ALIENWARE Challenge Entry: Michael Crawford

[nuclearman]

http://www.cgnetworks.com/challenge/entries/3/1312/1312_1073523036_medium.jpg (http://www.cgnetworks.com/challenge/entries/3/1312/1312_1073523036_large.jpg)

A bit behind schedule, but that tends to happen when your workstation is 2000 miles away and all previous attempts to come up with a workable concept before Christmas netted nothing but a progress on the "learning curve." :-) Anyway, I will have approximately ten days from the time I set foot back in Kansas to when I have to have the final image posted, and I already have a pretty good idea of what I *can't* do, so ...

The general idea is this: thanks to some major scientific breakthroughs (and our own deteriorating ecosystem), mankind has embarked on an era of interstellar travel and manned planetary exploration. On an expedition to investigate a series of extremely promising SETI signals originating from a G-class star system in Orion, the crew of the Hawking discovers what appears to be an advanced technological civilization in the final stages of preparing to leave its home world. Though the crew has yet to find any life forms, the source of the mysterious signals -- and the presumed location of the populace --becomes evident when one of the planet's small moons comes into view ...

[nuclearman]

That's looking waaaaay darker than when it was in Photoshop. Hmmm ... let me post a "lightened" version and see whether that helps.

[nuclearman]

http://www.cgnetworks.com/challenge/entries/3/1312/1312_1073523693_medium.jpg (http://www.cgnetworks.com/challenge/entries/3/1312/1312_1073523693_large.jpg)

Hopefully bumping up the light will help the "web version" looker more like it does in Photoshop.

[nuclearman]

http://www.cgnetworks.com/challenge/entries/3/1312/1312_1073543136_medium.jpg (http://www.cgnetworks.com/challenge/entries/3/1312/1312_1073543136_large.jpg)

Just playing with concepts right now. Since I won't have access to my workstation for a few more days, I've had to content myself with working on textures for the planet and the background elements. When I actually construct the models, I'll aim for a more "organic" look, using some things I worked out in Dark Tree and Photoshop. For the "moon ship," I want to make the part in shadow look like those pictures of the earth at night: pinpoints of light indicating the community clusters on the ship.

[eska]

VERY nice and piecefull.
GOOD LUCK:beer: :wip: :bounce:

[nuclearman]

http://www.cgnetworks.com/challenge/entries/3/1312/1312_1073612711_medium.jpg (http://www.cgnetworks.com/challenge/entries/3/1312/1312_1073612711_large.jpg)

For the past two weeks I've been doing quite a bit of experimenting in Photoshop and Dark Tree, trying to see what I can come up with in the way of interesting "alien" textures. As a result, I've now got a library of about sixty custom textures that should help speed things along once I get back to Kansas and can get to work on the "heavy lifting." Questions, comments and suggestions are quite welcome. I've found this contest to be enormously helpful in pushing me farther along the learning curve, and gather most folks feel the same.

[nuclearman]

http://www.cgnetworks.com/challenge/entries/3/1312/1312_1073613025_medium.jpg (http://www.cgnetworks.com/challenge/entries/3/1312/1312_1073613025_large.jpg)

More of the aforementioned textures. Should be interesting once I combine them -- and fine-tune them -- with the bump and specular maps in XSI's Render Tree.

[BClaw]

Wow! *Cool!* :bounce: Moving a whole planet is a big task! :thumbsup: Can't wait for your final! :beer:

[nuclearman]

Thanks, Eska! Thanks, Burt!

At least this approach should be easier to pull off than the concept I was testing out in December, which kept crashing Mental Ray (despite 2.5 Gb of RAM!). This time, I'm going to steer clear of enormous displacement maps, horrendously complex geometry, and other indications of pure madness. :-)

(Though the notion of a nomadic alien race that performs tremendous feats of macroengineering has me thinking about doing something with warring mobile "city-states" -- think alien cities with the proclivities of "monster trucks" -- which should be just enough to crash the renderer YET AGAIN!!!)

[Defonten]

Nukclearman! :beer: This is really great man! Lights are just great indeed but forms of sctructures are too simple, don't you think? Try rethinking your form approach in this composition and I am sure it would be one of the best !!! Great work mate! keep posting updates !!!:thumbsup: :applause:

[nuclearman]

Defonten, I agree completely; expect radical improvement to the elements once I get back home to Kansas later today. :-) The thing is, I'm not used to doing anything like concept sketches -- I'm much more comfortable experimenting with forms in 3D and then going from there. But I figured since I'm out on the west coast right now and don't have access to anything other than an old PIII with a copy of Photoshop on it that I might as well spend my Christmas break working on textures ... and getting that dreaded "concept sketch" milestone out of the way. Unfortunately, I haven't had much experience painting with Photoshop, so I kept the "city" (and the "moon ship," for that matter) as simple as possible so I wouldn't end up with just a mess. :-) The sketch is only intended to give a sense of the elements (moon, a bit of landscape, foreground city) and atmosphere; the "real work" starts tonight!

[nuclearman]

http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074057412_medium.jpg (http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074057412_large.jpg)

Ever since I found out about this contest (back in mid-November), I've been pushing myself to see how much I could learn in such a short span of time. Most of the experimenting I've done in 3D has been with shaders and particles and a little (VERY) basic modeling, so I've found it both heartening and inspiring to see all the folks who've pushed themselves in the quest for the "Wow!" factor. :-) Anyway, it seems like half my time has been spent trying to figure out ways NOT to crash my app -- I mean, you'd think that 2.5 Gb of RAM would be enough (or at least *I* did); but I guess that's what they mean by "steep learning curve." :-) All of which brings me to the simple image above: all part of a day spent testing out HDRI in order to get lighting similar to the concept art. The image is also a test of using a little light displacement (courtesy of the intensity node) to produce a more interesting -- and "alien" surface on this tower. (The tower is intended to be part of the foreground city shown in the concept art.) The HDRI sunset is courtesy of Dosch Design -- the first time I've tried any light probes beyond Debevec's collection. Hopefully this is the beginning of some rapid progress.

[nuclearman]

http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074058486_medium.jpg (http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074058486_large.jpg)

Yes, I *know* this doesn't remotely resemble the other concept art, but I found this image to be a good test of whether I could repurpose some of the elements from earlier test renders to come up with a more interesting foreground. Scattered throughout this image are a good dozen or so "dead ends" -- eg, plant life with such horrendously high polygon counts that I just *knew* I was doing something wrong (that "learning curve" again), or an attempt from late November to see whether 3.5.1 could manage Booleans any better than than the previous incarnations of XSI (answer: no). So ... just thought I post this before spending the rest of the night working on more of the alien city.

Questions, comments and suggestions are always welcome.

[vrf]

I like the rendering of the foreground. Looks like a photo. Add a funky alien color to that foreground shader (instead of the earthen brown) and you'd have one nice foreground. (And I like how you stuck to the "rule of thirds" on the composition.)

Oh, and the shallow depth of field on the foreground makes the image look like a miniature. You might consider getting rid of the depth of field effect. "Vast" landscapes, often taken with wide-angle lenses, usually have very deep depth of field, almost so that everything is (or appears to be) in focus.

[nuclearman]

Thanks for the advice on dof, vrf. I think I'm going to have to get Frischluft's dof plugin for Combustion, since the dof shader in XSI is awfully slow and not terribly good. I tried to fake it in Photoshop when I put together the composite, but I really need to read up on just *how* it works since I don't have any photographic experience to draw upon. If you can suggest any good books on the subject, I'd be quite grateful; the XSI manual really doesn't provide much info on the subject.

[nuclearman]

http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074155891_medium.jpg (http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074155891_large.jpg)

After all sorts of attempts to generate an interesting and effective sky (shaders, HDRI, Photoshop), I've ended up going back to something I developed back in December. I was playing around with the idea of doing an alien power station, and thought I might use the "gas flame" shader to good effect for that. I was also trying out a vbs script called "shatter" (based on a tool in Maya, I gather), seeing what I could use it for, and came up with the following effect, which remained pretty much stillborn until I purchased a copy of Combustion.

[nuclearman]

http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074156417_medium.jpg (http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074156417_large.jpg)

The "gas flame" shader produces a pretty nifty effect, but it has one major drawback: put anything behind the object (other than a really dark background) and the transparency mucks up the nice clean glow. There are probably solutions for this, but I'm still not far enough along on the learning curve to have a good sense of efficient workarounds. So ... I ended up with another dead end.

[nuclearman]

http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074157031_medium.jpg (http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074157031_large.jpg)

'Course, one of the many, many very nice things about this contest is that a neophyte like myself can learn gobs about what approaches (or apps) can help achieve certain effects. And since XSI doesn't have particularly good tools for getting glow effects, I was delighted to learn about the tremendous usefulness of Combustion and Trapcode's "Shine" plugin in Steve Green's excellent thread. As a result, fifteen minutes after installing the aforementioned tools on my workstation, I'd gone from a fairly cartoonish and flat output to a pretty neat effect that I've got a use for in another project. :-) Anyway, the image has been kicking around in my contest folder for about a month, and it wasn't until this afternoon, when I was working on the ground elements (for the umpteenth time) that I suddenly saw a use for it.

[nuclearman]

http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074160130_medium.jpg (http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074160130_large.jpg)

Hard to separate "modeling" from "texturing" in some areas. After days and days of using grayscale maps to generate displacement terrains and working with elaborate procedural shader networks to see whether I could get a basic landscape I liked, I finally started to make more satisfactory progress after switching to custom texture maps and low levels of displacement. Since I was up in Seattle and away from my workstation, I spent Christmas break doing what I could for the contest: churning out dozens of 2000 x 2000 texture maps while searching for interesting effects. Glad I did that, since now I have a MUCH better idea of the range of effects that are possible if you work with high-resolution maps. Yippee! Progress comes to Kansas!

[nuclearman]

http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074160899_medium.jpg (http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074160899_large.jpg)

A screen grab of the shader network for the ground effect.

[vrf]

Thanks for the post on my thread, Michael. As for resources that specifically cover DOF, I'm not really sure of any, other than general photography resources. Photo.net is a pretty good place for free tutorials on understanding how cameras, light and lenses work. I think a google search would probably dig somethign up as well. You might also consider Jeremy Birn's rendering and lighting book. It has several pretty good explanations for understanding the intersection of CG and photography. (And personally, I think that photography is one of the more valuable "traditional" art mediums to have in your belt when learning 3d.)

Good luck on the rest of your image. Time is ticking away!

Vance

[nuclearman]

http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074161636_medium.jpg (http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074161636_large.jpg)

Once I'd FINALLY put together a ground that felt ... well, alien ... a lot of other things started to follow from that. I can still see working in elements of the alien city into this, and some stuff from the old "power generator" concept may still relate to the glow effect in the sky ... but I can see this is starting to veer in another direction. *Sigh.* Maybe in a few years my skills will be sufficient to be able to follow a straight line from concept to completion, but at the moment I get too frustrated when things go wrong, so I'm just trying to find a path of "least resistance." At least this contest is teaching me what areas I need to practice, practice, practice. :-)

[nuclearman]

Vance, many thanks for the info on DOF. I will certainly dip into Photo.net after the conclusion of the contest, since I'm going to (finally!) be getting a camera later this month. Kept putting off the purchase hoping that the prices would come down -- and the resolution go up -- on digital cameras, and now that Sony has introduced an 8-megapixel model for under $1000 I can afford to splurge without completely using up my equipment budget. Anyway, the current contest has really whetted my appetite for digging deeper in the software and concepts so I have a better idea of how to achieve what I envision. :-)

And, yes, I'm quite aware of how little time is left in the contest. *Gulp.* 'Course for me the goal is just to complete the milestones and put together an image that doesn't look embarassingly bad. The problem, in this case, is certainly NOT the tools; just my lack of familiarity with them. :-) I guess I'll have to do something about that.

Good luck with your own image!

[nuclearman]

http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074321795_medium.jpg (http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074321795_large.jpg)

Well, I realize that time is fast running out and I need to sort out my foreground asap, so here's a start. This is a modeling shot for something I think of as an alien "tree," although the real inspiration was the dreaded "corpse plant," known for its horrific scent when it blooms. Of course, if lifeforms are *strictly* prohibited, then it's a strange rock formation, or an alien duster, or ... heck, it's ALIEN! Figure out what it is fer yerself. ;-) I'll post the test render in a moment.

[nuclearman]

http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074322072_medium.jpg (http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074322072_large.jpg)

A test render to see how the displacement maps, textures maps, and the volumetric hair interact.

[nuclearman]

http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074332270_medium.jpg (http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074332270_large.jpg)

Playing around with the foreground again. After trying out a bunch of different textures, displacement settings, and shader networks for the foreground, I finally got a promising result by using a 2000 x 2000 texture map I call "Organica." I set the bump mapping to zero, connected an intensity node to the "shiny" socket, and then fiddled with the hue in the picture editor. The result had none of the artifacts that were marring other attempts. (I swear, some high-res texture maps that look great from above suddenly look dreadful when the angle becomes oblique. I could probably solve the problem by going to a higher resolution -- say 8000 x 8000 -- but that would take quite a bit of time to recreate the textures for that res.)

Anyway, I then took a 960 x 720 render of the foreground (minus the plants and other stuff that will go on it) into Combustion and started playing with the Trapcode plugins, seeing what I could do with some radical post work. The results were interesting ... but lacked a sufficiently layered look to seem an improvement on my previous "ground" renders.

However, a quick trip into Photoshop solved THAT problem, with the net result that I now have a promising element that seems to evoke the contest line, "...it seems that reality here is somewhat obsure." As for the rest of the contest scenario ... at least I've got three more days to do something about that. :-)

[muyyatin]

Awesome alien tree.... very cool. I think the new ground has a lot of promise, maybe it just needs shadows w/ the displacement, maybe a little more variation in the geometry... whatever. Definitely an awesome tree though, I love the main shader with displacement and everything!

muyyatin

[monsitj]

very great tree !! I like the shading ,it's very high quality , but about the floor , i think it look a little tile ,
nice work , can't wait to see the whole thing come together:thumbsup:

[nuclearman]

Thanks, Jon. XSI (w/ its mental ray implementation) can do wonders with close-up, highly detailed work. I'd have to check the shader networks on the tree, but I think both the trunk and the pine-cone thingie have two or three 11 Mb texture maps working together to produce the effect. What XSI has *horrendous* trouble with is texture application to large surfaces -- like a landscape plain. I can get "clean" (if rather monotonous) results by using an incidence node, but most of the procedural shaders just don't fare well when cranking through a lot of space. So ... I could break up the ground into a lot of little pieces and attack the problem that way, but that's horrendously time-consuming -- even on a powerful system -- and I think part of the real value of this contest for me (other than meeting more experienced artists, being inspired by their work, and learning LOADS about new approaches and techniques from all these extremely helpful people) has been in forcing me to try to solve problems I hadn't encountered before. Most of the things I've tested out in XSI have been small, abstract pieces and individual "product shots" -- not massive scenes. So this has been a GREAT learning experience, even if it's cost me a lot of sleep. ;-)

Anyway, thanks for the suggestions concering the ground. I'll try that out sometime today -- probably using drop shadows in Photoshop, since it's pretty much all done in post. *IF* I manage to work up foreground elements (like the tree) quickly enough, I'll try interspersing little "islands" of differently textured ground in the overall plain -- or just let the foreground stuff dominate (which would certainly play to the strengths of mental ray), even if that impacts any sense of "grand scale."

*Sigh* ... if only I'd known about this contest sooner, or had access to my workstation when I was in Seattle for three weeks, I might be further on the road to doing a decent image.

Live and learn ... :-)

[nuclearman]

Thanks, Monsit. Still struggling with the concept of "ground," so I appreciate your concerns. I'll try out a number of different approaches today and see if I can come up with something better. I can see a basic "disconnect" between the quality of a render like the tree, and the approaches I've tried for the ground, but haven't solved the problem yet. :-(

[nuclearman]

http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074404123_medium.jpg (http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074404123_large.jpg)

"Situation hopeless ... but not serious."

That about sums up the state of affairs here. Can't say I'm thrilled about it -- I was hoping to be able to put together a decent composition by the end of the contest -- but I've learned sooooo much in the past two months, and have a much clearer sense of what I need to work on and learn in the months (and years) to come.

As things currently stand, I've managed to develop some potentially interesting textures, models and effects, but have encountered so many problems trying to get things to work TOGETHER that I'm back to compositing rendered pieces in Photoshop, just trying to get a sense of which models might be pulled together into a coherent scene and then rendered as a single image to try to get the lighting to match.

So ... I've spent the evening doing that, and have to say that nothing really grabs me. I like the look of the alien trees, but they don't mesh well with any background I've come up with. And every attempt to produce a bleak, almost empty landscape (easy to conceive of when you live in Kansas) just screams "lack of imagination."

I'll post three compositional attempts and then get back to work on putting together the full 3D foreground for the final shot -- the one that shows the most promise (and at gives me at least a TINY chance I'll actually complete it by the deadline).

[nuclearman]

http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074404343_medium.jpg (http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074404343_large.jpg)

This approach seemed to work a little better than the last, but still didn't seem like the right direction.

[nuclearman]

http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074404819_medium.jpg (http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074404819_large.jpg)

A slight improvement, but still not working. *Sigh.* I feel reasonably confident I can pull together the foreground elements, but since the "grand alien phenomenon" in the background is mainly post, and simply *can't* be lit like the foreground stuff ... well, that seems to completely screw up my attempts to make it feel "unified."

Maybe six months from now I'll know how to do this sort of thing. At the moment, it's beyond my capabilities.

[BClaw]

I really like your entry! :bounce: That alien tree is fantastic! The ground material on the plane is bizarre and very cool! That sun looks like it's shedding an outer layer or two, and I like the arched gate, it's very well done also. :bowdown: IMHO the only thing it might need is something out on the plain. :blush: (I'm a bit embarassed to suggest anything, you're such a fantastic artist, and have given me some very good insights!)

Good luck on the final entry! :thumbsup:

[nuclearman]

http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074499210_medium.jpg (http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074499210_large.jpg)

Spent the day trying to develop some more compelling models for the foreground. Wasted a over four hours on a single alien plant that was a little too complex for its own good. Live and learn. Anyway, thought I'd post a couple more renders of stuff; I had to check the lighting to make sure the models would hold up under the conditions in the scene. So far so good, but I'm waaaaay behind where I'd hoped to be a month ago. Aaaaargh!!!

[nuclearman]

http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074499424_medium.jpg (http://www.cgnetworks.com/challenge/entries/3/1312/1312_1074499424_large.jpg)

I now have a library of foreground models and will spend the wee hours getting them into position so that I can do a full-size render on Monday. Assembling the final image and doing the post work will go quickly, since I've already saved the settings for those operations.

[nuclearman]

Thanks for your kind words, Burt, but I'm _anything_ but accomplished when it comes to working in 3D. I've got about five years of experience with Photoshop, dating back to 4.0, but I've only started seriously trying to learn the high-end apps in the past few months. Before that, I'd done a bit of experimenting with Softimage 3D/Extreme (the Phoenix Tools shaders were wonderful, and it's just a shame they haven't updated them to work with mental ray 3.2) and now and then dipped into the different incarnations of XSI -- but nothing really systematic; I was just trying to get a sense of what they can do. This is the first time I've ever tried tackling something like a big scene, and I feel like I've tried to cram about six months worth of learning into just a few weeks (with the net result that I haven't had a full eight hours of sleep in over a month!). Anyway, good luck with your own entry. I'm glad I was able to pass on a few useful pointers, though the best I can offer are just general design tips -- nothing in the way of 3D knowledge.

As far as the background goes, I tried doing some renders that incorporate models into that colorful plain, but since the look of the plain is the result of a couple of Trapcode plugins and a bit of tweaking in Photoshop -- essentially a post effect -- I wasn't able to make them look truly integrated. I may give it another shot tomorrow, but I'm fast running out of time. If you have any suggestions, please don't hesitate to pass 'em along; you never know when some observation might set the tumblers spinning and unlock a solution that has long evaded me.

Thanks again!

[BClaw]

Nuke! I'm sorry to hear your computer freaked on the render! :eek: I had a lot of trouble with my computer crashing while I was trying to get things going too. :banghead: I didn't look too closely at all of the posts in your thread... :blush: Did someone suggest rendering in layers? That's what I ended up doing, in six layers! So, that way the computer only had to grind through a relatively few elements and a background image. :arteest: The disadvantage was that to make a change on layer 2 (which I ended up doing), I had to rerender layers 3-6. (sigh). I though about rendering with an alpha channel and trying to bring them together later in Photoshop, but I don't know Photoshop very well, so I was scared to try it. Anyway, maybe a new thought for the future?

[nuclearman]

Burt, believe me, if there'd been more time I would've gone through the manuals and figured out how to properly set up pass partitions; unfortunately, I'm so used to doing really small scenes -- digital product shots and experimental shader tests -- that I've never had cause to worry about that so of thing. As a result, I kept trying to do it all as a single pass until it was too late. :-(

So now I know better. But sometimes it's hard to see the forest for the trees: I swear I spent the better part of a month just trying to figure out how to get a workable "landscape" -- a bit of continuous topography that looked as good up close as it did in the distance.

In Softimage 3D/Extreme, which was bundled with the first version of XSI back in 2000, there were some very nice environment shaders which would basically allow you to just concentrate on the foreground of an image. Sun, stars, clouds, atmospheric dust and a realistic horizon line were all included in a single shader. When I started working on the first draft of my scene for the contest -- back in early November -- I tried to make use of these so-called "legacy" shaders in XSI, but they're all written for mental ray 2.1 and don't work nearly as well in mental ray 3.2, making it tough to get good results. So I changed my approach ... got nowhere ... spent three weeks in Seattle working on large texture maps ... and then arrived home in Kansas with only ten days until the deadline, and no solid idea of how to put together something spectacular enough to hold its own with the top 10% of work on display. I kept aiming for radical effects, rather than solid modeling, and only came to the conclusion that the one approach that played to mental ray's strengths was to put together a densely-packed photorealistic foreground and hope that sheer rendering quality would make the piece stand out. :-)

Well, here are a few important discoveries I made during that frantic last week of work:

1) You can get great results with the hair in XSI, but moving around a model (like the alien trees) that already has hair applied can slow even the most powerful computers to a crawl.

2) The displacement grid script (available on Ed Harriss's great XSI site) can be used to produce workable landscapes from grayscale images, but you only get really good results if you use sub-ds on the initial displacement, and that tends to push the polycount to around 35% of the mental ray limit, making it imperative that the other objects in the scene have fairly light geometry (which, unfortunately, is counter to how I normally work).

3) Procedural textures -- even the enormously flexible textures in XSI -- are not an efficient way of texturing large landscapes (not if you want anything that looks realistic!). Instead, texture maps are the way to go, but be prepared to build HUGE and highly-detailed maps (4000 x 4000 seems to work pretty well) otherwise you'll get lots of distortion and pixel-stretching with oblique camera angles.

4) HDRI is great for lighting an "intimate" view of some objects place at the center of a small scene, but doesn't work very well as a panoramic backdrop for large scenes. (Case in point: the Dosch collection, which has visible noise when used as a backdrop.)

5) XSI may have a great set of polygon modeling tools, but that approach leads to MADNESS when attempting to do large numbers of organic forms. (Each sub-d level ups the polygone count four-fold, making it hard to achieve the smooth, organic look when you have lots of vines or roots or whatever in the scene.)

Well, I'm sure none of this qualifies as news to veteran 3D artists, but these were certainly valuable lessons for me.

BTW, Burt, if you render a single object (or set of objects) in Targa format, the alpha channel is built-in and EASILY accessible in Photoshop. Just look under "channels" (in the same palette as the "layers" tool) and you'll find a black-and-white alpha channel listed below the R, G and B layers. It can be used to easily make a quick clipping path, though a good compositing tool like After Effects or Combustion are the ideal choice for handling alpha transparencies

[BClaw]

Thanks for the detailed explanation, Nuclear! I use 3D studio, so a lot of what you're saying sails right over my head. I'm not even sure if I can create procedural textures in 3D MAX (but I *think* you can!) I'll have to experiment off-line a bit and see if I can post-composite stuff successfully. Woulda saved me a *lot* of time!

Good luck in future challenges! :thumbsup:

BTW - You should finish up and submit *somewhere*... :bowdown:

[CGTalk Moderation]

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.