#Raytracer other than Ray
38 messages in this thread
Will 3DS ever come with raytracing or is there an agreement between the makers
of Rayman and Autodesk?
Will
Autodesk doesn`t discuss features of upcoming products <bg> and we also don`t
do if it hits this point. Sorry.
Michael
Thanks anyway!
Will
I read in a book about computer graphics that said the programers of 3D studio
were opposed to raytracing because it took to long and they believe you can
achieve the same effects faster with raycasting. I don't know if the book is
acurate or not but there it is.
—Will
I'm not sure what the point of your comment is but it is true that Raytracing
is considerably slower. As 90% of the time its a waste of rendering time, the
Yost Group has wisely IMO not chosen the Raytraced route, which in fact is much
easier to code. It is also considerably more difficult to get an artistic kind
of feel to your work with RayTracing. For none artistic types Raytracing is
easier to get a Photorealistic look. The down side is that it becomes a crutch,
and one doesn't fully develope one's artistic skills. If one plans to develop a
distinctive "style" this IMO is a real drawback. From a production standpoint
"speed" is an important factor, but its becoming less so, with faster
platforms. The "look" and range of "looks" is where 3DS really excells, all
else being equal. In the end it comes down to personal taste and talent. I
think if you've got the talent you'll prefer 3DS.
John
<<The down side is that it becomes a crutch, and one doesn't fully develope
one's artistic skills. If one plans to develop a distinctive "style" this IMO
is a real drawback.>>
strange, dull glass and chrome is a trade mark, do you want to stay with this
look for the next two years?
<<In the end it comes down to personal taste and talent. I think if you've got
the talent you'll prefer 3DS.>>
3D studio user's do not have a choice. only one option. [ with core rendering]
some artist like to mix
paints to get that look. since raytracing is one of the bigger wish's, thinking
like your post are
only setting us back two years or so.
Brandon
Hi John,
->"Raytracing is considerably slower. As 90% of the time its a waste of
rendering time,"
->"It is also considerably more difficult to get an artistic kind of feel to
your work with RayTracing. For none artistic types Raytracing is easier to get
a Photorealistic look. The down side is that it becomes a crutch, and one
doesn't fully develope one's artistic skills."
John, please understand that I normally believe your posts to be extremely well
thought out and accurate, and I've learned alot of valuable things from you
online. I really have the greatest amount of respect for you. In this case,
however, I must disagree completely with your comments. I think you are leading
Mr. MacDougall down a narrow path, despite his somewhat arrogant attitude.
Myth#1 – "Raytracing is considerably slower" – This is completely dependent on
which raytracer you use, and is a function of how well the code is written and
optimized. There are alot of really bad implementations of raytracing out
there, but there are quite a few excellent ones that don't lend to the myth.
For example, the watcom-compiled version of the Vivid raytracer absolutely
screams, especially with respect to raytraced shadows. Using the 3DS2POV
utility and outputting 3DS scenes to Vivid, I've found it's actually *much*
faster at rendering scenes with raytraced shadows. Other raytracers like POV
and PolyRay have also been greatly improved from a speed standpoint. I think
its somewhat flippant to automatically cast aside all raytracers as
"considerably" slower.
Myth#2 – "Raytracing is easier…it becomes a crutch" – Huh? Regardless of
whether you use scanline or raytracing, you still need to create a
fundamentally sound lighting and surfaces treatment to obtain decent
photorealism. And even if it's easier to setup reflection and refraction, this
isn't a crutch, it's a feature when considering this as a specific goal. Bottom
line is you have to work harder if you want this effect in 3DS. I'm not sure I
consider that an artistic "skill", more like an obstacle.
Lastly, your comment that "If one plans to develop a distinctive style, this
IMO is a real drawback" floored me. Have you seen the works of Truman Brown,
Mike Miller, Dan Farmer, or Steve Anger? They all have extremely well
developed, distinctive styles that I consider to be on par with any 3D artist
I've ever seen using any package.
->"I think if you've got the talent you'll prefer 3DS."
Aaaauuugh! It's not an EITHER OR situation. This is the thinking that spawns
the endless Lightwave-vs.-3DS type debates, and it's irrelevant.
Use *both* if you want, the technology is there. It's OK to wish for these
features in 3DS, they don't have to *replace* the scanline renderer, which I
consider to be excellent the way it is. I think all anyone wants is a toggle or
button that enables selective, and optimized raytracing features. How could
anyone object to this?
Best regards,
-Heinz.
Heinz,
I appreciate your recognition of my previous contributions.
Some people are interested in 100% photorealism. I am not. I do not think it
takes any particulary high degree of artistic skill to achieve it. Much like
with a camera that can be obtained. The compositional nature of the
"photograph" may or may not be art. So hopefully that clarifies that point.
Much like paint and canvas one has to be artistically creative to create
photorealism with 3DS. I personally feel that the days of photorealism are
somewhat passed, but there does seem to be a resurgence in Europe.
I think though that if CG is ever going to achieve status as a "fine art" its
going to be along the lines of 3DS / Renderman et al. not along the lines of
ray-tracing. You may not agree but at least I hope you get my drift. 3DS is
much more similiar to paint and brush than raytracing.
As far as leading Mr. MacDougall down a narrow path goes, I rather doubt it. I
don't exclude ray-tracers, I just don't see any need to include in 3DS. That's
just my opinion.
In terms of rendering speed you are correct that optimized code makes a
difference, even so as you noticed I down played the importance of speed as I
think its becoming a somewhat mute point. In context I think I gave it a fair
assessment.
My statements viewed as you have would appear to be not as complete
viewed in the light you've presented them. I do feel that 3DS
challenges one's artistic skills and development more than
"ray-tracers" do, unless you feel that it comes down to texture
mapping as you suggest. I feel there is more range in what can be
achieved in terms of a "painterly" look to a "photorealistic" look
than what can be had naturally with a ray-tracer. Most people who work
with them are going for "photorealism" exclusively. I am interested
in a less CG look and more of an airbrushed or paint box look and 3DS
does a beautiful job of not giving away that it is rendered. Call it
an anti CG generated look. But at the same time I can get that CG
look if that's what I want. I think that without exception the
potential range of 3DS is broader and more in line with Renderman in
that regard.
And of course the composition of a work can illuminate the "style" of an
artist, but most would guess that the artists you mention are using
ray-tracers, you may or may not recognise the same with 3DS.
Of course we all have our preferences and that is mine. Objectively I cannot
find fault with your comments other than the "myth" portion. The points I've
made are not in the myth category, possibly biased by relatively accurate.
Thanks for your comments, as I think they have helped to round out the
perspective. However, I find that for every comment it would be easy to explain
with a chapter. You've made some good points that needed clarification, and I
hope I've done that, even if the final result is not total accord.
John
Your discussion of ray tracing and speed seems to me to concentrate on relative
speeds of raytracers. There are ray tracers that are faster than others. Ray
tracers are being sped up. This one ray tracer is even faster than 3ds at
raytracing.
My experience has been that ray tracing is accepted as an _inherently_ slower
process for rendering than the other options out there. And is that not what
was originally said? That 3DS is not a ray tracer because to do so would be a
speed sacrifice not worth the particular strengths of a ray tracer?
>>This is completely dependent on which raytracer you use,<<
This is not true. Ray tracing is a known platform and implementation
independant technique that can be compared to other techniques without
considering specific instances.
>> and is a function of how well the code is written and optimized<<
This statement is true of anything. Substitute line drawing; z-buffer;
or egg scrambling for raytracing, and it holds true _inside_ that
given topic. If 3DS had been written to verify data integrity before
proceeding on to the next pixel, even Preview mode would be slower
than ray tracing on my pocket calculator.
Hi John (aka lightning rod),
Thanks for your reply.
-"I don't exclude ray-tracers, I just don't see any need to include in 3DS.
That's just my opinion."
Acknowledged, I see now where you're coming from in terms of your feelings on
photorealism not being your goal.
There's no question that 3DS is a considerably more versatile tool in
terms of the variety of different looks one can output. And without
any hesitation I would agree with you that it provides for a huge
variety of output styles. However, it would seem that alot of 3DS's
features are geared towards obtaining photorealism, and in my opinion
RT features are a natural fit towards this goal.
Since your work focuses primarily on animation and mine primarily on stills, we
have different needs hence our different viewpoints. For me, I find 3DS to be a
complete 3D production tool with the exception of it's raytracing capabilities.
I regularly find myself needing fast accurate shadows, reflection, and
refraction, and spend alot of time working around 3DS's limitations in this
area primarily by sending these scenes out to an external raytracer.
Cheers,
-Heinz.
Hello Heinz –
I was pretty impressed with Vivid 3.0's raytracing speed also! I can really
crank out the images on my P100. Did you notice that Stephen added field
rendering? Time for a little video animation…
Jim
Hi Peter,
Thanks for your thoughts on this thread.
->"My experience has been that ray tracing is accepted as an _inherently_
slower process for rendering than the other options out there. And is that not
what was originally said? That 3DS is not a ray tracer because to do so would
be a speed sacrifice not worth the particular strengths of a ray tracer?"
Yes that was what was said. To say that something is "accepted" as being an
"_inherently_ slower process for rendering than the other options out there" is
just that, heresay. What does it have to do with reality or specific facts?
Let me restate more clearly:
When comparing *PC* raytracers and partial implementations of raytracing (like
3DS's rt shadows), the comparitive speed is dependent upon how well the code is
written and optimized for the given application. Depending on which raytracer
you use, your rendering times may be vastly different.
>>and is a function of how well the code is written and optimized<<
->"This statement is true of anything. Substitute line drawing; z-buffer; or
egg scrambling for raytracing, and it holds true _inside_ that given topic."
Look, please don't disect my sentences and quote me out of context. My meaning
was confined to raytracing code and not scrambled eggs.
I have the greatest admiration and respect for 3DS, and have used it since r2's
release. I also use raytracing quite often, and have used RT programs since the
80287 chip was available. I have experience with most freeware, shareware, and
commerical raytracing implementations on the pc. Based on my experience, I can
unequivocably state that some raytracers are mature, optimized 3D rendering
engines that approach the speed of scanline renderers. And some are brutally
slow, poorly implemented and "not worth the particular strengths of a ray
tracer". To summarily dismiss them all as being not worth the trouble is, IMO,
an uneducated action.
Cheers and best regards,
-Heinz.
>> some raytracers are mature, optimized 3D rendering engines that approach the
speed of scanline renderers<<
That is of interest. I can assume that Vivid falls into or approaches this
category? I probably won't do anything with it, unless it's real real cheap
though. 3DS is a budget pusher for me.
– gruhn
>>> I probably won't do anything with it,
>>>unless it's real real cheap though.
>>>3DS is a budget pusher for me.
Vivid is shareware for the old 287 version. To get the newest, 387 version, it
costs $50. I wouldn't call that a budget buster <G>…
I wrote an article on raytracing 3DS scenes with Vivid in a recent Planet
Studio.
Jim
>> $50. I wouldn't call that a budget buster <<
Heck, except for right now, neither would I. I'll have to check it out. Thanks.
– gruhn
>>I wrote an article on raytracing 3DS scenes with Vivid in a recent
Planet Studio.
I just wanted to let you know I enjoyed this article very much. It has
inspired me to dive in to Raytracing. I have been holding off buying
Rayman, as I would hate to lose my $600 investment if it is not
supported (or included) in R5. I am going to give Vivid a shot.
As far as Mr. Coy accepting beer for his "beggarware", I have been a
devoted home brewer for some time and plan on sending him a 6-pack of
my best Stout.
>>I am going to give Vivid a shot.
Let me know if you have any trouble. I think you'll be impressed with
the output.
>>As far as Mr. Coy accepting beer for his
>>"beggarware", I have been a devoted home
>>brewer for some time and plan on sending
>>him a 6-pack of my best Stout.
Oh, that's going to be great! I'll have to ask him about that after
you send it! Only one other person ever actually did it, I think they
were in a foreign country, too.
Thanks for the nice comments on the article.
Be sure to get the 3.0 beta as soon as you're registered – it's much,
much better than the yucky 287 version of vivid 2.0.
Where does one find Vivid? How does it compare with POV Ray?
>> I think all anyone wants is a toggle or button that enables selective, and
optimized raytracing features.<<
I've looked at the ray trace work-around in the Special Effects book. I was
stunned by the technical merit and quality of the final image.
I would also like to have a selective ray trace button.
> I think all anyone wants is a toggle or button that enables selective,
> and optimized raytracing features. How could anyone object to this?
Excellent point! _NOBODY_ could object to having the option.
– Ernie Jackson, Written 18-May-1995 @ 12:40:44
Using: Windows NavCIS PRO 1.6
Hi John,
as Mike "RayMan" Bassier as want to write some words about your message <g>.
First of all, I`m a 3DS user and I make SFX for TV commercials and advertising
spots. Very often I have to combine real images with CG. For this point
raytracing is a great help, I don`t want to miss it. Raytracing is not a holy
cow. It doesn`t solve every problem. But if you want to make realistic images
you can`t afford to miss it!
If you want to make artistic images maybe you don`t need it. But it`s a nice
option. I saw great, artistic images wich were raytraced.
You`re right, raytracing is slower than scanline rendering. But raytracing can
be much faster than automatic reflection maps and raytraced shadows. It depends
on the program.
The production standpoint: If you have to fake reflections and refractions by
hand it`s much cheaper to put all that work to a PC. Rendertime is much cheaper
than manpower. If it`s too slow buy a faster PC (or a faster raytracer).
Michael
Hi Mike "Rayman"!
Been awhile since we've shared a thread or two. 🙂
>> First of all, I'm a 3DS user…. <<
Personally while not having used your product, I think the solution you provide
is the kind of solution that should be available for people interested in
having a "ray-tracing" capability with 3DS.
There is no doubt that "ray-tracing" has a place in animation as evidenced by
the variety of packages available for those who wish to use it as part of their
work. Whether it makes sense to fully incorporate it into 3DS as a matter of
course is highly debatable. (Obviously <g>)
>> If you want to make artistic images maybe you don't need it. <<
Well its really more than that. Its a fusion of artistry and realism as opposed
to stark "photorealism". And that is to a degree a matter of taste and style.
If you need to incorporate animation with real imaging then "ray-tracing" or
"radiosity" may be the answer. But not necessarily. Not so many years ago,
producers wanted it to "look" CG; they were paying alot of money for it and
they wanted people to know. <g> And I'm not sure that its changed all that much
except that no one makes a specific point of it as they once did.
But this really gets away from the main thrust of the debate which is whether
it should be incorporated into 3DS and my feeling is that third party solutions
such as your own are the answer for that segment of the market who has that
particular need, focus, or interest.
I don't expect to have the last word on this but I can express my opinion that
I would rather see the YG team continue to develop a "user-friendly"
implementation of the best of what is possible with scanline rendering (the
most versatile form of renderer in my opinion) as evidenced by the full set of
Renderman shaders (which is not friendly to non programmers) and leave the
"ray-tracing" or "radiosity" approach to those who choose to develop that
aspect of renderers. To me that mission is clear. And products like your own
bridge whatever percieved gap that may exist in the "super-realism" relm,
without taking away from that effort, or diminishing the greater/broader
capabilities which have made 3DS the highly respected product it is. I
pesonally think that the YG team has a clear sight on what needs to be
accomplished and this debate seems to lack an appreciation of where exactly we
are headed and what is involved. I suspect as with the advent of R5 it will
become more evident than ever. The ultimate goal is a laudable one. In the
meantime some people will choose to use both and there is nothing wrong with
that. I would hope that you would continue with your efforts to refine your
product as well, as there certainly appears to be an interest in the percieved
benefits "ray-tracing" affords. I hope those that have shown the interest will
pursue the option you provide and that it will be mutually beneficial to all
concerned. Your points are well taken in the scope that they encompass. But
lets not get away from the main focus of where 3DS is headed, which I think,
knowing what is possible; is the ultimate course, and the rest of this is
rather a moot point as there are a variety of options, including your own.
John
Hi John!
>>Been awhile since we've shared a thread or two. 🙂
Yes, I was very busy in the last weeks. I only had time to read messages but
not to write anything rich in meaning <g>. I`ll upload a little piece of my
work in the next days. Look out for "raytraced animation". It`s a very
realistic piece. I love realistic things, that`s my "stile of art".
If I had a wish I would like to have the feature of external shader. If anyone
want to have raytracing or radiosity, it doesn`t matter why, he or she should
have it.
The Yost group do it in the right way: they build an open animation package.
And that means that any kind of discussion like this thread is senseless. If
you want to do anything with you program you should do it. The only question
is: who gives us the possibility/feature?
I hope the new R5 allows us to plug in any IPAS seamless. And when this is
possible it doesn`t matter whether it`s an build-in-features or a third party
application.
Michael
Michael,
Your points are good ones and while I could expand on my perspective, your
zeroed in on the key point and that is that 3DS is open architecture and does
provide the facility to incorporate "special features".
The points that have been made in favor of ray-tracing being incorporated are
interesting, I don't think as issues go it quite merits the attention its
getting but I can see from some of the perspectives on this subject that some
people might have a real need for it. I especially like your idea of a
"plug-in" type ray-traced shader. I've really got to wonder at the
ramifications but from my standpoint it does have some excellent possibilities.
I can equally see more than a few potential hazards but fortuneately its not my
decision to make. <g>
In the meantime I wish you well with your endeavours and hope that you can
accomodate the needs of those who have shown an interest in this, as those who
do have the interest are quite passionate about it. 🙂
BTW I'm really not anti raytracing just real particular about the "look". FWIW
I think Ray Dream Designer is a slick little ray-tracer but of course its
primarily intended for stills.
John
Hi John!
The idea of plug-in-shaders is not very new but I don`t know much programs
with this feature. Renderman and Mental Ray are good examples and I`m sure that
3DS will be an other. As soon as 3DS has an open rendering interface, we`ll
might see three products from us: RayMan (for raytracing), furman (<g> for hair
and fur) and cloudman (for any kind of real clouds).
RayMan (specially Release 2) is already prepared for this kind of plug-ins and
the other two shaders, well I don`t want to say too much but I'm sure we`ll
surprise the 3DS community. We have to wait for the new Windows NT version…
BTW: you can fake plug in shaders with the video post queue but that`s not as
perfect as I want to have.
Michael
Mike,
>> The idea of plug-in-shaders is not very new… <<
Yes I realise that, but the idea that a ray-tracer might be a plug in is an
angle I hadn't thought about. If the whole rendering module was de-linkable and
there was a port to join a different rendering module, that would certainly be
different, not to mention unlike a shader which would still address the
rendering module and would feed it. Incorporating different renderers in the
same scene would be pretty radical too, especially with dual processors. Then
you could have a variety of renderers for different aspects of your animations.
Sounds interesting and something I'm sure you're alot more in tune with. I
don't think about that aspect as much because at present its a fixed entity,
but from your standpoint and because of what you do I'm sure its something you
think about quite a bit.
Sounds like you've got some interesting projects in the works. Its been awhile
but I wanted to be clear on this. Your Rayman is a standalone that runs in DOS?
It probably wouldn't hurt to run it down one more time, if you're willing. I
kind of tune out on things that aren't relevant but this discussion has raised
my awareness so I'd be receptive. <g>
Cloudman sounds cool, that's something I could definitely get into. There's a
new network open with some radical vapors …. I can just imagine the SGI's
cranking away on it… its pretty unreal.
Fake shaders in VP? No that's not what I had in mind either. VP is good for
image processing not shaders, at least in the terms that I think about them.
Keep me posted. I don't suppose you've got a crippled version of Rayman I could
check out? (I promise to keep my opinions to myself unless they're favorable
seeing as I do have somewhat of a bias. <g>)
Ciao,
John
John,
everytime I stand under my shower I`m thinking about new computer animation
stuff (software, concepts, ideas…)
Often I think a raytracer SXP would be great. Imagine you aply RAYTRACE.SXP as
an reflection map. You would have an dialog box where you can set all your
raytracing options. Now, everytime the 3DS renderer hits your object the SXP is
called by the 3DS API. Now the SXP calculates the pixels of your object.
This is only the first step of my thoughts and so there`re not perfect and
they might be some technical problems but it brings idea on the right way:
Let`s see what the A-Buffer of 3DS (should) do:
everytime a pixel on the screen belongs to an object with a SXP the renderer
calls the SXP to calculate the color of the pixel. It doesn`t matter for the
3DS render how the SXP calculates the color of the pixel (it may fire light
rays through the scene or use complex fractal algorithm), the result is the
same: the SXP returns the color value to 3DS. The 3DS render engine continues
with the next pixel and so on…
You can share this process over endless number of processors.
So, if you have an object like a coke bottle, apply the raytracer-SXP to the
bottle. If you have some cloud objects, apply a cloud-sxp. What are already
doing if you want to have water or smoke? You apply WATER.SXP or SMOKE.SXP. So
get a little step further. Think about RAYTRACE.SXP, CLOUD.SXP, FUR.SXP and
LIQUID.SXP…
Doesn`t that sound good? Let`s wait for the new 3DS version. I haven`t seen
anything from this version but I know it will beat Softimage or Alias and so
on… Why? Because the developer team doesn`t work for money (not only <g>) but
for their dreams and for the user community. More than 50,000 user have
deffinetly enough creativity to make everything possible.
>>Sounds like you've got some interesting projects in the works. >>Its been
awhile but I wanted to be clear on this. Your Rayman is >>a standalone that
runs in DOS?
Yes, Release 1 of RayMan is a DOS standalone program. Ok, you can run it from
the DOS shell of 3DS or in a DOS window of Windows, Windows NT or OS/2 but it`s
only a DOS program.
If you use RayMan in your production cycle you have to create your 3DS scene,
leave 3DS, start RayMan, make your rendering settings and let your image
render. I know it would be better to have it implemented in 3DS but it wasn`t
the time to do that. So let`s wait for the next version. Can`t say more but
that it`s faster, better, further <g>.
Michael
Michael,
I don't see a big problem in something running in DOS as long as it will run
under NT. I think DOS is going to be under the hood for awhile. The beaty of NT
is that you can strip it out at anytime and it has no effect on the OS. Pretty
clever really.
As far as your future plans go, the SXP's sound very interesting, I'll be
interested to see if its do-able.
>> that it's faster, better, further… <<
Sounds like you've taken a page from the Yost "how to" manual. <g>
Keep us posted.
Regards,
John
>> shell of 3DS or in a DOS window of Windows, Windows NT or OS/2
>> but it`s only a DOS program.
When I try to run RayMan in an NT command-line window it complains
about not being able to see the dongle. Am I missing something?
<<It is also considerably more difficult to get an artistic kind of feel to
your work with RayTracing. For none artistic types Raytracing is easier to get
a Photorealistic look. The down side is that it becomes a crutch, and one
doesn't fully develope one's artistic skills.>>
The inclusion of a "Raytracing option" on an object basis doesn't present this
problem. There are many projects where "raytracing" or "refraction mapping" are
NEEDED and not just a nicety, even for projects of a more *artistic* nature
than photorealistic.
3DS provides the realism in all other areas but is lacking in this one. To not
include the capability to accurately refract glass and other similar materials
would be very disappointing if not included in the next version.
Everyone I talk to say that 3DS is great BUT the still-images it renders aren't
as clean and neat as those produced by many other packages. I always respond
that 3DS is a production tool used to generate mass quantities of animations.
Topas/Electric Image do produce very clean pristen images easily. In 3DS it
takes a lot more tweeking to get such a clean net effect(TEXTURE-SHARPEN etc),
but it is possible.
<<If one plans to develop a distinctive "style" this IMO is a real drawback.>>
Too often it is the client that dictacts the style. If they want a
photorealistic glass than they *get* refraction but only through a lot of
fudging. An option to do selective ray-tracing/refraction mapping is the single
tool missing in the toolbox and would round out the capabilities offered by
3DS.
DaviD "It's all an optical illusion" GouID
David,
As usual you make valid points, but again its not my experience. Glass is not
something I have a need for typically except in terms of "glass lettering" and
I find that I can make a very acceptable animated glass.
Stills may very well be a different story. As I don't do stills very often
except as a "work in progress" so it really hasn't been a factor.
I have no idea what is happening with R5 and I wouldn't care to speculate on
the possibility.
FWIW I don't even use ray-traced shadows as in most cases there's an faster way
to do it. Although in some cases I know I would have to resort to it, if I had
to.
As there are a variety of alternatives many of which have been discussed here,
I don't see the need. I realise that that may be unique judging from the
responses I've recieved but then the vast majority haven't so I feel that this
issue is not as big as it might appear for most people.
BTW If I get inundated with responses, I won't try to answer them all. <G>
John "Lightning Rod"
<<As there are a variety of alternatives many of which have been discussed
here, I don't see the need. I realise that that may be unique judging from the
responses I've recieved but then the vast majority haven't so I feel that this
issue is not as big as it might appear for most people.>>
I think it all boils down to the fact that so many different people with
different needs, wants, and desires are using 3DS to meets their particular
goals. The goals of some are at the complete opposite poles of others,
"realism" in contrast to "abstract surrealism".
The fact that so many people can get out of 3DS the style they want is a
testament to the power and flexibility of 3DS. Some people see a lack in 3DS's
ability to cater to a particular style, in this case photorealism. 3DS at the
moment doesn't cater very well to *certain* realistic needs and the inclusion
of a ray-tracing option for objects would fill this gap.
Your comment concerning the fact that the Yost Group should focus on making the
product more user-friendly is a valid one but for others to have a realistic
reflection or refraction is of equal importance. It will be up to Gary and the
gang to decide where their priorities lie. As Gary often says "EVERYTHING is
possible given the time and resources".
Gary did comment some time ago that the most ideal graphics system would act
like an operating system, thereby allowing a lot of different
features(applications) to function under a single system. Under such a
mechanism you could add or subtract features as you felt they were necessary.
3DS goes part way there already, who know what they have in store for tomorrow?
In any case, the needs of the individual users are so diverse and varied that
an IPAS/Plug in architecture is the sole means to keep everybody happy. The
question is what should be core code and what should be external(Plug-ins).
The implementation of a different rendering method like radiosity or
ray-tracing would seem best suited to an external plug-in since they are
optional and not always needed. I briefly looked at the Quickdraw 3D
specification. It includes the ability to define external shaders.
So will 3DS be heading in the direction of a graphic operating system? I
personally think so, since this will provide the most flexibility and cater to
the largest possible market. You'll simply buy the building-blocks that suit
you and assemble a system that caters to your particular needs. And if you
can't get what you want then someone will provide a simple scripting language,
preferablly in the style of Hypertalk or OpenScript, so that everyone can
customize there system till the cows come home.
Now this last feature is the most exciting of all. If everyone were allowed to
tweak and change the graphics capabilities as they want, then the product would
become truely universal in the sense that everyone would NOW be contributing to
the evolution of the product. Instead of having a single group of programmers,
i.e. the Yost Group, you'll have the input of thousands of people throughout
the world contributing not just ideas but functioning building-blocks that push
the product in new directions. The MEIER IPAS is a good indication of this. If
this could be more firmly integrated into the product, something a graphics
operating system would provide, then we could all move off in this new
direction and from the input of everyone continue moving in different
directions evolving the product in new and different ways. The diversity of
features provided would be enormous catering to EVERYONE's needs and desires.
It would be a collective effort that would see a rapid explosion of
"building-blocks" uncomparable to the stream of new IPAS routines we are
currently seeing.
Everyone would be producing specific "building-blocks" that they could sell at
very low prices. TSE currently prevents this since the source scripts need to
be provided which means the author loses out since he has no competitive drive
to product scripts that HAVE to be given out freely if they are to be used.
Gary did mention that IPAS will be handled differently in the next release and
also the issue of pirating IPAS routines. An identification system can be more
easily integrated into Windows NT than DOS so with the assurance that your
"building-block" won't be *stolen* you can produce more for less cost since
you'll see more profits. These very low priced "building-blocks" will push the
product faster than any single development team ever could.
The power would be in the hands of everyone. W'ere heading into interesting
times and the diversification has only just begun.
DaviD "I'll take two ray-tracing shaders, a scan-line shader, and a side of
fries" GouID
David,
>> Your comment concerning the fact that the Yost Group should focus on making
the product more user-friendly is a valid one… <<
Actually that's not what I was saying. What I was saying was to implement a
more user friendly interface than what Renderman has to achieve the same
result. I've heard people say that others have said 3DS is not that easy to
use. I have no way of judging but from my standpoint its been the easiest
package I've ever used, so its all relative I guess. I'm talking about
professional packages btw not some of these hobbyist programs, like Truespace
which actually I don't care for.
>> Realistic reflection or refraction…. <<
Why not talk Mike Bassier and see if his ray-tracer doesn't work for you? Is it
seamless intergration that you want or the ability to get the desired result?
>> Gary did comment some time ago that the most ideal graphics system would act
like an operating system, thereby allowing a lot of different features to
function under a single system… <<
Yes and with DLL and DDE in NT all this should be possible provided that its
supported by software vendors which I expect it will be mostly. I don't like to
second guess what he had in mind but I think it would be fair to say that you
could have a ray-tracing program working through this linkage to obtain just
want you want without it necessarily being a component within 3DS. I
personally feel with the Plug-ins and NT we're getting very close to the ideal,
but it probably won't be a seamless as a pre-integrated system, but should get
fairly close.
So the thrust of your comments are totally in accord with what is likely to
happen and we both share the same vision of where we'd like it to go, with the
caveat being that Ray-tracing in all likelihood would be an add-on and not as
part of the main body of 3DS. I think more than anyone the Yost Group has
pioneered this focus in the realm of animation thereby allowing them to focus
on what they needed to do and providing the opportunity for others to support
the "special" interest. Even Bones Pro is not a product for everybody and I
think ray-tracing falls into the same category.
And personally I'd rather see it done that way as I don't want to pay for a
ray-tracing program I don't use, and I like the idea that I have options if I
do. Anyway, I'm sure you've registered it as a wish list item and more than
that I'm just expressing my opinion on the subject, which carries absolutely no
wieght at all. 🙂 You know who you have to talk to… <g>
John "Enjoying the option"
Will,
I have long wished for raytracing in 3DS. I have on occasion the need to
produce very acurate renderings of reflections in mirrors. This involves
reflections off of flat surfaces. As I understand, the 3DS method estimates
reflections, doing a fine job, but not 100% scientificaly acurate. Since my
work is mosty engineering visualization, I am more interested many times in
accuracy than artistic flair. My solution was to purchase Truespace by
Caligari (a Windows 3.X program, aprox <$500, I think), which imports 3DS
geometry (as *.3ds files, without textures), and allows for raytrace
renderings. I have run tests on the accuracy of the raytracing, and it does
very well. Truespace also allows you to render without raytracing, as the
raytracing does take much, much longer in Truespace. That sure is an option I
wish 3DS would have. I have not, howerver, found too much more in Truespace
that is not offered in 3DS, and since 3dS does everything else I need very
well, I haven't fully explored Truespace.
Hope this helps.
Rich Schwartz
NASA Langley
Thanks.
Will
PMJI: I don`t understand why you don`t use RayMan. This raytracer is developed
for exact the problems you discribed. You can import all mappings, lights,
cameras and settings and it`s even a few hundred percents faster than
Truespace. Why don`t you make your scenes completly with 3DS and render them
with RayMan?
Michael
Michael,
I wish I could get Rayman, unfortunately, our project can not afford
it. I got Truespace 1.0 on $99 into special, and I stuck with it. We
have struggled to get money to keep 3DS upgraded, and I drool over
IPAS routines on a regular basis. The sad fact of the matter is that
the US Government has decided to decrease the funding levels for
aeronautical developement and space technlogy by a significant
amount. (NASA has rumored 40% layoffs). Infact, we currently have a
computer moritorium at NASA Langley that severly restricts any
computer purchase. Therefore, I scrounge up what I can and make
do. Enough belly-aching. The bottom line is that Truespace does
what I need, even if it is not the most elegant solution. However, I
will take a copy of your letter to our next meeting. Thanks for the
amo!!
I thought the NASA has neverending money. Well, maybe we`ll have a
RayMan Light version very soon for, aehhmmm, maybe 348US$. That could
increase your chance to get a version…