One problem with 3d digitizing stuff is the output needs to be cleaned up a lot. Even then, a lot of the time that cleaned up mesh is just too damn dense to be used in anything more than a static object.
For example, say I model a really nicely detailed gargoyle using clay/sculpy/whatever. I then dump that into a 3d file using a 3d digitizer. Lets say for arguement's sake that the output weighs in at a a million triangles. There's a number of issues that come up at this point:
1, That's way too many triangles to deal with for stuff like unwrapping texture coordinates, weighting to bones, setting up morph targets for muscle/facial animation, etc. 2, The topology that was vomited out by the 3d scanner's software is going to be garbage topology wise. The current method of building a polygon based character is to build a low(er) res cage that is later subdivided to give you your high res final rendered model. These cages typically are constructed to have "edge loops" in them, that tend to mimic the orientation of muscle fibers in a real creature. I won't go into further detail on edgeloops, but they are an important consideration. Now, as I said the 3d scan software doesn't know anything about this, so you don't get these loops, and your mesh tends to deform _very_ poorly.
That being said, some people have found that you can use this high res mesh to generate a normal map which is used to drive a displacement on the subdivided mesh to achieve very, very nice results.
We're there already. Pricey little buggers, but if they're anywhere as good as the Intuos2 tablet I'd love one. Now if only I could convince my boss that it would in fact increase my productivity...
It is raster based, but it now can do raytracing as well. No need to use BMRT as a rayserver for PRman in frankenrender mode anymore. Check out Pixar's Renderman page for what its capable of now.
Pure speculation on my part, but it looks like the easiest way would be to output a.stl file (or whatever the compatible 3d file format would be) for a rapid prototyping machine. The machine would then build up a physical model that would then be the master for a mold so you could make billions of monkeys!
That's the whole point of using raytracing. The 1st image looks dull because its not reflecting anything. One could generate a reflection map and use that to approximate environmental reflections for the first image, but it wouldn't be very accurate.
My one issue with the second image is the lack of soft shadows. The shadowmaps in the 1st image look nicer than the overly crisp raytraced shadows in the second. If they could implement area lights or some other soft shadowing technique for raytraced shadows it would be good.
Actually, it was entirely self-guided. According to a few sites I've googled the V1 used a set of gyroscopes to keep it flying along a pre-determined heading and altitude.
There were guided bombs used by both sides, with some of the German types detailed here.
I, for one, welcome our new ganja-toking overlords! I'd like to remind them that as a trusted/. personality, I can be helpful in rounding up others to...toil in their underground hydroponics labs.
Its a common misconception that the computer does all the work in cg animation. A whole lot of work goes into making a workable character model. In my experience it takes a lot of time and effort to build somthing that looks good.
The basic process for a single character goes like this: 2d concept art, modeling of 3d geometry, rigging for animation, modeling of morph targets (facial animation, faking out muscles, etc.), setup of animation controls, texture coordinate unwrap, texture painting, material creation.
After all that, the character gets handed off to an animator who has the task of making this thing look like its alive. I havn't even gotten into what is involved in creating an entire environment. Suffice to say that there's a hell of a lot of time and effort spent by _humans_ in making this stuff.
I find it amusing that people romanticize 2d animation production. Pretty much all the ink and paint is done not by hand, but by a computer these
days . It's not some skilled artist wielding a brus h that's laying down those colors, but a computer, blindly following a user's inputs.
And I think even the RIAA would struggle to convince people they needed something that draconian - what if you dropped it and a wire snapped by accident?
Mugabe? We're doing nothing about him, unfortunately. However Canada is not infalliable, as a little digging on Google would show. When I first learned of this I had hoped some lesson would have been learned... but I guess not.
I have a sneaking suspicion for the price of a fully tricked out dual G5 mac I could build a very nice dual Opteron rig that would squash it. Not to mention I would have my choice of os to run on it, winxp 64, linux, bsd etc.
Slashdot Mirror
Did you start with an S, and then draw a more different S? Also, did you use consumate V's?! If not... well... Strongbad will not be pleased!
For example, say I model a really nicely detailed gargoyle using clay/sculpy/whatever. I then dump that into a 3d file using a 3d digitizer. Lets say for arguement's sake that the output weighs in at a a million triangles. There's a number of issues that come up at this point:
1, That's way too many triangles to deal with for stuff like unwrapping texture coordinates, weighting to bones, setting up morph targets for muscle/facial animation, etc.
2, The topology that was vomited out by the 3d scanner's software is going to be garbage topology wise. The current method of building a polygon based character is to build a low(er) res cage that is later subdivided to give you your high res final rendered model. These cages typically are constructed to have "edge loops" in them, that tend to mimic the orientation of muscle fibers in a real creature. I won't go into further detail on edgeloops, but they are an important consideration. Now, as I said the 3d scan software doesn't know anything about this, so you don't get these loops, and your mesh tends to deform _very_ poorly.
That being said, some people have found that you can use this high res mesh to generate a normal map which is used to drive a displacement on the subdivided mesh to achieve very, very nice results.
We're there already. Pricey little buggers, but if they're anywhere as good as the Intuos2 tablet I'd love one. Now if only I could convince my boss that it would in fact increase my productivity...
As long as I get to venture forth from the vault to find the replacement water chip its all good.
It is raster based, but it now can do raytracing as well. No need to use BMRT as a rayserver for PRman in frankenrender mode anymore. Check out Pixar's Renderman page for what its capable of now.
Pure speculation on my part, but it looks like the easiest way would be to output a .stl file (or whatever the compatible 3d file format would be) for a rapid prototyping machine. The machine would then build up a physical model that would then be the master for a mold so you could make billions of monkeys!
That's the whole point of using raytracing. The 1st image looks dull because its not reflecting anything. One could generate a reflection map and use that to approximate environmental reflections for the first image, but it wouldn't be very accurate.
My one issue with the second image is the lack of soft shadows. The shadowmaps in the 1st image look nicer than the overly crisp raytraced shadows in the second. If they could implement area lights or some other soft shadowing technique for raytraced shadows it would be good.
Actually, it was entirely self-guided. According to a few sites I've googled the V1 used a set of gyroscopes to keep it flying along a pre-determined heading and altitude. There were guided bombs used by both sides, with some of the German types detailed here.
Ok, so maybe no trusted, but whatever.
The basic process for a single character goes like this: 2d concept art, modeling of 3d geometry, rigging for animation, modeling of morph targets (facial animation, faking out muscles, etc.), setup of animation controls, texture coordinate unwrap, texture painting, material creation.
After all that, the character gets handed off to an animator who has the task of making this thing look like its alive. I havn't even gotten into what is involved in creating an entire environment. Suffice to say that there's a hell of a lot of time and effort spent by _humans_ in making this stuff.
I find it amusing that people romanticize 2d animation production. Pretty much all the ink and paint is done not by hand, but by a computer these days . It's not some skilled artist wielding a brus h that's laying down those colors, but a computer, blindly following a user's inputs.
Block 3: In Soviet Russia Natalie Portman has hot grits for Profit!
Profit!
- Infinium Labs
Sounds like the recipe they're using in the vending machines here...
nacctdreport.pl
-All queried networks: 100% (52083 MB)
3 Gigs? I averaged just over half that much per day this month. I wonder if my provider will be amused...
2. Extend
3. Extinguish
4. Profit!
Fun times ahead!
Last I checked the Canadian Census results the combined poplulations of New York and Los Angeles were more than all of Canada.
Mugabe? We're doing nothing about him, unfortunately. However Canada is not infalliable, as a little digging on Google would show. When I first learned of this I had hoped some lesson would have been learned... but I guess not.
Don't go there eh? Well, we'll just see about that!
*click*
Oh my god! My eyes! I'm blind! Damn you! Damn you all to hell!
Right after Mozilla Pinto, which will be a precise clone of ms outlook, security issues and all!
They forgot that you can only choose 2 of the above.
I have a sneaking suspicion for the price of a fully tricked out dual G5 mac I could build a very nice dual Opteron rig that would squash it. Not to mention I would have my choice of os to run on it, winxp 64, linux, bsd etc.
So, basically what this Lau guy's saying is:
1. Cut hole in object
2. ???
3. Profit!
Nobody will EVER need more than 640 kb!
Slashdot editors must have the memory spans of a gnat, or Homer...
"Ooh look a butterfly! Whee! Oh a bird! Wow! Look at the funny kitten!!!!"
Seriously though... do the editors even READ the articles that get posted?!
If you did not get 7/7 on that quiz please line up in the morons aisle, and ready yourself to be banished from /.