rebo…HEY, don’t spoil the surprises!!! 
Jinian go there: ZbrushCentral
Julez: Actually the mouth DOES have enough detail, only towards the cheeks it’s a little thin, yet, it would only require a bit of clever management to fulfill practically any desire for the animation…if I find time, I’ll show you what I mean…
BTW, I prepared the messiah post: Technique behind the Neckling
really nice work. i love when someone does something different with available tools, makes us all think “why didn’t i think to do it that way”.
i would love to see this method brought into the game industry by using Normal Mapping. sure, he uses 4000x4000 textures, but using game friendly map size i’m sure we can get the same end result [on a smaller scale].
if a game engine was made to process more textures then polys, we might have something.
Man! That’s crazy bro. I gotta add ZBrush to my toolset.
Tech like that is ground breaking. So real how the skin moves.
Yes, in fact, more so I am going to share this entire scene/project with the release of messiah:Studio2.0c and the TextureDeform plugin! An entire tutorial will follow up…I might just make a DVD with an even more evolved workflow that goes a bit further than that.
Taron333
I have never heard of animated displacment maps working like this. Nice job. Here is a question for you. Can you do animated normal maps and have this high detail in video games??
Hey dansgarbage, well, in theory of course it is possible, whereby this depends on the kind of access the hardware allows. I’m not sure about the amount of control from the software side when it comes to the normal mapping so I rather don’t make too many assumptions at this point.
However, in theory it is no problem whatsoever and if I had any way to support what’s happening inside gfx cards, I am certain it would be rather easy for me to implement that sort of function. It is practically a mapping type that can be interacted with. The routine to remap space as I do it in the textureDeform is just a tiny bit more tricky, because it can not only rotate or stretch space but also twist it from origin to the target of the bone. But still extremely quick to process. 
As for subdivisions and also texture resolution…well…those are the two things that obviously limit the possibilites for the brutal amount of detail our eyes begin to demand eversince ZBrush…hehehe…but it could nontheless become highly impressive with the given limitations if we see the type of motion underneath skin in an interactive environment…that’ll be fun!
Thanx for asking, however…I’d like to think about it and see if there’s anything I could even do myself to find out…not sure. We’ll see…
GREAT work taron !!
and the texturedeform on messiah is very Impressive but (how many bones just for neck !)
keep up the good work
Thanx! On this one I goofed around a little bit to test out performance and handling and, well, just to have some fun of curiosity with it so I ended up with…let me see…9 on each side and one in the middle, but since they were just for the Shading and the actual routines behind the TextureDeform are fast as can be, it was no real extra effort for the rendering times. I also paid great attention to avoid any extra computation for anything out of reach to any bone, although I havn’t done any particular sorting to clip even the check for possible proximity of the current rendered spot to each bone. In a future release I may do that, too, but the rendertimes are really extremely satisfying right now. To put it like this…if you add about 10 bones and use them within reason (not everyone covering the entire screen so to say…highly unlikely), it’s the equivalent of a single UVtexture readout. Also depending on the kind of use you’re looking for, it can be even more efficient. The TextureDeform feels very much like animatable brushstrokes in ZBrush and feel practically like an extension or continuation of the work from there.
However, it’s really easy to get carried away with such setups, because it’s really fun to think of interesting constructions and curious deformations…but again, the great thing is, it really doesn’t hurt!
Damn it,Taron; you’re at it again :rolleyes:
Beautiful ,beautiful work! Really love this head’s unusual expressions…did you store different MT’s for this one? Looks great - how are you sir BTW?
Atwooki
I just made one test morph with the lips. I’m thinking about actually making a little system for it to get some lipsync done, but not sure yet…I may just move on to the next step altogether. I already have a next step for this one that really a cool surprise…just did something a few hours ago…I’m rendering a nice radiosity clip of it now, but it takes 3minutes20sec per frame…60 frames…yeah…it’ll be a moment…
THANX!!!
The biggest limitation I think is the relatively small texture cache, where all textures go while they’re being used. It seems this isn’t going to get much larger in the next-gen hardware either. Normal maps can be compressed, but I’m not sure if they stay compressed in the cache.
Hardware-accelerated displacement and tesselation methods are becoming more common, so I think we’ll see more of this in real-time apps like games. In that link, click on the grey alien picture for a similar methodology to the Neckling. Memory is still an issue though. But displacement maps can be palettized or use cache-friendly compression like DXT, so that’s an advantage, and they also change silhouette and self-shadow properly, things that normal maps can’t do.
We’re working on some real-time technology that blends displacement mapping, and allows you to deform meshes per-pixel, like if you shoot a character in the arm, you see the flesh pucker inwards exposing flesh. Eventually with more hits you see bone. The surface is tesselated on the fly around the wound, so you get all the gory details. Cool stuff, can’t wait to show it. If you go to our video page, you’ll see a snake monster… he’s blending his whole body in the video, but will be per-vertex based on collisions soon, then per-pixel after that. One step at a time.
Neato stuff Taron, thanks for sharing the how and the why. Looks very promising.
I just saw the character animated
it’s simply amazing these are the best morphs i saw
but i really like critisizing so i still think you can make the area between the uper lip and the nose and the nostrils move a little more.
HOLY CRAP!
This is something I never thought possible
I have 3ds max,now modeling tools is a time waster
A good reason to check messiah studio
I just shit my pants:eek:
How long did you work on this? Is there any way you can post the actual displacement map?
JHope
Good job…
Hey mister Taron!
Always obscede by the time of rendering!
:rolleyes:
This is really an amazing face!
I really love the real sensation of muscle behind the flesh.
Is he love to frighten the littles dogs?
Incredible stuff…
I really wonder how the animation would look, how the deformation textures fade into eachother… (Do they fade, or do they warp+fade etc)
This technique is excellent for lots of reasons probably, but here is one: Now matter how much you care about the topology, with different expressions, some wrinkles that form on the face can’t be realized, as their directions are quite arbitrary, esp. in the forehead and between the brows, not to mention the wrinkles the nose when you lift your nose wings up (like when you smell something bad and show your disgust…)
Sorry if these are mentioned, couldn’t read 5 pages…
