…violet blue creature 
Pilou
WOW!
Excellent concept, modeling, Z and awesome mr
It strikes that all of those really cool high-level details couldn’t have been there unless they were modeled in one of the normal or disp. maps. For the second render, did you modify the disp map to include all of those details?
Thanks,
Ken
sunit, This is cool! great job! I wish you could do subsurface scatteringin ZBrush
monstermaker
hi sunit…excellent thread and model…fun to watch your experiments!!
I am just basically average joe user who hasn’t had the time yet to get into displacements or normal maps etc in other programs. But hope you can answer a question for the particular program you are using or any other you may have tried this experiment on. yes I know i am too curious for my own good!
Anyway, in this particular program or any others you use, I am curious to how the renderer evaluates the various displacements…by that I mean…i see folks post that they used a displacement map and a normal map and a bump map or any variation of those. My question is how well does the renderer you are using apply those so one does not wipe out detail from the next…if that makes sense…eg does the order of how they are applied matter? etc.
I do not wish to take away or interupt your work with the question but am curious and probably others are too…if you are up to answering and it’s too involved…start another thread so it will not detract from what you are doing here…and please don’t feel you have to explain if you don’t have the time…specially if it is too involved.
Thanks for your time, and sharing your work!
Awesome!!
Really inspiring stuff! This picture is motivating me to learn Zbrush in depth as well…,
Also, to get out the Dark Crystal and labyrinth dvd’s again, and those Brian Froud books, and… well, just to get to work, really 
Brilliant stuff, thanks for sharing!
PS: I’m also thinking forest spirit…, a really dark forest…
Thanks!
-Stijn.
hi aminuts,
the three (displacement, normal, bump) rely on two different methods of detailing - when a renderer calculates a surface, each pixel is given a surface normal (eg. a vector/line perpendicular from that pixel) that tells the renderer how to calculate light, shadows, etc. (all the things renderers do that i don’t anything about :)). a bump map will trick the renderer into thinking that the point is pushed in or out along that perpendicular vector (the z-depth to that pixel) without actually displacing the geometry.
a normal map will rotate that vector in all three axes - so you get a lot more information about that pixel. both of these methods though have nothing to do with the geometry. they’re illusions (introduced by Jim Blinn in the late 70’s) so when you look at the silhouettes, you can see the model geometry doesn’t actually change.
displacement mapping, on the other hand, actually moves pixels (and in the process, depending on your displacement type, adds detail to the model). so, at render time, you actually render a much higher res model. there are many different ways of subdividing the geometry to get the fastest and nicest geometry while still retaining the detail - the reyes renderman specification is probably the fastest i’ve used (although i don’t know the specifics) and mental ray these days has some nice spatial subdivision algorithms to make displacement faster than it used to be.
you can also rotate displacement vectors (like the horns for that “dinosaur” cg feature by disney a few years ago), but that’s a whole new area that i don’t yet know much about.
does that answer the question? bump is pretty much a pared down normal map, so those two will conflict unless you’re combining them before feeding them into the normalcamera of your shader (at least this is the way it works in maya…).
Thanks for taking the time, Sunit!
You explained it all well, some I knew and some I understand better now because you put it in a way I hadn’t read before so all is well!!
And yes you answered my question!!
Thanks again!
Aminuts?
i should just clarify - i think my post is actually a bit wrong above.
i don’t know about ohers, but at rendertime, mental ray will tesselate your models into triangles. what displacement does is move the vertices of these triangles (not the pixels, as i posted above). you can then also choose to subdivide the geometry to help deal with the added detail of the displacement render. mental ray uses a bsp tree to subdivide models (i’m just now reading a book on writing renderers, and what they actually do), which basically partitions your model into smaller and smaller bits at render time depending on where the detail is.
chees,
sunit
ok now am even more curious ahahhaa
why tri’s…am guessing that’s cus of older conventional thinking but really don’t know. if zbrush can displace on quads rather well…what would be the advantage? (i mean converting to tri’s ) more points to push? I believe I have read that some renderers don’t subdivide or smooth til render time on displacements which I reckon is good cuz you save eons setting up your animations without the added cost…but more questions hahaha
which brings up smoothing questions etc etc…perhaps hahaha it would be easier to point me to the book you are reading?
and am curious who wrote it…am wondering if it’s one of the stone soup guys.
I like what you have here very much. Although I am more interested in what you stated about about possibly animating him. Are you using real SSS or are you faking it? Have you had much experience animating objects with SSS applied to them? I can’t wait to see what you do with this guy, I agree with the gentlemen above, why look to the past for monsters. Or anything for that matter.
D. Ryan Reeb
Texture lead
Double Edge Digital
cool model man! love the s.s. scattering
Thanks for taking the time to explain some particulars SUNIT,
I know how busy you are