" in the Stage editor. Then when I load your objects, they
will be all placed in relation to one another, and I can just select them all
at once and move them to their new 'home'.

'Snapshot' should also be used if you had to resize an object in the stage
editor after loading it. 'Snapshot' will save it at the new size. Delete the
object from the stage editor and reload the newly 'snapshotted' object in
order to get rid of the old sizing info in the action editor. Having the
object the correct size rather than resizing info in the staging file will
greatly reduce potential problems.

Cycle objects are accepted. It would be real neat to have things bouncing
around or whatever when passing through the room. It would help clue me in
that you have a cycle object if you name it with an extension of '.cyc'

NOTE - All submissions should be your own original work. Objects obtained from
PD libraries or commercial sources will not be accepted.




9)SUBMITTING THE STAGING FILE

The reason I am suggesting that you submit a staging file is NOT to show the
placement of objects, they should already be saved in their appropriate
positions relative to one another as stated above. The reason instead is to
show what your desired light placement and light attributes are. Since this is
one of the most important elements in achieving an effective scene, I shall
leave this creative aspect up to you and will only copy your settings to the
new staging file that includes the other rooms.

NOTE - As there is the possibility that the animation may wind up being
       rendered in 24bit SCANLINE mode, think about using the conical and
       cylindrical lamp types when wanting to achieve 'lampshade' shadows
       and the like. (Hopefully this will not be the case.)

Another purpose for the staging file is for you to use it for a test rendering
of your own, that way when you submit it you should be confident everything
will go smoothly. When I first receive the files I will conduct the same test
rendering with the staging file in order to discover any problems before
icorporating the files into the finished staging.

Yet another purpose is an option that you may choose whether or not to take
advantage of- Create a motion path spline and save it to the OBJECTS directory
with a unique name (certainly not 'path' or 'campath', but maybe 'APcam.path')
This would not neccesarily be the exact path that will be used in the finished
staging file, but it can be used as a guideline to show the path you would
like the camera to follow as it traverses your room.

Other paths could be used if something ( a toy, for example) were to move
across the floor as the camera passes through.

If a cycle object is employed in the room, be sure and note how many cycles it
should repeat over the length of time it takes to pass through the room.  (In
the Actor requestor in the Action editor, ofcourse.) If you need a number,
base it on the 7 second example in section 2 above. The cycle number will be
altered with the appropriate ratio if the length of the fly-through changes.
(Simple example- if you want your cycle object to perform 4 cycles in the 7
second scene and the scene length is changed to 6 seconds then the cycles will
be changed to 3.43)

Cycle objects and motion paths would be much more preferable to key-framing
any moving objects - It will be much easier when copying your stuff over into
the final staging. Don't expect me to copy over 'Position' and 'Align' bars
that change every 5 frames over a span of 100 frames!

I will be employing key-framing for the swinging open of doors. You don't need
to mess with that.

Concerning the 'Highest Frame #' setting - Don't worry about it since the
animation is not being rendered directly from this staging file. In most cases
you can leave it '1', unless an object morphs or something and you want to
represent that in the action editor screen. If something seems like a special
case situation, include a text file with your submission that further explains
what you have in mind.




10) SUBMITTING THE ILBM FILES

All ILBM files will be copied into a directory called 'WRAPS', so make sure
that's the path that your objects want to see when loading.

This type of file is the biggest memory hog of them all, both in disk storage
and when loading in for rendering. If you insist on submitting HiRes 24-bit
ILBMs for brush wraps, then maybe you ought to be one of the people helping to
render all this because you obviously have the hardware and/or memory for it.
Otherwise try to keep the ILBMs as economical as possible. If it's just the
pattern on a vase in the far corner of the room, you might get away with a 8
color lores ILBM.

A limit on the number of ILBMs for each room should be set, too. Let's make it
four for now and you can argue later. Try making use of the fact that the same
ILBM can be used for more than one object and only has to be loaded into
memory once. Try to make more use of the procedural textures, as they require
much less memory and don't need to be uuencoded and sent over e-mail.

NOTE - all ILBM files should be your own original work. ILBMs obtained from
       PD libraries or commercial sources will not be accepted.


11) DEADLINES

A good idea, no? A tentative deadline for the submission of the rooms would be
the end of September. The deadline will be modified according to the amount of
activity related to the project.




SEND SPECIFIC QUESTIONS TO: [e-mail]
            OR: [e-mail]


SEND GENERAL QUESTIONS, OBSERVATIONS, IDEAS, SUGGESTIONS, ETC., TO:

                [e-mail]



HOPE TO SEE YOUR ROOM SOON!


AP.



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

Endnote from Udo Schuermann ([e-mail]):

Alan and I have discussed some of the issues that went into this revised
proposal.  The vast majority of the work, however, is his and his alone.  Due
to his absence for the next few weeks I am posting this for him and will offer
myself and my mailbox for questions.

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

##

Subject: Using Anim5 - help
Date: Fri, 19 Jul 91 09:24:10 JST
From: [e-mail] (Manjit Bedi)

Hello all,
excuse me if this has beeen asked before.  I have been off of the list for
a few months.
Over the last few days, I had been fiddling with a fairly simple
animation.  I first rendered the animation using the Imagine format
which seemed to work fine.  Then I rendered it using Anim5.  The
animation is in HAM lor-res interlace.  When I play the animation back
in Amigavision it seems ok at first but then as one the moving objects
is flying around it leaves some white crap behind and parts of the screen
start to flash.  Is this what people refer to as HAM fringing?  Is there
a cure?

Manjit Bedi ( [e-mail])

Recommended listening : Electronic 'Electronic';
            Morrissey  'Kill Uncle'.
##

Subject: New Rooms project...
Date: Fri, 19 Jul 91 1:49:35 PDT
From: Daryl T. Bartley <[e-mail]>

 Okay, silly question, but is there any _minimum_ for rooms? Also, would it
 be okay to upload a room without testing it? I might be able to get it to a
 friend's system to test it out, but I can probably put a room together I can't
 render at all. Also, if there are any brushmaps, a test render for me is right
 out. I was just wondering, as I would _like_ to participate, but I don't know
 if a room of the content that I could handle on my weenie system would be
 worth it to include.

 Oh, and assuming that I CAN get into the project, how do you take the device
 specs off the objects? Just go in 'by hand' and change the paths in the
 requesters?


Daryl Bartley
[e-mail]

##

Subject: help me.. help me.. I be hypnotized!
Date: 19 Jul 91 09:11:00 EST
From: "SYSTEM MANAGER" <[e-mail]>

Greetings folks!

Thanks to Steve's wonderful idea and Imagine's trace mode I was able to
creat the basic floor plan of my house using Deluxe Paint III and
Imagine.  It worked wonderfully!

Here is the question though.  After I extruded the floorplan, and saved
it I went into the stage editor.   I loaded the object, set my perspective
window to be the camera view and then created a new AXIS to aim my
camera at.

I then placed the camera inside the building and moved my AXIS in there
as well.  I looked at the perspective window and discovered that my
camera magnification must be too high, and so not knowing how to
reduce the magnification, I scaled the building up.   It helped, but
did not fix it to my satisfaction.  So how do I change the focal
length of my camera?  Am I using the correct term?


-mark=

---------------------------------------------------------------------------
Mark D. Manes                       //           CIS  :  74030,744
System Manager/Programmer          //            Email:  [e-mail]
Norfolk State University       \\\// Amiga       Phone:  [phone]
---------------------------------------------------------------------------
"AmigaDOS 2.0 is the future of the Amiga -- be a part of it!"

##

Subject: RE:  Help me.. Help me.. I am hypnotized..
Date: 19 Jul 91 10:23:00 EST
From: "SYSTEM MANAGER" <[e-mail]>

>To change the focal-length, try the size gadget in the action requester.
>(for the camera that is.)  8)

I tried that.  I guess I really don't understand the size gadget.
Can you explain it a bit better ?

>-- Steven Webb

-mark=

---------------------------------------------------------------------------
Mark D. Manes                       //           CIS  :  74030,744
System Manager/Programmer          //            Email:  [e-mail]
Norfolk State University       \\\// Amiga       Phone:  [phone]
---------------------------------------------------------------------------
"AmigaDOS 2.0 is the future of the Amiga -- be a part of it!"

##

Subject: Rooms Project: Clarification
Date: Fri, 19 Jul 91 11:04:56 -0400
From: Udo K Schuermann <[e-mail]>

I send this to the list just in case that there exists a more
widespread misunderstanding due to some lack of clarity in the
proposal.  There are a few things to which you _must_ adhere,
otherwise you will create us much grief!

Daryl T. Bartley <[e-mail]> asks:
> would it be okay to upload a room without testing it?

Negative.  No.  What you submit to Alan (or me) should have been
tested, i.e. rendered to your satisfaction.  We will assemble what you
send us, and perhaps make minor adjustments (only if we absolutely
have to).  Please don't assume that we will do your work.  We have
enough to do assembling the stage and rendering the animation!

I suggest you use Scanline Rendering both to prevent surprises if the
final project isn't raytraced, and to speed things up on your end.

> I might be able to get it to a friend's system to test it out, but I
> can probably put a room together I can't render at all. Also, if there
> are any brushmaps, a test render for me is right out.

A few quick tips for saving memory:

1. Do NOT use the [MakeAnim] button to create the rendering, as this
   will open (by default) 3 screens.  Total waste of RAM if all you
   want is a still picture to test your room.  Use [Generate] instead.

2. Use the algorithmic textures instead of brushmaps.  We ask that
   you limit yourself to 4 brushmaps for now.  We don't really care
   how many textures you use, as these are far less costly.

3. If you're absolutely desperate for RAM, there are programs out
   there which reduce your workbench screen from 2 bitplanes to 1
   bitplane.  Looks ugly, but it just might help make the difference.

> how do you take the device specs off the objects? Just go in 'by hand'
> and change the paths in the requesters?

NO!  We will not alter the paths which your objects specify.  The
proposal detailed this in section 7.  Please read it again!  Synopsis:
Specify relative paths (not absolute ones), meaning that all your
object filenames look like:
    OBJECTS/WALRUSmyTub.iob
    OBJECTS/WALRUSmyPlug.iob
    OBJECTS/WALRUSmyWater.iob
    OBJECTS/WALRUSmyDrain.iob
all your ILBM brushmaps look like:
    WRAPS/WALRUSmyStainedGlass.transparency
    WRAPS/WALRUSmyBathroomJoke.color
etc.

Do *NOT* specify them like
    Imagine:OBJECTS/WALRUSmyTub.iob
or
    DH0:Imagine/WRAPS/StainedGlass.transparency

Ok?  All you need to do is
    a) go to the directory from where you load Imagine.
    b) makedir OBJECTS
       makedir WRAPS
    c) store and retrieve all your objects from there
       without specifying anything more of the path
       than absolutely necessary.  Upper/lower case
       doesn't matter, of course.

Btw:  in my examples above, I named all brushes and objects such that
they were preceeded by a (hopefully) unique ID (walrus).  Do something
similar.  Perhaps use your initials.  Whatever.
    The proposal as posted to the list yesterday is authoritative.

I hope this clears up any sort of misunderstandings.

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

##

Subject: The 1000th post
Date: Fri, 19 Jul 91 18:16:24 EDT
From: [e-mail]

The 1000th post was by Udo K Schuermann ([e-mail]). True, he
was talking about Alan Price's project, but it was his post.
Congratulations, Udo!

The prize is a set of floppies containing lots of fun Imagine files, mostly
copies of the files on hubcap, but with a bunch of new toys that I haven't
uploaded yet. These are copies of disks I'm sending to our friend
Rick Rodriguez.

As a bonus, unannounced prize,  copies of these disks also go to
Sean Cunningham ([e-mail]) and Rick Tillery
(drtiller@something) just because a few of their posts were especially
well thought out and added a lot to the discussions on the list.

A very honorable mention goes to Mark Thompson, who also regularly
contributes a great deal to this discussion. No prize, though, because
he doesn't own Imagine! :-) [Sorry, Mark!]

I thought it would be interesting to see the major contributors by
volume. In the past 150 posts (the number I keep in my personal on-line
archive) the most prolific posters were:

Steve Worley    26
Mark Thompson   22
Sean Cunningham 10
Rick Tillery     8
Udo Schuermann   5
Juan Trevino     5
Steven Lee Webb  5

The rest of the posters has less than 5 posts. The top 7 posters
combined made up about half of the total posts.

Udo, Sean, and Rick, send me your US Mail addresses. I promise to send
your disks within 5 weeks. :-)

-Steve

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

##

Subject: New object stagnation
Date: Fri, 19 Jul 91 18:28:22 EDT
From: [e-mail]

I've noticed that nobody likes putting anything on hubcap, especially
objects and projects. I'd really like to encourage people to upload
their work: it gives the rest of us new toys to play with or scenes
to inspire new ideas. I'd especially like to get a HUGE object
library. There are about 100 objects on hubcap right now, but I'd like
everyone to upload their best couple of objects just for something new
to play with. Please?

To egg people on, I have uploaded a 115K imagine object of a hypersonic
space transport. It has a great, rounded, slick feel, unlike the X29
which is incredibly overused. :-)

I am working on a couple very complex models: one is a space station,
and one is a jet engine. [Imagine putting this on the back of the
wimpy VW bug model!]

-Steve

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

##

Subject: Camera Focal Length
Date: Fri, 19 Jul 91 14:55:17 -0400
From: Udo K Schuermann <[e-mail]>

The ration of the camera's X/Y size determines the focal length.  I
don't know which ratio widens the angle, off the top of my head, so
play with it.  I don't think the Z size does anything.

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

##

Subject: FAQ
Date: Fri, 19 Jul 91 18:33:44 EDT
From: [e-mail]

I forgot to say that hubcap is an anonymous FTP site. Its
IP number is 130.127.8.1, for those without a host name database.
The Imagine goodies live in /pub/amiga/incoming/IMAGINE/*.

If you don't know what FTP is or how to use it, ask one of your
local computer experts: they should be able to help you get
access to the GIGS of free software and information out in netland.


-------

We should make a FAQ list to post monthly with the intro to the list
message. Any suggestions on questions to include?

1) What is FTP, where is hubcap?

2) Why can't I make transparent glass?

3) My object renders fine in scanline, it gets chopped off in trace!

4) When is 2.0 coming out?  :-)

We can hash out  questions and the "best" answer on the list. Thats what
its here for!

-Steve

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

##

Subject: Re: attribute assignment idea
Date: Sat, 20 Jul 91 12:34:34 PDT
From: [e-mail]

Steve Worley writes:

> Talking about hi-end features, the author of 3D Pro is bragging that
> 3D Pro 2.0 will have radiosity. Yow! Radiosity is a very complex method
> of lighting scenes that is particularly realistic; areas like room corners
> and alcoves are more dimly lit than the middle o> f the room. It is a VERY
> subtle feature that makes startlingly lifelike renderings. The trick is
> that it is so expensive (hard to code well, and a LOT of extra compuation in
> the render) that an Amiga (or Apple or NeXT or IBM) implementation that
> would actually be useful seems nearly impossible. Of course, what do I
> know?  One other feature of radiosity is that if you have a scene with
> the radiosity lighting computed, you can render different views without
> re-computing radisity. This makes room fly-throughs very popular..

That's not all radiosity is good for. It is now widely being used in
Virtual Reality worlds for interiors. This way they save quite a bit
of the real time rendering expense. The rooms are rendered once with
radiosity and are then available for real time use anytime in the future.
I spent quite a bit of time playing in VPL's worlds (I have had some
private demos) and their three newest worlds were all done with radiosity.
One was a floor of an office building, one use a subway station the other
was a surreal room with a giant teapot you could fly into with lots of
strange image maps (yes, besides radiosity they can now use worlds with
image maps in real time also). This is quite an improvement. It is
quite a leap in the reality factor to be able to have worlds with
shadows and images to play in.

It should be noted here that radiosity is not all THAT wonderful.
It certainly does not even come close in reality factor/accuracy
to full raytracing and even scanline tracing with a good renderer.
There are unrealistic factors, shadows are not always there or
correct and as you move around things look incredibly static. However
for flythroughs, as Steve mentions above and VR radiosity can be
quite nice.

=======================================================================
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: Imagine to Lightwave
Date: Sat, 20 Jul 91 12:33:22 PDT
From: [e-mail]

  Randy Carris writes

> I'm looking for a way to cpnvert Imagine objects to Lightwave objects.  I
> know that with Interchange you can convert Turbo to Sculpt and load them that
> way, but do you loose anything.  I obviously don't want the shape distorted
> etc.  Can you keep the attributes or do they get lost or screwed up?  Is
> there any dirrect way to do these conversions?  I like the object creation
> abilities I get in Imagine (when I ever get better control over them), but
> I con output to tape a lot easier with our Toaster.

Two things to mention here:

1) I spoke with Alan Hastings (the author of Lightwave) and his intent
     for version 2.0 of the Toaster software (which is very close now)
     will almost certainly be able to DIRECTLY import Imagine/Turbo created
     objects. He is well aware the the modeler for the Toaster is ... well
     ... er... lacking. He isn't responsible for writing that program and
     has not been able to convince the programmer to revamp the program.

2) I have been able to successfully render animations in Imagine (the last
     one was about 30 seconds) and use the Toaster to single frame them
     to videotape. I even did one where I sequentially grabbed (24 bit)
     frames from video using the Toaster (programmed it myself) and then
     used the images as an animated brushmap in Imagine.
     If people are interested in the procedure, and the ups and downs/pros
     and cons I ran into let me know and I will post it.

> P.S. ,
> I got my plane tickets to Las Vegas for July 28-Aug. 2  I'm a student
> volunteer at SIGGRAPH this year, who all is going to be there.  Anyone else
> on here going as a volunteer?

I'm not going as a volunteer, but I will be there from Sunday through
Saturday. Maybe some of us should get together. By the way, Newtek
is throwing an AWESOME party. They have a sincere desire to blow Wavefront,
Iris and others out of the water. And they would like anyone who has done
any Toaster animation work to bring a tape.

See you at Siggraph.

=======================================================================
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: Rooms Project: Clarification
Date: Sat, 20 Jul 91 13:54:21 PDT
From: Daryl T. Bartley <[e-mail]>

Okay, then I guess I can't participate at all. Any 'room' I could put together
and actually RENDER would be so barren as to look like a mistake.

Oh well. Maybe if there is ever a 'Rooms II' in the far distant future, I might
be able to join in. I might still want a copy of the tape though.


Daryl Bartley
[e-mail]

##

Subject: Re: How to precisely add objects (and all that)?
Date: Sat, 20 Jul 91 14:00:15 PDT
From: [e-mail]

Juha Kallioinen <[e-mail]> writes:

> I'm making a house which has many rooms and thus many walls, what is the
> best way of adding walls to the house so that they leave no cracks to the
> corners ?

> Now I always have to zoom as much I can to move a new wall precisely to a
> corner. But it is pretty tedious to zoom around many many times and then
> back again.

> Stupid question probably, I'm sure there's a simple solution to this, but
> I can't figure it out, so heeelp.

Two fairly simple solutions.

1) Use "Snap to Grid" or "Lock" setting up your grid appropriately for
    the size of your walls.

2) For pure accuracy, don't use "move", "scale" or "rotate", I almost
    never do now. Add a wall, use the Transformation requestor to
    place your walls exactly where you want to.

=======================================================================
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: FAQ
Date: Sun, 21 Jul 1991 05:59:03 GMT
From: [e-mail] (Stephen Menzies)

Donald Richard Tillery Jr <[e-mail]> writes:

>How about "Why do I have to re-boot after showing a picture?"

>My U.S. mailing address is 314 E. Boyd #7 Norman, OK 73069 until August 15
>and after that I can be reached through 5000 Ryan Dr. OKC, OK 73135 (but I'll
>really be in San Diego, but I don't have a place to live yet)

>Thanks for the positive feedback on my posts.  I didn't think they were all
>THAT helpful, but it's good to be appreciated :-)

>BTW, did my post on ColorBurst get out there?  I haven't had any responses. Or
>is everyone waiting for my review to appear?

It got to me , and I look forward to your review. I did hear some
negative things said about it's output quality and would appreciate
any comments you have about that (among other things). Glad to hear
you got it Rick, and I'll look into getting a pic or two to you
as promised. Got some *really* clean and detailed Caligari Broadcast
pics now, too.

>Rick Tillery ([e-mail])

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

##

Subject: wood grain on brush
Date: Mon, 22 Jul 91 1:25:17 EET DST
From: Juha Kallioinen <[e-mail]>

 I'm making a painting with a wooden frame to be hang on a wall.
There popped up a problem with the brush and the wood texture.
If I put the brush inside the frame like the awful picture down there
shows, I get the wood grain rendered on the brush, and that is not what
I meant. What should I do with my texture parameters ?

Yes, I can remove the back of the frame, so that there would not be
anything to draw the grains to, but can this be solved by altering the
texture-stuff ?

               ---.
                | |<----- Frame
                | |
                | |
                | |
                |<+------ Brush
                | |
                | |
                | |
               ---'

Did anyone get that :-?

- Juha

--
o-----------------------o----o---------------------------------------------o
| [e-mail]/  __ |          Old programmers never die.         |
| Juha Kallioinen     /_ /// |      They just branch to a new address.     |
o--------------------o \XX/  o---------------------------------------------o

##

Subject: AmiExpo and Extruding
Date: Sun, 21 Jul 91 18:53:20 -0400
From: [e-mail] (ESTES,JON-PAUL)

This is my first post, so howdy.  I have a couple of questions.
The first has to do with a feature that, for some reason, I
simply cannot get to work.  I am trying to extrude an object
along a path.  Nothing I do seems to work.  I am using a path
created in the Detail Editor, both the path and the object are on
the screen.  No matter what I do, nothing works.  Help!

The second question has to do with AmiExpo in Orlando.  Anyone
going?  Anything I should know about (seeing as how I have
never been to an AmiExpo)?

Well, for now that is it.  Happy trails, pardnerDD

##

Subject: An Idea?
Date: Thu, 18 Jul 91 09:19:30 EDT
From: [e-mail] (Bob Lindabury - SysAdm)

Daryl T. Bartley <rutgers!ecst.csuchico.edu!dmon> writes:

> I was just thinking of more stuff I could ray-trace (On a regular system, no
> less!), and came up with an idea, sort of. There are lots of people on the li
> I have been toying around with logo ideas (like a small object I could hide a
> the bottom of my pics), and was just curious.

That sounds like a good idea.  Would be interesting to see what
people come up with.  Whatever happened to the house of rooms?  Did
that die out or is it all over already?

> Also, have any of my postings to the list been getting through? I know the er

Well, I guess they have been getting through as I received this one!
<grin>

> Hmm, this message prolly doesn't qualify towards #1000, does it?

Sure, why wouldn't it?  It has some useful information in it and
isn't just fluff as far as I can tell. B^)
>
> Daryl Bartley
> [e-mail]

-- 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: Morphing
Date: Thu, 18 Jul 91 09:16:05 EDT
From: [e-mail] (Bob Lindabury - SysAdm)

rutgers!pro-party.cts.com!seanc (Sean Cunningham) writes:

> I've just about determined that I'll have to continue work in another
> package to progress much further.  It's not that any other package does
> better morphing (they all do point morphing), it's that the smoothing
> sometimes turns off on faces around the "lip" of my puddle, giving a
> faceted look.  Also, I need to be able to bend points, where the DETAIL
> editor will only let me drag points.

How do you bend points?  As far as I know you can only bend edges and
not points.  Wouldn't dragging points or using magnetizm do what you
are looking for?

> An interim solution may be to do all of my distortion modelling in
> Modeler3D and then import the object into Imagine via Interchange.

And Modeler3D will bend points?  Strange.

> Sean

-- 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: Imagine to Lightwave
Date: Thu, 18 Jul 91 01:10:49 EDT
From: [e-mail] (Bob Lindabury - SysAdm)

rutgers!shumun.weeg.uiowa.edu!rcarris (Randy Carris) writes:

> I'm looking for a way to cpnvert Imagine objects to Lightwave objects.  I kno
> Interchange you can convert Turbo to Sculpt and load them that way, but do yo
> anything.  I obviously don't want the shape distorted etc.  Can you keep the
> do they get lost or screwed up?  Is there any dirrect way to do these convers
> the object creation abilities I get in Imagine (when I ever get better contro
> I con output to tape a lot easier with our Toaster.

Interchange will convert your Turbo Silver or Imagine objects over to
Lightwave compatible objects (Videoscape) quite nicely.  That's what
I use and I've never had any problems with it.  I wouldn't suggest
going to the Sculpt format however.  I haven't tried that format
conversion for Lightwave but since Allan Hastings (the author of
Lightwave) also wrote Videoscape 3D, I would imagine [sic] that the
videoscape objects would be the most compatible.

All my objects maintain thier color (more or less) and never get
distorted.

> Thanks,
> Randy Carris
> [e-mail]

-- 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: wood grain on brush
Date: Mon, 22 Jul 91 00:44:28 -0400
From: Udo K Schuermann <[e-mail]>

Juha Kallioinen <[e-mail]> writes:

> I'm making a painting with a wooden frame to be hang on a wall.
> There popped up a problem with the brush and the wood texture.
> If I put the brush inside the frame like the awful picture down there
> shows, I get the wood grain rendered on the brush, and that is not what
> I meant. What should I do with my texture parameters ?

The solution is to separate the frame and the backing upon which you
map the brush (front view):
     _________
    | _______ |<--frame    (object 1)
    ||       |<---backpane    (object 2)
    ||       ||
    ||       ||
    ||_______||
    |_________|

The Backpane gets the IFF brush mapped onto it; the frame gets the
wood texture.  Group the two objects together and save them as a
group.  Voila!

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

##

Subject: Colorburst and wood textures
Date: Mon, 22 Jul 91 01:03:06 PDT
From: [e-mail] (Aaron Tucker)

        Well, I am no longer affiliated with M.A.S.T., the makers of the
Colorburst. I will, however, continue to address any questions concerning
the video device as I probably know it as good as or better than anyone else
in the US.
        The meaning of 48bits is misleading. 48bits does not mean 2.8E14
colors. It is still 24bits. 48bits is similar to 32bits. 32bits is what the
TARGA advertizes. All it is is 24bits with an 8bit alpha channel. The
Colorburst is capable of 48bitplanes. That is 24bits with a 24bit alpha
channel, so MAST says. IMHO, that is not true. The other 24bits is an overlay
similar to an alpha channel, but it cannot be separated. You can, however,
acheive a 28bitplane effect by seperating the Amiga graphics from the
Colorburst graphics. You create a 4bitplane hires screen and use the info
from the composite port (sorry A3000 owners). The 48bitplanes allows you to
fade from one 24bit picture into another completely smooth, something most
board are not capable of, including the ones due to be out soon.
        As for the post about bad image quality...whoever you heard that
from either A) Needs glasses B) works or is influenced by the competition
C) used a normal 1 to 5 bitplane IFF D) Didn't have thier monitor on E) Has
no idea what they are talking about, or F) Any combo of the above.

        The picture quality is nothing short of spectacular. 24bit RGB is
wonderful. I love my IMAGINE pictures...sorry, OUR IMAGINE pictures. The
Colorburst is equal in picture quality with the Firecracker.

        Anyways, I will be uploading some 24bit brushes of wood textures
in the somewhat near future. I downloaded some HAM framegrabs of some wood
and will upload them when I get them converted. There are some really good
ones. Lots of different kinds of wood.

        See ya guys later.

Juan Trevino
[e-mail]

##

Subject: Config editor
Date: Mon, 22 Jul 91 01:24:53 PDT
From: [e-mail] (Aaron Tucker)

        Is anyone experiencing problems running the Imagine Config File
Editor from the Workbench? I can run it fine from the SHELL, but it gives
me the ols guru when I 2xclick it from the WB.

System: A500 (68000 phewy)
        4.5MB RAM
        290MB Hard from old Supra controller
        WB 1.3.2 with ARP installed

BTW, running it from WB not only causes a GURU for me, it also devalidates
my 180MB drive. It takes FOREVER!! to validate upon bootup.
Any questions, comments, help, or celery?


Juan Trevino
[e-mail]

##

Subject: C Hackers
Date: Mon, 22 Jul 91 16:32:54 EDT
From: [e-mail]

Glenn Lewis and I have been working with a lot of ideas in writing C
code to perform VERY complex object manipulation. Our plan is to write
a library of object manipulation functions to read and write object
descriptions (using TTDDD) and then perform manipulations of these
objects in software. Already the read and write routines have been
completed, and now we are working on manipulative ideas.

One is a dynamic motion simulator: It allows points and edges to
respond to gravity and each other, allowing you to collapse objects or
melt them slowly, or make them jiggle like they are made of Jell-O.
[An F-16 wiggling like this is lots of fun!] We also plan to make a
routine to output Rayshade object files. [Rayshade is a PD renderer,
mostly for UNIX boxes].

We also are designing algorithms to morph objects into one another.
ANY objects. Smoothly. Even if they have different #s of points.
Fun, eh?

Other ideas involve an alternate input file format that is easier to
use, building an ILBM outline tracer with autofacing, creating
exploding objects that explode MUCH more realistically than the
built-in triangle vomit effect, making objects collide and interfere
with one another, and many other ideas we've thought of that just
cannot be done without low level coding.

We decided to post to the list in case there are any C hackers who
want to contribute... we are just starting our library of functions,
and other people's input (as well as coding!) would be really useful.
This invite really doesn't apply to those who don't hack C code: this
is a pretty low-level (though incredibly powerful!) approach to
animation.

If you are interested, E-mail me ([e-mail]) with a
CC: to Glenn ([e-mail]). A special invite goes to
Rick Tillery and Mark Thompson: I know both of you are code hackers.

-Steve

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

##

Subject: Re: AmiExpo and Extruding
Date: Mon, 22 Jul 91 17:16:24 EDT
From: [e-mail] (John J Humpal)

> I am trying to extrude an object
> along a path.  Nothing I do seems to work.  I am using a path
> created in the Detail Editor, both the path and the object are on
> the screen.  No matter what I do, nothing works.  Help!

This may be a stupid question, but does your path have any points?  It should
have at least 2 points.  Also, extrusions always take place in the extruded
objects y-dimension.  I'm not sure what could be causing your problem. Path
extrusion has always worked for me.

John J. Humpal -- [e-mail] -- short .sig, std. disclaimer

##

Subject: Re: Radiosity
Date: Mon, 22 Jul 91 19:29:53 CDT
From: Wayne Haufler 283-4160 <[e-mail]>

Hello all,

Since there has been some discussion on radiosity, lately, I thought I'd
type and share with you the contents of a handout on radiosity that I
received at the Feb. 20, '91 meeting of the Houston Area ACM/SIGGRAPH
meeting.  The presenter was S. Mike Goza of NASA at JSC, and his speech was
titled "Radiosity Algorithms for Animation".  He also showed some very
impressive animations of a ray-traced robot servicing/repairing some
electronics in a 'radiositied' space station interior.

But since this is nearly 100 lines long, please E-Mail requests for this
presentation.  It is rather involved, and I do not pretend to really
understand it all.  I will try to get permission from him before
distributing this.

===============================================================================
          ____  Wayne A. Haufler     ([e-mail])
\\   /\\   /\\   //    McDonnell Douglas Space Systems Company - Houston Div.
 \\ /--\\ // \\ //--        [Christian/SW Engineer/Amigan]
  \//   \//   \//___   "Exploring the Use of Computer Graphics and Animations"
        //                   "To Serve and Support Christian Endeavors"
       //
            Standard Disclaimers Apply.
===============================================================================

##

Subject: Re: AmiExpo and Extruding
Date: Mon, 22 Jul 91 22:12:53 PDT
From: [e-mail]

> This is my first post, so howdy.  I have a couple of questions.
> The first has to do with a feature that, for some reason, I
> simply cannot get to work.  I am trying to extrude an object
> along a path.  Nothing I do seems to work.  I am using a path
> created in the Detail Editor, both the path and the object are on
> the screen.  No matter what I do, nothing works.  Help!

Here's what you do:

You do NOT use "create path". Create your object outline to be
extruded. Add a new axis and make it correspond more or less with
the axis of your object. "Add lines" to the new axis from the y axis
in the shape of the path you want your object to extrude along.
Name that new axis "path", then extrude your object along the path.

Good Luck.

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

##

Subject: Imagine
Date: Mon, 22 Jul 91 23:09:31 CDT
From: mit-eddie!harvard!scubed!pro-party.cts.com!cbownds (Clayton Bownds)

   Hello, I don't know if you are the correct person to ask, but I would
like to be part of the Imagine mailing list and any information you could
give me would be greatly appreciated.
   My partner (Sean Cunningham) and I have been working on some fairly
complex animation and have run into some snags...
So we have talked back and forth quite a bit about what things we would
like
improved, and came up with a crude list. I know we have not corresponded
via E-mail or the net before, but Sean gave me your address and I wanted
to see if you had any input on our list. ...so here goes...



          Motion Blur
          Depth of Field
          Particle Systems (with tradjectories and life cycles)
          Displacement Mapping
          Spline support as well as polygon
          True object morphing
          Transformation and Molding on the point level
          Soft Shadows
          Ability to manipulate scene in perspective
          Ability to run in a window (anticipating mega-hires displays that
                                       are coming real soon now)
          Radiosity
          Scene Gravity and Mass for objects
          Free form object paths
          .RIB support
          Light and Environmental effects (Fog, Fire, Smoke, Clouds, etc.)
          Free form deformation
          Built in support for outline and postscript fonts (for converting)
          Network Support (distributed processing)
          True "object" lighting (not just the axis)
          Rubberbands



Granted some of these things are a few updates away, but seriously, what
are a few things you would change or add to Imagine, and do you agree with
some
off our list (or disagree).

And last but most certainly not the least...
Have you heard *ANYTHING* about an upgrade from v1.1 to v1.x? I have
noticed that Lightwave hasn't stopped improving since it's initial
release, nor Animation:JOURNEYMAN, so what should make Imagine any
different? I don't think it's very fair to the users for Impulse to
stop busting their tail to improve their platform (in todays its a matter
or staying competetive)


Clayton Bownds
Vision Graphics
----
ProLine:  cbownds@pro-party
Internet: [e-mail]
UUCP:     crash!pro-party!cbownds
ARPA:     crash!pro-party![e-mail]

##

Subject: Bending points?
Date: Tue, 23 Jul 91 12:49:17 EDT
From: [e-mail]

Sean Cunningham complains that you cannot rotate or scale points.


>When you select a group of points in the DETAIL editor, all you can do is
>drag them...it doesn't let you ROTATE or SCALE that clump of selected
>points.
>
>What I've run into is this: I stretch up the center of a puddle with the
>magnetism feature and I can move that "lump" around, but that's it.  I can
>pull up other "lumps" but what good are they for except for a bit of added
>detail to the surface...I've just got a bunch of straight up and down, or
>slightly curvy lumps on a surface.
>
>With Modeler3D, I could take this surface, select the lump at about half
>way and do a rotate: Wah-lah, instant bend.  I can do this a couple more
>times, varying the degree of rotation to get a completely new shape.
>
>This would be most tedious, if not impossible, to do in the Detail editor.
>Add this with the touchiness of Imagine's PHONG shading (several times I've
>had triangular facets show up on the edges of some of my objects) and
>you've got some major headaches to deal with.
>
>This is something that needs to be addressed in a later release of Imagine.


You'll be happy to know, Sean, that you are mistaken. Imagine will
let you move selected points, as well as rotate ans scale them. The
limitation is that you cannot do this interactively by using the mouse:
the Tranform command does the manipulation.

Quote from my Detail Tutorial:

>Picked points can be manipulated with the Transform command. The
>picked points can be translated, scaled, rotated, and positioned
>INDEPENDENTLY of the rest of the object. Rotations and scalings all
>use the object's axis as a reference point. Absolute positioning will
>move the FIRST point you pick to the location you choose, and the
>rest of the picked points will be translated an equal amount. Interactive
>dragging is accomplished using the "drag points" mode.


I've used this method to make paper airplanes out of flat planes. Lotsa
fun!

-Steve

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

PS- To inprove Imagine's Phong shading, make sure to MERGE your object
to make sure that all of the parts of your object are joined.

##

Subject: Re: Morphing
Date: Tue, 23 Jul 91 17:14:15 EDT
From: [e-mail] (Bob Lindabury - SysAdm)

rutgers!pro-party.cts.com!seanc (Sean Cunningham) writes:

> Well, maybe I should have worded "bending points" alittle differently.
>
> When you select a group of points in the DETAIL editor, all you can do is
> drag them...it doesn't let you ROTATE or SCALE that clump of selected
> points.
>
> What I've run into is this: I stretch up the center of a puddle with the
> magnetism feature and I can move that "lump" around, but that's it.  I can
> pull up other "lumps" but what good are they for except for a bit of added
> detail to the surface...I've just got a bunch of straight up and down, or
> slightly curvy lumps on a surface.
>
> With Modeler3D, I could take this surface, select the lump at about half
> way and do a rotate: Wah-lah, instant bend.  I can do this a couple more
> times, varying the degree of rotation to get a completely new shape.

This is more like it.  I fully understand what you are talking about.

I would love to do the group movements (scale, rotate, move) on
selected points.  It would make modeling certain objects much easier.

-- 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: Colorburst
Date: Wed, 24 Jul 91 00:19:23 PDT
From: [e-mail] (Aaron Tucker)

No, you do not need a special monitor. Any monitor that plugs into the
standard 23pin RGB connector will work with the Colorburst. The Colorburst
works with any Amiga (Some old A1000s have problems). The PAL Colorburst
supports PAL monitors. The NTSC Coloburst supports NTSC monitors. Actually,
it doesn't matter. The monitor will sync to the signal. It does matter that
you have the correct machine for the Colorburst you buy. The price varies
from place to place. In the U.S., MAST retails it at $599. That was the price
when I left. In other countries, the price is much higher.

Juan Trevino
[e-mail]

P.S. I am going to upload the wood maps to hubcap as HAM, not 24bit. Anybody
who needs to convert themfor other purposes can do so.

##

Subject: Re: Bending points?
Date: Tue, 23 Jul 91 23:16:46 CDT
From: mit-eddie!harvard!scubed!pro-party.cts.com!seanc (Sean Cunningham)

Hey, thanks Steve...I'll try it out.  This would help out greatly (and save
me a few bucks :)

Alas though, until your utilites are out the MERGE feature can't be
implemented since I'm MORPHing between objects.

Sean
                                        /\
 RealWorld: Sean Cunningham            /  \     "Doing our business is what
      INET: [e-mail]   VISION                Amigas are for."
     Voice: [phone]             \  /
                                 //     \/      "Hasta la vista, baby."
    KEEP THE COMPETITION UNDER \X/   GRAPHICS         the Terminator, CSM101

##

Subject: Re: FAQ
Date: Mon, 22 Jul 91 15:28:11 EDT
From: Mark Thompson <[e-mail]>

Stephen Menzies writes:
> It got to me , and I look forward to your review. I did hear some
> negative things said about it's output quality and would appreciate
> any comments you have about that (among other things). Glad to hear
> you got it Rick, and I'll look into getting a pic or two to you
> as promised. Got some *really* clean and detailed Caligari Broadcast
> pics now, too.

Hi Stephen,

A while back in one of your posts to me, you mentioned being rather enthralled
with Caligari Broadcast. I was wondering what it is about this package that
makes it really shine above the others. I have largely ignored it in the
past because of its huge price tag and very poor modeler (at least it appeared
poor back when the program was first released). I would be very interested
in anything you have to say about it. Also, I read an article the other day
in Computer Pictures in which Keith Nealy from the Nealy Group had stated that
Caligari "may be the best renderer in the Amiga area". Care to comment on
what makes it so special?
|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
|      `       '        Mark Thompson                 CONCURRENT COMPUTER  |
| --==* RADIANT *==--   [e-mail]        Principal Graphics   |
|      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   |
|     Productions       [phone] [phone]   & General Nuisance   |
|                                                                          |
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

##

Subject: Re: Imagine to Lightwave
Date: Tue, 23 Jul 91 10:33:42 EDT
From: Mark Thompson <[e-mail]>

[e-mail] writes:
> 2) I have been able to successfully render animations in Imagine (the last
>      one was about 30 seconds) and use the Toaster to single frame them
>      to videotape. I even did one where I sequentially grabbed (24 bit)
>      frames from video using the Toaster (programmed it myself) and then
>      used the images as an animated brushmap in Imagine.
>      If people are interested in the procedure, and the ups and downs/pros
>      and cons I ran into let me know and I will post it.

I'd be interested to hear about your experiences with this including method
and equipment used.

> I'm not going as a volunteer, but I will be there from Sunday through
> Saturday. Maybe some of us should get together.

I have already organized a get together of Amiga graphics types at Siggraph.
The time and place are not yet set, but I will leave a note on the message
board under "A" adressed to Amiga Gathering or something like that. I intend
on getting some video equiptment so definitely bring a tape of your work
(I'm not sure but I assume VHS is all we will have to work with).

> By the way, Newtek
> is throwing an AWESOME party. They have a sincere desire to blow Wavefront,
> Iris and others out of the water. And they would like anyone who has done
> any Toaster animation work to bring a tape.

Yes, I had planned on attending. I'll be bringing a tape for that party
also but that one will be Umatic SP. I don't know if they will have a VHS
machine there. Also, the LA Video Toaster Users Group is planning some
sort of party. If you are interested, drop me a line.

> See you at Siggraph.

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

##

Subject: Re: New object stagnation
Date: Tue, 23 Jul 91 09:53:21 EDT
From: Mark Thompson <[e-mail]>

> There are about 100 objects on hubcap right now, but I'd like
> everyone to upload their best couple of objects just for something new
> to play with.

Well since I am testing this InterChange LightWave conversion module, I could
convert a few of my LightWave objects over to TurboSilver format and upload
them. Unfortunately, most of the surface attribute information will be lost
with the current version of the converter.
|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
|      `       '        Mark Thompson                 CONCURRENT COMPUTER  |
| --==* RADIANT *==--   [e-mail]        Principal Graphics   |
|      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   |
|     Productions       [phone] [phone]   & General Nuisance   |
|                                                                          |
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

##

Subject: Pic on hubcap
Date: Wed, 24 Jul 91 09:41:04 PDT
From: [e-mail] (Steve Worley)

I've put a picture of a German WWI era triplane on hupcap.  (FTP site
hubcap.clemson.edu, 130.127.8.1). I will happily post my triplane
object as soon as some other people upload some of THEIR objects.

Childish? Yep. But I figure I've already uploaded about 100 objects
already, and this way I can get other people to share their creations.

I spent eight hours last night (7 pm to 3 am) finishing my space
station. It's cool! Anyone have a pic of the NASA logo and/or an
American flag?

-Steve

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

##

Subject: black triangles
Date: Wed, 24 Jul 91 22:44:02 EET DST
From: Juha Kallioinen <[e-mail]>

Some strange things have happened to some of my scanline and color shade
renderings, that is , strange black triangles appearing in places where should
not be anything like that at all. The latest, a black triangle on the ceiling
of my scanline-overscan-lace-ham-room. What is causing these ?  Sometimes in an
animation the triangle appears in a frame, but not in the previous frame or
the next one.

I haven't been tracing too much, because of the time it takes, but in those I
have experienced no problems.

I could upload my latest triangle to hubcap.clemson.edu if someone is interested.
-Juha

--
[e-mail]
Juha Kallioinen

##

Subject: Re: New object stagnation
Date: Wed, 24 Jul 91 15:07:00 EDT
From: Mark Thompson <[e-mail]>

> Well since I am testing this InterChange LightWave conversion module, I could
> convert a few of my LightWave objects over to TurboSilver format and upload
> them. Unfortunately, most of the surface attribute information will be lost
> with the current version of the converter.

OK, they are now on hubcap and called markLWob.lzh. Here is the README file:

The following objects were created by me using LightWave and converted
using a beta-test version of the new LightWave converter module for
InterChange. They have been converted to TS 3.0 format and should work
with Imagine. Let me know how well they converted. One object, the reels
for the audio tape deck, would not even convert. You are free to redistribute
the objects if you include this file with them.

audio.reel.to.reel              14858
fan                             24444
fan.blades                       5378
gun                             21170
hubble.telescope                30374
magic.marker                    16166
movie.light                     23160
movie.light.stand                6672
rocket                          30626
synth.kybd                      62926
toy.plane                       17914
toy.plane.prop                   4934
toy.train.car                    8990
toy.train.engine                16052
zenith.star.laser               53896

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

##

Subject: Re: Imagine to Lightwave
Date: Wed, 24 Jul 1991 16:40:23 -0500
From: Donald Richard Tillery Jr <[e-mail]>

Hey Mark,

Rib NewTek about a standalone version of LightWave.  I've about given up on
them (this is the second time they've missed a marketing niche), but it
might not hurt to mention it in person and see if you can get a more personal
answer than "we are not planning to release a standalone version of LightWave"
which tells me exaclty what I already know and not what their reasoning is.
If they really do have a good reason in mind, I'd like to know for reference
in my future marketing plans.  They have more experience than I.

Rick Tillery ([e-mail])

##

Subject: Re: New object stagnation
Date: Wed, 24 Jul 91 19:19:40 EDT
From: Mark Thompson <[e-mail]>

> OK, they are now on hubcap and called markLWob.lzh.
> audio.reel.to.reel              fan
> fan.blades                      gun
> hubble.telescope                magic.marker
> movie.light                     movie.light.stand
> rocket                          synth.kybd
> toy.plane                       toy.plane.prop
> toy.train.car                   toy.train.engine
> zenith.star.laser

For those who would like to see what these objects look like, I have
uploaded a HAM pic containing all the above objects as they appear in
LightWave. It is in /pub/amiga/incoming/IMAGINE/PICTURES/markLWpic.lzh
on hubcap.
|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
|      `       '        Mark Thompson                 CONCURRENT COMPUTER  |
| --==* RADIANT *==--   [e-mail]        Principal Graphics   |
|      ' Image `        ...!uunet!masscomp!mark       Hardware Architect   |
|     Productions       [phone] [phone]   & General Nuisance   |
|                                                                          |
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

##

Subject: Re: black triangles
Date: Wed, 24 Jul 91 20:06:23 EDT
From: [e-mail] (John J Humpal)

I've never seen this missing or black triangle bug but I had heard about it.
Impulse was supposed to have fixed the bug in Imagine 1.1.  Is that the
version you are using?  I believe Impulse said the bug appeared only when you
tried to render objects with large faces.

John J. Humpal -- [e-mail] -- short .sig, std. disclaimer

##

Subject: Colorburst and wood
Date: Wed, 24 Jul 91 18:56:01 PDT
From: [e-mail] (Aaron Tucker)

Well, maybe I misworded something. The wood pics I have are in HAM. They are
the ones I downloaded from a BBS here. I was going to convert them to 24bit,
but figured that people who wanted to possibly use them for other purposes
or view them without ADPro would probably want them in HAM. I also don't see
if 24bit would really help since Imagine does that anyways.

As for Colorburst:

Yes, the Colorburst is capable of animation. BUT!, there is no animation
software available for it yet. You would have to program it yourself or wait
for animation software. The animation it is capable of depends on the number
of bitplanes and the resolution. In NTSC, in 384 x 480, you are capable of
25 to 30 fps animations from AMIGA RAM. The Colorburst communicates with
the Amiga at 5.5MB/sec. This is 24bit, BTW. In 8 bit mode, you can do quite
a bit more obviously. The CB only has 1.5MB RAM. So, large (XxYxD) pictures
will not fit. A 768x480x24 picture requires approx. 1MB RAM. That is only
one frame. For the other ones, you would have to go from Amiga RAM to
the CB  RAM.

Yes, you can have one monitor on the RGB port displaying a 24bit pic while
you use your 31kHz port for whatever. You will see garbage (flashing white
/grey screen on the 31kHz port when you are sending data to the CB. I am not
sure if this can be avoided or not. I know that the routines that are doing
that now are hardware brut force and just shovel the data through pretty
fast.

A program called BACKDROP24 is sent with the CB that allows you to mix CB
and Amiga graphics, display one or the other only, or seperate them to the
different channels. It is pretty powerful.

Juan
[e-mail]

##

Subject: The Imagine Companion Review
Date:    Wed, 24 Jul 91 20:04 PDT
From: ivan i <[e-mail]>

okay, as promised, my review of The Imagine Companion.
a quick summary of what follows:
(this may well sound like an ad, but since i have found
the book to be enormously helpful in my coming to understand/use Imagine
more fully, it seems only reasonable)

if you're a relative newcomer to Imagine, and feel that a lot of what
can be done with the program is missing you, or that you're wallowing
around in inadequate documentation, send your $29.95 (plus $2.17 sales
tax if you're having it mailed within California) to:

Motion Blur Publishing
915A Stambaugh Street
Redwood City, CA  94063.

additionally, i have found Dave Duberman, the author, to be extremely
helpful.  i had a problem getting the book delivered correctly, and he
took the time & expense to phone me to set things aright, and sent the
book out again.  so you don't need to worry about being screwed.

okay, as for this book:

it's divvied up into a number of different parts, for different levels
of expertise.  he starts off with a NOTES section detailing hints and
observations about the various Imagine Editors, which seem mostly geared
to helping you make the most out each one after you understand the basic
setup & operation.  These consist of a number of undocumented features
ranging from how to put a self running animation on a disk to faster
zooming to tweening attributes to what happens in the cycle editor when
you resize axes.  i spent the least time on this section, but it has
already given me quite a bit of useful information that i could
eventually have found out, maybe, but have condensed and available to me
here.  some of it i knew, most of it i didn't.

this leads into a section of TUTORIALS, some of which follow his Turbo
Silver tutorials for AVID magazine of not too long ago pretty closely,
but most of which are new at least to me.  They run from the very simple
- he starts of with lathing a cup - well, maybe i should list them all.

1.    lathing a cup
2.    painting & spinning the cup in an animation
3.    extruding along a path - which has come up recently on this list;
      this is the object extrusion type of path, not a path for motion to
      follow which is covered in the Imagine manual
4.    altitude brush mapping, which covers flat plane, cylindrical, and
      spherical brushmapping as well
5.    a lighting tutorial, covering the variations of lightsource options
      and the effects to be gained by placement in a variety of
      situations
6.    "fluted column" - which i have not done, but which is explained as
      using various detail editor functions, primarily extrusion, in
      intermediate object creation
7.    cycle tutorial #1, which covers the way object sizes & shapes are
      dealt with when passed from the detail editor to the cycle editor
8.    cycle tutorial #2, which details the creation of a walking figure
      less complicated than the walker in the manual, but which
      additionally covers how to get it to do its cycle in the stage
      editor when it "forgets" what it was told to do in the cycle editor
9.    a ripple effects tutorial which i also did not try, but which is
      said to cover using the ripple effect on a flat plane, concentric
      rings, and multiple objects.
10.   a clown's head tutorial, which really runs a gamut of techniques
      involving object creation, conforming to spheres, grouping,
      attributes, camera movement to achieve an effect, etc.  what it
      does is detail the process of creating a clown's head, with
      separately moveable eyes, and panning the camera back and forth to
      make the illusion that the head is shaking while the eyes are
      moving independently.  this one is pretty nice as far as i'm
      concerned, since i've mostly been using imagine to try out
      character animations. this is split in two parts, with part two
      giving approaches to head construction and a couple of extra
      explanations about why he did things the way that he did; the
      reasons were not at all obvious to me.
11.   spotlight tutorial, in which is specified how to make a beam of
      light visible as though it were travelling through a field of smoke
      or fog.  a nifty effect, i think
12.   depth of field tutorial, which requires a 24 bit display device
      liek DCTV or colorburst or whatever.  what it is is a method for
      getting the stills to look more "realistic" or photolike, in that
      you simulate the narrower depth of field that a camera would have
      as opposed to the superdeep focus of the computer.  there is a
      picture of this effect on the companion disk in DCTV format, and it
      is an interesting if perhaps kind of limited use effect

then there's a list of appendices, all of which should have appeared in
the Imagine manuals themselves.

appendix a details the updated features of Version 1.1, such as bug
fixes, and explains a bit about Fracture, Split, and Taut, as well as a
couple pros and cons of Cycle Setup & Cycle Shuffle.

appendix b is an addendum to the Imagine reference manual, and gives a
lot of information that should have been included in the refernce
manual. ditto for appendix c, which covers additions to the tutorials
manual (and also tells you how to get rid of the slow Imagine.pic
screen)

appendix d is the index to the reference manual which for damn sure
should have been in the reference manual in the first place.  according
to the book, this is floating around in the public domain somewhere, so
if you don't get the book, latch onto this if you can find it

all of these sections come in a spiral bound book, much easier to lay
open before your computer than the stapled binding of the Imagine
manual

lastly, there's a companion disk which has all the objects & such
necessary to do the tutorials if you a) want to skip actually making the
stuff & get to the meat of the matter, or b) want to check out how your
objects look vs. the ones the author has made.  it additionally contains
an Imagine.Config file which assigns various features to keys, so that
Pick is now F1, Attributes is now F7, etc.  i know i could have done
this myself, but it's a nice addition anyway.  these keys are used
throughout the tutorials, although the non-reconfigured way to do it is
presented as well, so you get a decent amount of practice in
using/memorizing them.

so that's the facts.  as for my opinion, i think i have made it pretty
clear that i really like this book and think that it's been tremendously
informative for me.  Dave Duberman apparently got thrashed for his Turbo
Silver manual pretty soundly, which surprises me because his AVID
tutorials & the material in this book are not only thorough, but
extremely well presented.  when he runs you through what to do in his
tutorials, he doesn't just list steps ("Now pick all; now select
attributes; now set dithering to...) but goes through his reasoning,
what he's trying to achieve with that particular manuever, and sometimes
the larger thinking process behind it.  i found that with the Imagine
tutorials, i was handfed a bunch of steps which (75% of the time) left
me with a cool picture but no real understanding of what it was i had
done.  the imagine companion, on the other hand, takes a nice teaching
approach to the lessons in an attempt to get you thoroughly familiar
with what is going on.

and it sure beats all hell out of that "Imagine: A guided Tour" video.

in addition, and you don't need to buy the book for this one, you can get a
free sample copy of AVID, The Amiga-Video Journal by sending your name &
address to Avid Publications / 415-112 Mary Ave. #207 / Sunnyvale, CA / 94086
its a pretty good magazine, certainly worth a stamp.

##

Subject: Colorburst
Date: Wed, 24 Jul 91 17:50:06 EDT
From: [e-mail] (Bob Lindabury - SysAdm)

rutgers!cs.unr.edu!tucker (Aaron Tucker) writes:

> P.S. I am going to upload the wood maps to hubcap as HAM, not 24bit. Anybody
> who needs to convert themfor other purposes can do so.

Huh?  Wouldn't 24bit wood maps be better?  Wouldn't it be easier to
reduce colors from the 24bit file than increase colors from HAM image?

Seems to me that uploading 24bit images exclusivly is the way to go.

-- 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: help with cycle editor
Date: 24 Jul 91 (18:30)
From: mit-eddie!uucp.cs.toronto.edu!lsuc!canrem!james.kewageshig

HI Again,
Was wondering if someone could explain how I could attach a rotor to
the helicopter body with the CYCLE editor in such a way so that I just
specify the number of rotations for the rotor and so that the body
becomes the PARENT object.

PLEASE E-Mail DIRECTLY!!           THX, James
+-------------------------------------------------------------+
|  NETWERK  |  James       UUCP: [e-mail] |
|  V I I I  |  Kewageshig   CRS: [phone]                |
|  1 9 9 1  |               CLI: [phone]                |
+-------------------------------------------------------------+
---
 ~ DeLuxe} 1.1 #8086 ~           *** UNDER CONSTRUCTION ***
--
Canada Remote Systems.  Toronto, Ontario
NorthAmeriNet Host

##

Subject: fog effect in imagine??
Date: 24 Jul 91 (18:29)
From: mit-eddie!uucp.cs.toronto.edu!lsuc!canrem!james.kewageshig

Hi All,
Was wondering if anyone could explain to me how the 'FOG' effect works
in Imagine v1.1.  There's a brief mention of in in the readme.docs but
no real examples.  In case you don't know what effect I'm talking about,
its in the readme.docs about the colour white completely dissipating
in a scene. (thats about how descriptive they were in the docs ;) )

PLEASE E-MAIL me directly since I don't have real access to this group
or newsletter. If this has been covered in a previous letter, please
E-MAIL me the reponses.            THX, James
+-------------------------------------------------------------+
|  NETWERK  |  James       UUCP: [e-mail] |
|  V I I I  |  Kewageshig   CRS: [phone]                |
|  1 9 9 1  |               CLI: [phone]                |
+-------------------------------------------------------------+
---
 ~ DeLuxe} 1.1 #8086 ~           *** UNDER CONSTRUCTION ***
--
Canada Remote Systems.  Toronto, Ontario
NorthAmeriNet Host

##

Subject: wood
Date: Thu, 25 Jul 91 00:36:34 PDT
From: [e-mail] (Aaron Tucker)

I DON'T HAVE 24BIT WOOD BRUSHMAPS! I DON'T HAVE 24BIT WOOD BRUSHMAPS!
maybe that'll work.
I HAVE HAM WOOD BRUSHMAPS! I HAVE HAM WOOD BRUSHMAPS!
that should definately work.
If you want 24bit wood brushmaps, you'll have to get them yourself. I don't
have a digitizer nor do I have access to 24bit textures.
Yes, 24bit would be much nicer, Bob. I agree with you 100%. Unfortunately,
I don't have any. I was merely asking if it would be needed to convert them
to 24bit before uploading. I decided that it wasn't necessary. Agree? Thanks.

Juan Trevino
[e-mail]

##

Subject: Why does SCANLINE take more RAM than TRACE?
Date: Thu, 25 Jul 91 00:27:24 CDT
From: mit-eddie!harvard!scubed!pro-party.cts.com!seanc (Sean Cunningham)

I'm sure this has been asked before, so please bare with me :)

An object I'm working on (human figure...sorta) is progressing nicely, but
I've just run into a major problem.  When I try to do a scanline rendering
of the object, not all of the pieces render (I'm assuming that a low memory
situation causes truncation of parts), BUT I can render the whole thing if
I go to full trace.  Why is this?  Why would scanline take more memory to
render than raytrace, when I'm dealing with the same object under the same
staging file?

Note: I'm running on a 4M A3000-25/50 (2M Chip, 2M Fast).  Because it won't
      render at all if I run from WorkBench, I have a script in my S:
      called "IMSHELLSTART" that makes all the necessary assignments,
      runs CPU, and starts the program.

Thanks!

Sean
                                        /\
 RealWorld: Sean Cunningham            /  \     "Doing our business is what
      INET: [e-mail]   VISION                Amigas are for."
     Voice: [phone]             \  /
                                 //     \/      "Hasta la vista, baby."
    KEEP THE COMPETITION UNDER \X/   GRAPHICS         the Terminator, CSM101

##

Subject: Re: C Hackers
Date: Thu, 25 Jul 91 00:18:19 CDT
From: mit-eddie!harvard!scubed!pro-party.cts.com!seanc (Sean Cunningham)

Mark, you mentioned modelling stuff like fire, clouds, etc. with Lightwave
using particle systems.  Do you know if Allen is planning on adding
"lifecycles" to the particle systems generator?

I seem to remember he had mentioned that he planned on adding trajectories
to the particles (so that you could grow grass, etc.) in an E-Mail post a
while back (is this in v2.0 of Lightwave???).  This would be really cool,
and only make me want it all the more.  Not really for grass (which would
be useful), but I keep thinking of the whiskers on the Chinese dragon kite
from the "World of Discovery" opener...lotsa possibilities for all kinds of
effects there (for anyone that hasn't seen this opening (I believe it'll be
shown at SIGGRAPH), it is amazing.  It uses _TONS_ of particle systems).

Sean
                                        /\
 RealWorld: Sean Cunningham            /  \     "Doing our business is what
      INET: [e-mail]   VISION                Amigas are for."
     Voice: [phone]             \  /
                                 //     \/      "Hasta la vista, baby."
    KEEP THE COMPETITION UNDER \X/   GRAPHICS         the Terminator, CSM101