I had a thought while ago. I read about bitBoard and I was curious, could be the concept be applied in 3D?
What I mean is to use a 3D matrix of boolean value, which correspond to the voxels. So that one could use the and, or, not, ecc bit operations to calculate if a position for example is occupied or not.
for example let's consider the two objects (represented by 1's):
object1 object2 result: no collison
0000000000 0000000000 0000000000
0001110000 0000000000 0000000000
0001110000 & 0000000000 = 0000000000
0001110000 0000000011 0000000000
0000000000 0000000011 0000000000
0000000000 0000000000 0000000000
would it be useful? Are there already games that uses this trick?
>>1
You could do it using a 1-bit 3D volume texture and a pixel shader or compute shader. But its probably overkill compared to just using a hierarchy of cheap collision primitives and then geometrically solving for the collision position and normal against a low-LOD collision mesh.
Name:
Anonymous2011-12-24 13:58
>>1
enjoy trying that for circular objects in 3 dimensional space.
enjoy the memory bloat
Name:
Anonymous2011-12-24 14:39
>>4
Geomenry is harder to do than a cell-automation and geometry requires heavy math knowledge with Set Theory and the shits. Graphics would be much simpler if everything was made of solid cubes.
Name:
Anonymous2011-12-24 14:40
>>5
"Circles" don't exist in real life, as well as this "number Pi" pseudoscience
Name:
Anonymous2011-12-24 15:19
>>7
Yes we all know down at the super small scale circles are just a series of ridiculous points that look nice and circular at our level.
You can go suck a dick someone, no one cares.
Name:
Anonymous2011-12-24 15:22
>>8
At small scale there no fucking "infinitesimal points", there are probably just a simple cellular automaton that operates on a finite 1s and 0s.
>>6
Geometry precedes the Jewish influence of mathematics by thousands of years. And 3D bitmasks for collisions would be more difficult than geometric solutions when it comes to animated models.
Name:
Anonymous2011-12-24 19:21
>>13 Geometry precedes the Jewish influence of mathematics by thousands of years.
But only jews add infinitesimal points into it.
bitmasks for collisions would be more difficult than geometric solutions when it comes to animated models.
it depends on definition of "animation".
basically, what you understand under this term is a pseudo-science.
Name:
Anonymous2011-12-24 19:34
>>14
Solving for collisions between geometric primitives does not require the invocation of infinitesimals. It's simple algebra.
By animation, I mean skeletal animation.
Name:
Anonymous2011-12-24 19:43
>>15
>By animation, I mean skeletal animation.
simulate it with cellular automaton
Name:
Anonymous2011-12-24 19:45
>>15 Solving for collisions between geometric primitives does not require the invocation of infinitesimals. It's simple algebra.
Simple algebra already uses Set Theory (groups, monoids and the shit).
Name:
Anonymous2011-12-24 19:48
Here's one semi-practical way:
You have your volume and all voxels existing, either in a 1:1 bit mapping or a sparse octree.
Every time you add a new voxel or move an existing one, you & it into the space in which it will exist. If any truths result, then you don't move it, otherwise you remove where it was and create where it now is. This process is more memory intensive as you have to remove all of it and replace it, though it might be less intensive to have a "was" map and a "will-be" map, and then you just & those to determine if you have to change the state of any of it's existing volume (though you may not want to do this unless you're working the game in 1-bit monochrome). Also, use a fixed-dimension universe, that way accessing a voxel cell will always take O(1) operations.
The problem with any voxel-based engine, is everything in a collision-oriented world is basically the intersection of surfaces. To make a collision-engine for voxels, you have to abstract a surface, which will always take more processing or memory than just using geometric surfaces. Even worse, usually it's impractical as fuck to try rotating objects in voxel space, as you either have a lossy volume, or you take the voxel object, convert it to coordinate geometry, then convert it back to a voxel object, which is adding a retarded step compared to just have a coordinate object in the beginning.
It might be easier if I recommend the information be voxel # then palette color or something, then use a keycolor to indicate empty, such as 0, which would in theory allow you to use the bitwise operation on the color bytes, and if it's anything other than 0, it's occupied, while using color-changing operations on the same information space, to conserve memory and complexity. Fullbright would also make a good choice, as you could just use NAND which is likely more primitive in the cpu.
Name:
Anonymous2011-12-24 19:48
>>16
You're obviously out of your league. Skeletal animation can be described in terms of cellular automata! It's a bunch of bone vertex weights (cells) and they influence their adjacent vertices through what's known as vertex tweening, which are just some simple rules using linear algebra. No infinite sets, not infinitesimals, no non-computable real numbers. If did use any of those, making it non-implementable in Turing-equivalent computers, the past couple of decades of 3D games would be vastly different.
>>20
the definition of "computable" already depends on set theory http://en.wikipedia.org/wiki/Computable
>A decision problem fixes a set S..
That is all. It a set pseudoscience. There is a jew somewhere behind curtains. Call the Nazis!
>>21
>the definition of "computable" already depends on set theory
Not in Computer Science faggot. Set Theory is INFINITELY more complicated than value-dependent mathematics, which makes it require more rules and data types, i.e. it takes more to compute.
Your Jew Math is at the bottom, not the top of both realism and fantasy, it's only at the top of people who think making up rules and making other people obey them is Correct Math.
Name:
Anonymous2011-12-24 20:17
>>22
There is no Computer Science without Set Theory. Even O(N)/O(log2(N)) crap is defined using some "Set of Shitominals". Computer Science is a Pseudosciens, that explores imaginable objects, instead of real world.
Name:
Anonymous2011-12-24 20:24
>>23
The real world is computable. Deal with it, kike.
>>24
the definition of "computable" already depends on set theory http://en.wikipedia.org/wiki/Computable A decision problem fixes a set S..
That is all. It's a set pseudoscience. There is a jew somewhere behind curtains. Call the Nazis!
Name:
Anonymous2011-12-24 20:35
>>25
Mathematics is the emergent pseudoscience. Mathematics is nothing more than distilled declarative computational knowledge from our own imperatively computational minds, which themselves are emergent artifacts of the computational Universe in which we reside.
Universal Computation (Turing machines) can be implemented in terms of nothing more than Peano arithmetic, or vice versa. No set theory required. The Göttingen Jew may have infected mathematics and much of the sciences last century, but try as he might, he can't distort Computation, for to do so would be to deny the truth of our Universe.
Name:
Anonymous2011-12-24 21:27
`
To have a Computer Science you must first code a Hamburger.
>>35
I don't like the Jews, but I'm afraid Putnam and his philosophy of reductive physicalism is correct. What will you do when we can construct substrate independent minds?
Name:
Anonymous2011-12-24 23:10
>>37
Why should I do anything? I'm a lisper and a subjectivist, so I've strong perception, that the lambda is controlled by it's environment. Even your crappy Unix app sees it's world through chroot jail. Everything is fucking subjective!
Name:
Anonymous2011-12-24 23:20
>>38
Subjectivism is an emergent consequence of a discrete reductive materialistic world: subjective minds naturally arise as a consequence of Computationalism, Occam's Razor and The Pigeon Hole Principle.
Don't confuse the reductive materialistic ideologies with the other materialistic macroscopic ideals of objects like "chairs" or "forks" existing independently of the mind--they do not.