Re: Even Lightwave gets in front of XSI for a Maxwell plug-in

["Matt Lind" <speye_21(at)hotmail.com>]

Interesting.

Personally I was never interested in seeing how long I could work on 
something for the sake of saying it took so long.  I was more interested in 
looking at something that was completed and figuring out how I could do the 
same or better with less time, fewer resources, and smoothest workflow. 
That's largely where I derive my techniques such as using nulls as deformers 
instead of shape animation.  It always has, and still does, irk me to see 
models built with gigantic operator stacks and excessive numbers of 
polygons, or textures, or animation with Fcurves that look like they've been 
plotted when in fact they've been keyed by hand.

I think the longer timelines back in the day made for better product because 
it forced you to think more about how you go about things as there was a 
smaller margin for error.  When a render took 3 days and the deadline was 4 
days away - kind of put priorities in perspective to get things right the 
first time.  My very first full length project using Softimage Creative 
Environment back in 1993 took me 5 weeks on an old Indigo to produce 3.5 
minutes of footage.  It was originally a stop-motion animation project, but 
due to lack of time I decided to give the computer a crack at it.  I had a 
bunch of classes using various 3D softwares on the Amiga and saw Softimage 
as a step up, so I took the leap.  My characters were simple teddy bears and 
roundish characters which had simple phong or lambert shaders and very basic 
raytraced shadows.  No reflections, soft shadows, no mental ray, no FX, but 
frames took from 4-15 minutes each to render at 640x480.  I learned real 
fast how to use sub region renders, the compositor, and other shortcuts to 
reduce workload.  While renders were going I'd sit by the machine and scour 
through the software manuals looking for every possible corner cutting 
maneuver I could find and learned how to use tools for purposes they weren't 
intended.  For kicks I exhumed that project last year and loaded up a few 
scenes.  On my Dell Precision Workstation 470 here at the home, those same 
frames rendered just short of realtime using mental ray.  My how far we've 
come.

One of the downsides of today's faster computers is people whip out trash so 
much faster and make, in my opinion, many more stupid mistakes because they 
don't have the patience to plan ahead or stop what they're doing for a few 
minutes to look at the progress (or lack of) and evaluate what truly needs 
to be done and why.  Many developers are in the same boat.  It's amazing how 
many people in this industry flounder or panic when the electricity goes out 
because they don't know what else to do with their time. ;-)


Matt

-------------------
Matt Lind
Animator / Technical Director
Softimage certified instructor:
   Softimage|3D
   Softimage|XSI
Matt.Lind(at)Mantom.net




Date: Fri, 06 Apr 2007
To: XSI(at)Softimage.COM
From: j.g.ponthieux(at)LaRC.NASA.GOV
Subject: Re: Even Lightwave gets in front of XSI for a Maxwell plug-in

This is an interesting take. Everything is relative I suppose. Normally I 
would tend to disagree with this assumption but I guess it really reveals my 
age and how long I've been doing 3D. 19 years ago when I set out to start 
rendering what would become my first broadcast piece, the render took in 
excess of two weeks. Mind you the scene had no HDRI, no GI, no Caustics, no 
Final Gathering, no raytracing, no shadows, no Phong shaders(we had Phong 
and raytracing but didnt dare use it unless we had a time machine), no Blinn 
shaders, no Anisotropic shaders, no textures from disk, no reflections, no 
refractions, oh yeah thats right I did say no raytracing.... sorry, no 
motion blur, no displacement, no bump maps, no area lights, no hair, no 
particles, no volume shaders, no well....no alot of things, you get the 
picture.

Basically everything was rendered in Gouraud(Lambert would be its general 
equivalent today I think) and I only had a 525x488 live framebuffer(not the 
XSI kind, a real honest to goodness hardware framebuffer) to store my 
texture maps ....uhh map with a bunch of tiny images painfully inserted 
between each other. All hot textures had to be stored in that space and this 
"global" texture had to be swapped for new textures when the old ones that 
had been used were no longer needed. There were render bugs galore, the 
frame would render in about 45 minutes to an hour and get sent immediately 
to this enormous VPR-80 tape machine. There was no render to disk, heck the 
disk was only 40MB, my ram was only 8MB and my cpu was a 386DX with a 24 bit 
"above board" graphics display and small monochrome menu monitor beside it. 
Pretty exciting huh!
Well, back then to say you managed to keep the computer, tape machine and 
render going for two weeks straight for something that took 3 months to 
model, 2 weeks to animate and another 2 weeks to render 24 hours a day at 
about 45 minutes a frame, yes that was something to be very proud of. You 
just completed something which few other people had ever done, and you 
wanted to do even more. Heck you really looked forward to rendering 
something in Phong! Wow that would be really kewl! Just, well....come back 
in several months, it might be finished, or maybe not. Whats really 
incredible is how much we looked forward too which was right at edge of our 
fingertips but still light years out of our reach. Features we had but 
didn't dare use because time prevented us from using it. Features which 
would have pushed the quality of our work to a new level not frequently seen 
in that day and time, but were so inefficient that their use would have 
prevented us from producing anything of worth in a timely manner.

The point I suppose I am getting at is that the industry has definitely 
changed. There was indeed a time when there was much to be proud of if you 
were able to model & texture something that would take days upon days to 
render. Even today I think there is a place for that, especially with all 
the rendering features which exist today which we did not have 19 years ago. 
Mind you that in general the goal for most artists today is to get the 
highest quality at the lowest render expense. That is indeed something to be 
proud of, but there is a time tested tradition in this industry of packing 
in as much as you can just for the sake of seeing how far you can push it, 
how much detail, how many parameters, shaders, etc, how much can improve the 
appearance, how many features do I have and how man features can I shove in 
there before I blow it up! An enormously long render may(or may not :)) be a 
testament to that success. In a production environment with deadlines it 
definitely may not be of much value, but strictly from an artistic viewpoint 
I think it still holds great merit. The trick is knowing where to separate 
the two.

Joey Ponthieux
NCI Information Systems Inc.
NASA Langley research Center
____________________________________________________________
Opinions stated here-in are strictly those of the author and
do not represent the opinions of NASA or any other party. 

---
Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in body:
unsubscribe xsi