|
||||||||
|
|
|
|
|
|
|
|
|
|
|
|
|
|||||
|
|
||||||||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|||||
|
| The Comet CG Frequently Asked Questions (F.A.Q.) |
In general a really good place to ask Generic MAX questions is http://support.discreet.com/ That site is an official Discreet MAX forum site, and there are a ton of people there who read it and help answer basic MAX questions.
If you are looking for artwork feedback, then for modeling there are places like the Spiraloid Forums or Cg-Talk links off my links page. There is also the CG-Char Web Forums which are great if you have a Character Animation you want critiqued. I do have my own forums for all of this stuff, but you will generally get better, faster, more answers/responses by posting at a larger site as mentioned.
With all that said...here is the FAQ.
| Animation and Rigging Questions |
Note too when you change the controller, you can also click the "Set Default" button which will let you keep Euler as a new default, so that all newly created objects will use Euler instead of TCB. See information in the Manuals on how to change controllers etc... (it can be done via the motion panel, or trackview).
Generally speaking, Euler_XYZ is far better for character animation than TCB.
The controllers I use so that I get 3 separate curves for X Y and Z are: Position_XYZ, Euler_XYZ and Scale_XYZ.
You can either find the controller range in Trackview, and drag out the range slider to work on a longer timeline, or download a free script from many of the MAX script websites, that will help you automatically readjust all controllers to work.
The best general solution is to Make a dummy or helper object, and use the "Align" tool to align the helper to the bone/object you want to have zeroed. Then put the helper into the same hierarchy where the real bone/object is. Then parent the bone object to that helper. Effectively you are "inserting" the helper in the middle of the hierarchy between the bones parent and the bone itself. The catch is the helper is first aligned to the same values as that child bone.
What this does is now cause the child bone to appear to be zero. This is because everything is in parent space if you remember. Therefore if you rotate now, it will in -trackview- appear to be from 0, and even if it was in some odd angle in world space, rotating on one axis should now only show one axis affected in trackview.
You may also wish to see my cartoony eye rigging tutorial on my site under my 3D Help Page. If you use Maya and not MAX, try the free "Rivet" script on the net which does the same thing.
For a more step-by-step directions that work on any package, check out Keith Lango's excellent animation tutorial on his site at http://www.keithlango.com/.
For MAX you can use scripts like keyManager on my site to easily change tangents for keys from stepped, to linear, to smooth etc... You should also make sure that the rotation controllers for the objects are set to Euler_XYZ and not TCB, so that you actually get fcurves and tangents to work with.
Pretty much you make sure you set a key on the parts you want ALL at the same frame. So if you rough it out and you decide you have poses on frames 1, 15 and 20, all the parts for the -body- that you might key in that case would be keyed, all at the same time. You can use Linear or Step keys too to help with the timing and blocking idea too as discussed on his site.
Make sure also that you are keying ALL of the parts/controls that you want to maintain at both the start and end of the hold range.
You may also want to investigate my "charRigger" script on my 3D Help page, which lets you automatically create an entire character rig, all from placing ptHelpers.
There is one other way tho...which is do it by hand using the Reactor Controller. You can set a morpher up by making it have a Reactor Controller on the morpher channel and driving the morph by hand from the rotation x y or z as you want, assuming u are using Euler XYZ rotations... The catch is that in many cases rotations may be in world space or be changing angles/axis...so you can't always count on it.
Worst case you make the targets and set up hand animated controls for the animator to use to fix by hand.
Typically I block out the horizontal hip movement for an entire walk in top view, then add a bounce, and then tweak it and add each foot position/key by hand for every step of the entire walk.
Just select the keys in the trackbar and use the "Affect OBJECT Keys". That will affect all selected keys for all the selected objects. If you do the first one, all, even if the trackview is closed, or tracks not selected, selected-KEYS will be adjusted. You have to be careful, because even if there is a "selected key" outside of the current time range, it will be adjusted.
The last mode, TV Keys is the most precise one, in that you have to have TV open, and highlight the track/controller and the keys to affect. The first one is more dangerous in that keys you may not even know are selected can still be affected.
Normally I use the regular OBJ key affect and use the trackbar as mentioned. Note tho, if you do forget to unhighlight a key, like you select on frame 20, then set the range to 50-100, if that key is still technically "selected" it will still be affected.
Beyond that, for rigging, if I have a cartoony eye setup as in my cartoony eye tutorial off of my 3dHelp page, then I use wiring to drive multiple morphpers. That is, typically you will have morpher on your head/face. In addition if my eyes are in an FFD lattice (the spacewarp, not the modifier), then I will not only create face morph targets, but also duplicate the FFD's and create FFD targets. Each object will have morpher. The face has morpher. The Left Eye FFD has morpher, and the Right Eye FFD has morpher. Then I just add the morph targets for each. In the end I will wire the face one, or make a separate control object with Custom Attributes, and just "wire" the spinner from one to the other.
So the main smile custom attribute is wired into the face morpher smile, the L-Eye FFD smile morpher target and the R-Eye FFD smile morpher target. That way I can animate just one value, and everything is animated together properly.
Another thing you can do is to keep things separate and just use Edit Poly|Attach function at the end to attach the teeth, gums etc all together. Just make sure you attach everything for the base and each target in the same order. ie: Attach teeth first, then gums etc... all exactly in the same order...so that point order doesn't change.
Alternatively you could setup a separate morpher on each one, and wire them together as I mentioned in the question above.
If you like to use bones, most people will use just 1 bone for the jaw bone, and then still use morph targets for the rest of the stuff. You just setup morpher below and then skin above in the stack and do your bone weighting.
Anyhow you can delete morpher targets later, and it should store only the changed point information, so the data kept isn't too large, even if you include full body/head seamless targets.
On the other hand, if you have a way to hide a seam, such as if the character is wearing a shirt, and you can make the head+neck object be a separate object from the body, then by all means, it makes sense to keep the head as separate.
I typically keep my entire body as separate as I can even arms/legs etc...if I can. Obviously though in some cases you need or want a seamless body.
If you MUST have a seamless body but really only want head targets, what you can do is take your base head, create morph target copies and add morpher onto the base head. At this point you just have the head and morpher working. Then add an Edit Mesh modifier above morpher on the base head. Then click the Edit Mesh "Attach" button and attach the body and then weld the vertices etc...using Edit Mesh. Now you have only one object. Then add Skin above the Edit mesh like normal, and then meshsmooth or whatever above that. You're effectively leaving the "Edit Mesh-Attach-And-Weld" in the stack above morpher and below skin, so that morpher and below you literally just have the head, and skin and above, it's a complete seamless mesh.
But skinning is one of those things that is difficult and just plain old takes time and effort. Don't forget to check out the MAX script sites linked off my pages too for tools that other people have written. Some of these save time by allowing you to mirror weights over, or save and reload weights, etc...
Beyond that for a walk etc, you would animate the hips, and then animate the feet to match. Same with the arms.
So effectively the hand will move a bit when chaning Local/World the two. When you are animating, first set the value how to want it to act for that frame, either locked or free, and then pose the hand. For example if a hand was about to push against a wall, and then lift back off say on frame 0 it is free, 10-20 locked to the wall and then 30 free again. You would first do a free pose on frame 0 with local mode. Then go to frame 10, set World via 100 and then pose the hand on the wall. Set an exact same key on frame 20. Then on frame 30, put local/world back to 0 and pose the hand away from the wall as desired.
You might also want to use the align tool to align some of the hidden dummy objects for the hand...that will allow you to snap the local dummy to the world, or vice versa, so that the local/world value causes it not to slide when changing the values.
| Education Questions |
In general the key is persistence. Don't expect going to school for 4 years is going to make you anywhere near good enough to get a job. Unless you work your butt off and teach yourself as well during that time, it probably won't. Expect to have to keep working and working and improving.
Mostly all you need is a really good demo reel on VHS tape and resume and cover letter to mail out to the places you want to work. Some places now accept DVD, but check first. Many still require VHS. If you don't follow each places explicit submission guidelines (and each is slightly different) expect to simply have your reel dropped in the trash. Places have far too many submissions to waste time with someone who can't follow directions listed on their website.
The main thing to do is read all you can and mostly spend as much time as you can practicing what you want to do. Build up a basic set of skills and understanding in each area from story to concept to boarding to layout to modeling and texturing, rigging, animation, lighting and fx and rendering. I mean basic...you don't have to be great, but you ought to know and be able to do a little of each. THEN after that, you'll probably have a feel for what area you want to focus on in that. Spend most of your time working on 1 or 2 areas. Most places want specialists. Now that may change over time and it does vary place to place. For example our modelers also texture. Some places tho, modelers only model and they have separate texture artists. Generally though the more you know and can do, the more valuable you are, but some places only have you do one thing. An animator at Pixar will only animate. All they want to see is animation on your reel if that is what you are applying for. They don't care if you can model or rig etc... Other places may be different.
If you have a C.S. background you may or may not want to focus on a more "Technical Director" type role, which might be more rigging or setup or progamming based. Like writing 3D tools or plugins or things like that. Writing hair or fx scripts or programs etc... In general it's easier to get into the job as a tech guy than an artist if you really are good at programming etc, since less people want that. It's also generally easier to get in as something other than an animator, since many many people want to be an animator.
But in the end do what you want. After 94 or so the industry was really booming. By 96 you could still practically walk into a studio with almost no experience and make tons of money. That's changed. At that time I was pure tech almost. I had C/C++, Perl, Unix etc skills. I also had done some decent modeling and texturing etc...but not by todays standards. I had several job offers from major studios to be a Tech Director/T.D.. which means doing some art work but mostly scripting or helping to do tech stuff. Exactly at that time I decided I wanted to be an animator so I turned down all those offers. I had to work several MORE years to get there but I eventually got in. Now it seems I'm heading back towards rigging and scripting since that's what I decided I really truly like.
But either way, your skills and VHS tape demo reel will HAVE to be really good. If you want a job as a modeler at a video game company...work on low poly models and textures. Make them as good as the best games out now. If you want to do models for a movie, make your models as good as the movie models. And be realistic when critiquing your work...are you REALLY that good? Is the work on your reel the level that would be done at the place you wan to be at? That's what it takes.
Max and Maya are VERY close. You ought to be able to pull off a reel that is as good as what you'd do in Maya in Max. Especially for modeling, texturing, lighting, I'd say Max is ahead. For rigging and setup they both offer very similar things, same for scripting. Though Maya is arguably better for setup and animation workflow. But when you get right down to it, animation is pretty much being able to pose and set a keyframe. And tweak fcurves. There's not much more to it. Almost any package will do that. There are guys at Pixar who were hired based on work done at home with Hash:Animation Master for $300. It's the skills/demo reel not the package. Many of us got started with programs before things like Meshsmooth/SubD's or I.K. even existed. We had to learn workarounds and techniques and skills to get around that. Toy Story was animated with out any IK at all! Chances are any program you have now will be way more powerful than what was out there 10 years ago. So you ought to be able to do something that much better, even if it is not the top top packages.
But remember the bar is that much higher now as well. We laugh at demo reels we get now that are better than reels we had when we got into the industry ten years ago. So pretty much practice practice practice and be persistent and don't give up.
Most students reels I have reviewed at work for people who have applied, get rejected. This is because the quality is simply not good enough, especially when there is a lot of competition and saturation in the industry now. What that means is, even after you graduate, you may still have to take time and continue to work on your reel at home, work on building your skills on your own time with your own system. If you have already been doing this, you already have the software and system. If you are coming out of a school, can't find a job, and don't have a home computer system, you are in trouble.
If you want to know what schools exist, I'd recommend taking a look at "Animation Magazine", there are usually many schools listed in it, and you can find it at most bookstores. So basically there are 2 divided camps on the subject. Some of the pros of school are, you may get to use software you can't afford, you get advice on traditional art principles, you may get placement/recruiting help, you will be around other people doing the same thing and can work together to improve.
Doug Kelly once gave me a great quote about students and teachers. He said a teachers job is to make sure the student has the opportunity and knowledge to learn. BUT, the students job is also to make sure they work really hard to learn as much as they can from the teacher. It works both ways.
Finally, check out my new Links Page which has links to other sites that have good resources for those looking for schools.
One important note: Only include traditional work with your reel/resume if it is REALLY GOOD. There is nothing worse you can do, than to have a really great 3D demo reel, but then ruin it by showing poor 2d skills. With both 3D and 2d, only show your best stuff!
My top 3 animation books would probably be:
"The Animators Survival Kit" By Richard Williams
"Disneys Illusion of Life" by Frank Thomas and Ollie Johnson
"The Animator's Workbook" by Tony White
In general, if you have a good strong traditional portfolio, bring it in. IF your drawings are good. If they are not good, then don't bring them.
Do the same thing with any "extra" material on the reel. If you have more material that is really good or at the same level of quality as what you showed earlier, then bring it. If it is not as good, leave it off.
In general if you have something that is good and great, then it's worth showing. Just remember you are judged by the worst piece that you show.
Regardless of all of this, you should always bring extra copies of your resume and at least one extra copy of your demo reel with you on an interview. Many times places will have lost your tape even if they call you in.
| Modeling, Texturing and Lighting Questions |
Typically I just start with a box, and just like modeling the character itself I model a solid piece of geometry for clothing. Then on the ends, like holes on sleeves, I extrude back in a bit to make it appear as if it is thin.
From a technical standpoint all methods can yield the same results, it's just a matter of how you get there, and how the tools work.
A background character or a game with a ton of characters may be more around 400-600 polys each. That's pretty normal. If you have a fighting game with just 2 characters on screen and some flats and simple background, it may be 2000+ polys. Typically tho game companies want to see demo reels with characters around the 500-1000 poly range. It'd probably be good to do a few in that range, like a 300-400 version, then a 500-600 version/different character and a 1100 version character. Or ASK before sending in a demo reel.
Summoner used 300-500 for very simple village people that were filler. The main characters and bosses had about 1100 when i was there, which included generally a somewhat hires face where a mouth was cut out for facial animation.
Cinematic Model Poly Count is of course higher, since it is all pre-rendered. For Freespace we had models that were quite high, it really taxed our system and MAX, and this was Max 1.2!!!! Our space pilots were about 20,000 polys for the body, and then another 20,000 for the head! Quite high. I'd say now most places would probably have 10,000 or so or maybe less, but use meshsmooth on top to smooth things out if needed. So in essence you get that detail, but the model itself is a medium res poly cage.
Obviously if you have something like Deep Paint or Texture Weapons, it would probably make more sense to do painting that way. Don't forget too you can also unwrap using the UVW Unwrap Modifier, and arrange polys how you want, just like you do for games. Getting seamless textures tho then can be difficult, which is why the cylindrical map for the head works pretty well.
The anime head itself was mapped mostly with a solid color, and only had a little detail here and there, so it was pretty easy to map and paint. Most of the color feeling comes from using warm lights and cool lights. Lighting and texturing are hard to separate.
For photoreal texturing, you can go and paste parts of real photos in photoshop and distort them to match the polys you are painting on.
If you imagine a triangle, each point is a light, pointing towards the center of the triangle. Tho usually the triangle is a little closer together for the key and bounce.
Rim
/\
/ \
/ c \
/ \
Key-------Bounce/Fill
The "c" is the center of interest, usually a character, or center of background etc....
Each light I make is usually a "spot" or "directional" type light, all pointing at the center of interest. In addition the triangle is usually rotated several degrees...so it's not flat on the bottom as show. This puts the rim not exactly behind the center, but to one side or the other. Of course once that starts it, you manually move and adjust each light as needed.
Vertically my Key light is usually placed above the center and aims down on it. The Bounce is placed below or at center and aims up. The rim is usually flat at center or maybe a b it up depending.
Color wise I'll start with a warmer slightly, but still near yellow/white key light. My bounce is usually a cooler blue or blue white and half intensity. The rim is always pure hot white, maybe even with a raised intensity to really cause the rim to pop and show up. The key and rim usually cast shadows while the fill will not.
That's my basic lighting setup. From there I'll tweak and add/remove lights as needed. One thing you mention is flat planes...this almost never works well where the planes touch, such as at the back where the ground touches the bottom of the sky plane. To solve this I'll usually model the ground to at the back it literally curls up. This gives a softer look to the horizon line. But that's mostly for simple shots of a character literally on a flat ground.
There is one other CRITICAL thing about lighting as well, especially for sets. The above works pretty well for just getting light. But for real sets it's best to think about light sources. And to actually place lights -simulating- where they might actually be. For example if you are in an office, then stick lights coming from the ceiling or windows like they would be. Even better, literally model the light object and such and it will look even more real. If you have a fire, again you'd probably have a yellow-orange light coming from that. Probably a point light.
So essentially, your scene should already have an idea of lighting. Use that. I'd still use fill lights, since light reflects. You can simulate color bouncing around etc... But figure out your keys, and where lights might bounce from, and that's about it.
| Personal Questions |
You might want to read my "About Me & How I got Started In 3D" page that gives more info about me and how I got into CG.
By the time I got into middle school I managed to snag a Super8 stop mo camera. At that time I loved Disney and WB stuff as well as some of the Will Vinton stop mo and such. I knew a little about Computer Graphics, and they were becoming more prevelant, especially for TV commercials.
Mostly I wanted to create pretty pictures. Things like the opening to "Amazing Stories", or the HBO logo even, Listerine commercials, Young Sherlock Holmes stained glass knight, etc...all that started me out. In many ways I was inspired by just CG in general. I loved computers, and loved to draw (or at least trying to draw).
As mentioned I loved cartoons, so as I progressed that is what I wanted to move towards. This is a lot of why I love Big Idea, it's one of the few places that does what I feel are really high quality CG cartoons all hand animated.
Certainly I also felt driven to help people etc...which is another part of what Big Idea is. I wasn't actually Christian when i started at Big Idea...and most of my conversion actually was do to my wife, then my girlfriend really introducing me to it, as her dad is a Pastor.
Artistically I've gone from more of a Salvadore Dali and Magritte type inspiration to more impressionist work like Monet. Not sure why but I love the colors.
Obviously Dinsey and especially Chuck Jones have inspired me a lot. What they both brough to the field is incredible. I also get inspired a lot by just other current artists, whether it's coworkers and their animations or models, like Keith Lango, or other artists and their still image work such as on raph.com.
I'm not sure how much religion has influenced me. Before I even started at Volition, I was doing a few oil and digital paintings. Interestingly most all of those had some religious tie in to them. So perhaps in a subconciouss way I was affected by that, but I think it ties more closely into my "fine art" works vs. my "entertainment art" works and shorts.
| Software Questions |
Another option, get cheaper software. This can actually be a good thing. Many people who have been doing 3D for a while, like myself, learned on old old old Amiga's and ray tracing programs. These pieces of software had NOWHERE near the bells and whistles of today packages. You might think this is a bad thing. But, consider this: It's a lot EASIER to learn something with less complexity. You might be better off with something a little less expensive but still good, you will also learn how to work around problems and technical limitations, something even pro's have to do during work. As a suggestion, Hash Animation Master at http://www.hash.com/ is pretty good for learning character animation and such in 3D, and is only a few hundred dollars. There are a LOT of online resources, and I have seen people do really nice work with it, and get hired at places like Pixar, Klasky Csupo and Big Idea. I've actually seen Dennis Murren from ILM standing out front of a Hash booth at Siggraph watching some of the guys stuff on the monitors there...
One other note for those who simply will pirate regardless. Do yourself a favor, and go to the bookstore, and spend the $50 or so on a good book. There are a lot of good books out there, especially for MAX. They basically are the manual re-written. You'll save yourself a lot of headaches.
Very VERY generally speaking, Lightwave is used at many TV/Commercial houses, 3D Studio MAX enjoys a lot of popularity in the video game industry, and Maya/Softimage is used mostly in the film industry. There is of course overlap.
| Tutorials and Files Questions |
If you are using MAX 4.0, then DO NOT USE THAT TUTORIAL. Instead use the tutorial in the manual which shows how to setup a much much better IK system with the newest ik types.
For example, if you are having problems with facial or body modeling, go over the basic polygon modeling type stuff in the manual and manual's tutorials. Also I'd suggest studying/obtaining basic anatomy and other traditional art type books. 3D modeling is still art, and it is still tedious and difficult. You might be able to get the basic model/points created, but tweaking is a big part of the work. Expect to spend about 80% of your time just adjusting points to get things to look right. Also, having a good anatomy book etc, will teach you where the muscles/bones are and where definition needs to be, so you know where to add more detail/points and how to do it. Experiment and try things out. Don't expect to be the best artist in the world right away. Practice practice practice.
As a note, the basic "Bug Modeling" tutorial in the MAX 3.1 manuals should help with the body modeling stuff I have on my site. In addition, there is a new tutorial on box modeling that I myself actually wrote that should come with MAX 4.0. That has much more detail on the specifics of what to do for meshsmooth type tutorials.