This course will look at the fundamentals of rendering in Arnold. We'll go through the different light types available, cameras, shaders, Arnold's render settings and finally how to split an image into render passes (AOV's), before we then reassemble it i
Yeah retopo, dismaps and Normal Mapping. Its a mine field mate. Everyone has different opinions on all 3.
Topology: Always a fun topic, I love it as I've always thought myself of being a hardcore modeller, even more so now I am in the industry. I think most people with tell you these days, you gotta have quads (moreso in film) and your edgeloops have gotta be 'there', without edgeloops people may aswell forget good deformation, particularly in a Face/head model. If you look at a good mesh, and the artist knows his or her stuff then the loops will near enough match the direction of the muscles that lay beneath the surface of the skin. Its great to sculpt from a blank sphere or cube in ZBrush but you dont get good loops from the offset, and in my opinion doing something again to create the proper geometry is a waste of time and effort when you could have done it the first time round with an importedmesh from Maya,soft etc....
Dismaps: Yeah great as long as you can get them to work in Mental Ray properly, if you want them to work in a superior way without arsing about, then renderman is the answer. Failing that theres always Cyslice. Just import the low rez mesh and the hi rez one from Zbrush and it will work out the difference between the two and produce a stunning map for little or next to no effort.
Having recently read that the characters in Avatar had no displacements whatsoever because of dissolve issues between the different expressions I can actually understand that - why put yourself and the team thru so much pain in order to get it right when the info can be built on the mesh already without resorting to image maps at rendertime....afterall a few years back what the hell did we do?? We did it the very same way and applied a damn good bump map!
Normal maps: AKA the Super Bump Map! great if you are into realtime stuff or just wanna add extra kick to your models without the aid of Dismaps pushing the Geo out left right and centre and without the long render times. According to a Siggraph paper this year Normal Maps can be used in a specific way to now drive Subsurface Scattering in a skin shader, again all based off the normals....
Overall though I would use try and use these methods and see what feels best, always try and think about the end result, whats it for etc...what will benefit me
remember too that alot of work (as good as they are) on the forums are really highly detailed just to show it off, none really have a workable factor built in for animation etc and most end up in portfolios 'posed to sell'.
At the end of the day the knowledge gained in these areas will help you move forward and make the vital decisions when needed as Ive mentioned above
When I first started learning 3D i use to hear and read that alot, the idea that you use whats necessary at the time, small texture space if in the background, LOD etc, it makes more and more sense to me all the time.
Its comforting to know that disp maps are regularly an issue. I shall google at some point renderman and cyslice. I have reverted to going for normal maps generally as they seem to be a good trade off between security and quality.
I know if I am going to continue 3d I need to understand edge-looping and topolgy much better than i do.Its one of those areas Ive not singled out and studied with any real effort yet.
Just seeing your work in this thread makes me see the benefits of a good imported mesh, as opposed to sculpting from afresh in Z.
Cool, thats good. Ive jus been trying to grill CTBRAM on his Viper thread because of his use with Boolean ops in Maya and the mess they tend to make. So yeah a good mesh output regardless of hardsurface or organic is important, especially if going into a sculpting package.
Yeah normal maps are a good trade off for now, its all swings and roundabouts.
I don't know bro, it might be me, but i think the eyes look a bit off......Just....something about them doesn't quite seem......right. ....LoL
[edit]N/m that part of this post about his skin, i look at it just wakign up and i see it doenst look scaley at all, amazing what sleep can do for you[edit]
Also, will this cat ever receive a body? or is he someones trophy on a wall? or....Is he just going to have a lot of mechanical legs pop out of his neck to walk around on like John Malkovitch's waist in Hitchhikers guide to the galaxy?
damn that jaw for no apparent reason i just feel like tugging at it anyways its looking great jay been wondering though when will we start to see some animated work from you its only the next logical step
SomeGuy: I actually have stubble lined up for him via shave and haircut.
G-Man: I think the subtleties of the skin will come thru onnce the subsurface shader is on it. He's just a manky alien dude that will be running a hotel in my Outlaws Comic, and yeah there'll be a body..
Murambi: Animation? Me? No mate, though he will be rigged and poseable, but I wont animate him, it would look awful
Jay, i was wondering.
With your work flow, at least for your personal projects, How often do you have to go back and tweak the mesh of your character models after posing them.
I remember somewhere in some post, you were talking about people having to do that sometimes to make the model look right for the scene it was in, is that a normal thing for organic modeling or more of a project / scene specific thing? mind to enlighten me on that aspect of organics some my friend?
G-man
Generally tweaks on actual personal meshes I dont as such. When Ive rigged my own stuff I know its limits and because Im doing stills and not animation I can get in there and simply paint it out in Zbrush or even in PS after the renders have popped.
But When it comes down to rigs and stuff in production, theres a zillion and one things that are going on, theres blendshapes, muscles, cloth, and even displacements etc......plus theres more than one dude involved in all the processes and everyone has their say. Generally the model meshes are fine but whether they fit a camera is another matter. Sometimes refernces can be misleading as opposed to a final shot
At some stage any of this stuff can just explode in a rig depending how far they are pushed....its not all perfect. So sometimes in specific areas something may not work - period, so its necessary for any team to evaluate the entire thing. This means start with the modeller and finishing at the final comp, its a matter of whittling it down until the prob is found and then you tweak it...then go again and so on until it works.
it does,
if i understood correctly, to sum it up, a perfectly good model, in a production, can under the correct (or wrong as it were) circumstances, completely and fall to pieces. Also while this were happening, it is not necessarily the fault of anyone, but only limitations of the amount of different forms of 'art' thrown upon the poor mesh in question?
Would this be the main driver of Weta, ILM, ect. producing their own software / plug- in a lot of times for this movie, or that to solve some problem or another they are having?
Sometimes stuff straight out of the box simply wont stand up.
This is really why maya and most of the apps are the way they are, they can be re-written under the hood. Im using stuff here at work just for modelling that kicks maya's standard stuff into touch.
In that case, here's the kicker then old friend.
and i ask this here, hopefully so that it will help others as well.
In my case, you know my skill sets, and my forte, what i am good and and what i am not so good with, at least i would think you do after this long anyway. Granted, i know you or no one else, really has seen anything new form me in a while unless you've gandered at that environment i am working on based on a Feng concept.
That being said, and hopefully being correct, What would you offer me (or others) as advice, being a organics guru such as you are, on my next step to actually being able to model something of the flesh?
I think you've heard the complaint form me before " i could seem to get the part after 'lets set up reference images' to click". To elaborate on that, for me it was getting the edge loops for eyes and mouth and nose right, and being able to play them in 3d space so that they looked like more then lumpy dumplings.
(sorry for hijacking your thread bro, i just think it fit with this one semi ok )
You may not post new threads |
You may not post replies |
You may not post attachments |
You may not edit your posts |
BB code is On |
Smilies are On |
[IMG] code is On |
HTML code is Off