Sam Hocevar

11,995 reputation

22654

bio	website	sam.zoy.org
	location	Paris, France
	age	34
visits	member for	2 years, 2 months
	seen	25 mins ago
stats	profile views	917

I live in Paris, France.

I work on PlayStation, Xbox and computer games for a living. I write free software and all kind of crazy shit for fun. My interests: image processing, video coding, game development, maths, physics, compression, cryptography.

	11,995 reputation	bio	website	sam.zoy.org	visits	member for	2 years, 2 months
22654 badges		location	Paris, France		seen	25 mins ago

summary answers questions tags badges favorites bounties reputation activity

439 Comments

suggestions reviews revisions comments badges posts accepts all

1d	comment	Hexagon collision detection for fast moving objects? @madshogo You're right. I assumed that the timestep was small enough compared to object sizes that this wouldn't be a problem, but it's certainly not very robust in the general case. I'll look into fixing it.
Mar 29	comment	Why is my sky color calculation in Mathematica incorrect? The question is really too localised and might get closed, but thanks for the paper link, it's interesting.
Mar 28	comment	Why are graphics engines writern in C/C++, when games are GPU bound? This question really lacks acceptability of premises. A lot of the games I know are CPU bound.
Mar 20	comment	Why do game engines convert models to triangles compared to keeping it as four side polygon Being coplanar doesn’t “ALWAYS define a flat plane”.
Mar 11	comment	Handling window resize with arbitrary aspect ratios possible duplicate of Scaling a game view with letterboxes?
Feb 25	comment	limiting jump forces to achieve a maximum jump height Note that applying a force for a long time is not jumping, it’s more like some kind of thruster effect.
Feb 17	comment	How can I replicate the color limitations of the NES with an HLSL pixel shader? This is not the NES colour palette at all. The NES didn’t use 8-bit RGB, it used a fixed palette of about 50 to 60 colours in the YPbPr space.
Feb 17	comment	What's the best way of translating a 2D vector into the closest 8-way compass direction? A `map` with `float2` as the key? This doesn’t look very serious.
Feb 14	comment	What's the best way of translating a 2D vector into the closest 8-way compass direction? +1 by the way, I really think this is the most straightforward and rigourous answer.
Feb 14	comment	What are the main requirements to port a game to Linux About Valve, note that they hired Sam Lantinga, the original author of libSDL.
Feb 14	comment	Animate part of .3ds object using lib3ds library @Sidar I approved your edit and further removed the signature at the bottom of the post, so it's that second edit that appears. You can still see yours in the history.
Feb 14	comment	What's the best way of translating a 2D vector into the closest 8-way compass direction? It's perfectly OK to answer your own question. But I would consider not using `atan2` to be a bad decision; your code is a lot longer and difficult to understand. It's also... very wrong: where did you get these 0.5 and 2 values? You would probably need sqrt(2)/2 and sqrt(2)+1 instead.
Feb 14	comment	What's the best way of translating a 2D vector into the closest 8-way compass direction? Note that it's usually `atan2(y,x)`, not `atan2(x,y)`.
Feb 13	comment	Using two FBOs results in the second FBO having nothing drawn to Have you tried gDEBugger maybe?
Feb 12	comment	3D isometric depth sorting The reason no algorithm seems to handle what you are looking for is because there is no such algorithm. It's not possible to Z-sort tiles in the generic case. Even in a situation where there is a solution, it may be hard to find it (look for "topological sort"). The obvious option is to use a Z-buffer.
Feb 12	comment	Jagged transparency when rendering translucent textures (OpenGL 3.2 + GLSL) You are right. I used the wrong argument order. Thanks, answer fixed!
Feb 11	comment	Jagged transparency when rendering translucent textures (OpenGL 3.2 + GLSL) `smoothstep` may not be necessary at all; but its purpose here is to avoid any visible discontinuities when the alpha is near 0.1, by stretching these values to 0.0 using a smooth curve.
Feb 11	comment	Lerp vs Vector math: one better than the other? In your first example you mention a "desired distance", yet in the second one you use a "speed value". Can you maybe clarify? Anyway, lerp() is only one scalar-vector multiplication and a few additions; it should be extremely fast.
Feb 3	comment	What book/guide should I follow for GLSL I second the above comment. From what you explain it seems you’d also need something like this book (don’t pay attention to the ugly cover; the content is excellent).
Feb 2	comment	vec4 for vec3 multiplication while calculating the specular value @RamyAlZuhouri Usually the alpha value does not come from the same light calculations as the colour value. You can certainly set it to 1.0 if your object is supposed to be opaque.