Search found 23 matches

by Christer Ericson
Tue Oct 21, 2008 6:32 am
Forum: Research and development discussion about Collision Detection and Physics Simulation
Topic: what does "sweep" and "prune" mean ?
Replies: 5
Views: 8708

Re: what does "sweep" and "prune" mean ?

I think it was Christer Ericson who said the algorithm should really be called "sort and sweep" instead of "sweep and prune". To me it looks more like "sort and prune" :) I have said that, and the reason I've said that is because Baraff was the first one to describe th...
by Christer Ericson
Tue Oct 21, 2008 6:21 am
Forum: Links, Papers, Libraries, Demos, Movies, Comparisons
Topic: Universal SIMD Mathlibrary
Replies: 6
Views: 11187

Re: Universal SIMD Mathlibrary

Mathfunctions for single precision floating point vector datatypes are provided for the SIMD-platforms x86 (SSE2), PowerPC and Cell. In most cases, speed and/or accuracy compare favourable with existing SIMD-libraries (MacOS Accelerate Framework, Cell SDK). Most of the algorithms are based on those...
by Christer Ericson
Sun Sep 09, 2007 8:36 pm
Forum: Research and development discussion about Collision Detection and Physics Simulation
Topic: generic swept sphere collision detection algorithm.
Replies: 4
Views: 8394

Re: generic swept sphere collision detection algorithm.

All that sphere collision talk got me thinking a little. I am thinking of an algorithm to collide spheres against over convex shapes, which is similar to the GJK algorythm, and is an iterative method. However it requires square roots to find the point of contact (due to the sphere's nature). It's u...
by Christer Ericson
Thu Aug 02, 2007 5:44 am
Forum: Research and development discussion about Collision Detection and Physics Simulation
Topic: Multi SAP
Replies: 22
Views: 45172

All I'm saying is that "multiresolution grid" is a well-known data-structure, e.g. described in Christer Ericson's book (he calls them "hierarchical grids"). The reasons I called them hierarchical grids and not multiresolution grids were: (1) hierarchical grid was (and is) the m...
by Christer Ericson
Sat Jun 16, 2007 6:12 am
Forum: Links, Papers, Libraries, Demos, Movies, Comparisons
Topic: Rotational penetration depth
Replies: 5
Views: 10996

Re: Rotational penetration depth

Interesting concept, calculating rotational penetration depth, rather then just translational[...] Here's another on the same topic: http://www.geometrie.tuwien.ac.at/nawratil/gpdcbokg.pdf This one is outstanding in the sense that it manages to make the presentation just about unreadable by overloa...
by Christer Ericson
Sat Mar 10, 2007 11:16 pm
Forum: Research and development discussion about Collision Detection and Physics Simulation
Topic: accuracy of collision detection: Bullet vs. Solid
Replies: 17
Views: 26646

I think you should also mention that those bullet points are not facts, they are your own personal opinions, and it is unlike they have any base on any rigorous analysis. I see nothing objectionable in Erwin's post, which Erwin very clearly marked as "my thoughts" at the start of the arti...
by Christer Ericson
Sat Nov 04, 2006 7:21 am
Forum: General Bullet Physics Support and Feedback
Topic: Stackless trees
Replies: 10
Views: 14682

I can imagine different definitions for "stackless trees." How are the stackless trees you guys are discussing set up? What are the perceived benefits and usage-cases?
by Christer Ericson
Sun Oct 29, 2006 11:45 pm
Forum: Research and development discussion about Collision Detection and Physics Simulation
Topic: Interval arithmetic.
Replies: 1
Views: 6546

Re: Interval arithmetic.

Does anyone know of any papers/text regarding interval arithmetic? My maths skills are fair, so the easier to read the better. Oh, there are lots of papers! For starters, try the papers by Tim Hickey . E.g. this one: Hickey, Timothy. Qun Ju. Maarten van Emden. ?Interval Arithmetic: From Principles ...
by Christer Ericson
Sun Jun 25, 2006 7:35 pm
Forum: Research and development discussion about Collision Detection and Physics Simulation
Topic: Fast Realtime bounding volume approaches ?
Replies: 4
Views: 7711

Re: Fast Realtime bounding volume approaches ?

I can build an optimal sphere quickly, but most clothing doesn't really conform to this kind of test well. most cloth is planar majority of the time. AABB too is not optimal in world space.. I think I need to use an arbitary BB. That is Convex hull of 6 planes. Hi Andi! I probably wouldn't try to c...
by Christer Ericson
Sun May 28, 2006 6:33 pm
Forum: General Bullet Physics Support and Feedback
Topic: collision with very big static object
Replies: 13
Views: 24834

You are probably right and I am wrong. :( Almost humble?! Doesn't sound at all like the child who first posted this belligerent post: Wow, thanks you for letting it me know about that algebraic identity (I feel so much empowered now). Is this a principle you came up with? Would that be the Ericson?...
by Christer Ericson
Wed May 24, 2006 9:22 pm
Forum: General Bullet Physics Support and Feedback
Topic: collision with very big static object
Replies: 13
Views: 24834

Maybe I should had said they are numerically equivalent You could say that, but that would be wrong, assuming you by "numerically equivalent" mean having the same error numerically. (If you mean something else, you need to define what you mean by that nonstandard term.) For example, consi...
by Christer Ericson
Wed May 24, 2006 7:15 am
Forum: General Bullet Physics Support and Feedback
Topic: collision with very big static object
Replies: 13
Views: 24834

As I remember it was said that by using Voronoi space search the numerical errors due to lost of precision in the analytical distance calculation are avoided. In fact it can be proven that both methods are mathematically equivalent. You misunderstood. It is trivially the case that we can have two e...
by Christer Ericson
Sat Apr 29, 2006 6:01 am
Forum: Research and development discussion about Collision Detection and Physics Simulation
Topic: Ageia PPU convex-convex performance
Replies: 1
Views: 6353

Re: Ageia PPU convex-convex performance

- Peak instruction bandwidth: 20 Billion/sec - Sphere-sphere collisions: 530 Million/sec max - Convex-convex (complex collisions): 533,000/sec max [...] The convex-convex collision information is interesting. I guess it refers to a convex-convex case of up to 32 vertices. It doesn't tell if that is...
by Christer Ericson
Fri Feb 03, 2006 7:55 am
Forum: Research and development discussion about Collision Detection and Physics Simulation
Topic: Gino's Book - Robustness section
Replies: 4
Views: 9150

Re: Gino's Book - Robustness section

So, from page 59 I'm not understanding what Gino is trying to explain in the phrase that starts with "Assume". I'm not Gino, but perhaps I can explain anyway. Let's start by nothing that a typical floating point operation 'op' introduces an error along the lines of: fl(a op b) = (a op b)(...
by Christer Ericson
Fri Jan 20, 2006 6:46 am
Forum: Research and development discussion about Collision Detection and Physics Simulation
Topic: Interval arithmetic, interpenetration, repositioning
Replies: 8
Views: 14489

Re: Interval arithmetic, interpenetration, repositioning

2. Why is it preferable to check whether the point of impact is inside the edge/triangle only after the time of impact (with the infinite line/plane) has been determined? The way I see it, this allows some collisions to be missed. For example, a vertex that's supposed to collide with a triangle mig...