# pickups

Member

90

122 Neutral

• Rank
Member
1. ## Using TKinter with directx

Thanks, i was planning to look at the D3DUT gui. It looks really good. About that crazy eddie, it looks good too. I thought it is opengl only. It turns out i was wrong. Thnks!
2. ## Using TKinter with directx

Hello. Im trying to find a good GUI library i can use for my DirectX9 engine. I thought about using tcl/tk, because my general idea is to use python for the editor, and C++ for the engine's core. This approch has many advantages. But, im not sure it is possibile to create a D3D device for a tk widget, or window, becuase as far as i understand it, its not a real win32 window. Do you know whether it is possibile, or know of another directX GUI lib? Thanks.
3. ## Ocean water using FFT - The equations

No ,this is an open source FFT solver. i was asking about an implementation of the ocean solver. thanks anyway.
4. ## Ocean water using FFT - The equations

Maybe somebody knows where can i find a reference implementation?
5. ## Ocean water using FFT - The equations

Hello. Im working now on implementing the FFT algorithm to simulate ocean water, based on the deep-water article in gamasutra and on tesendorf's course notes. I understand the method and the math, and when wrote down the equations and developed them, I found out they turn out to be pretty simple, so im not sure im right. Maybe someone here can take a look: The main equation is: h(x, K) = H0(K)*e^a + H0*(-K)*e^-a Well, as far as i can tell, P(K) == P(-K), so the only difference between H0 to H0* is the imaginary parts are negated. So the final equation to compute H(x,K) is this (Before FFT) - H(x,K) = [ (H0(K).real + H0*(K).real))*cos(x) + (H0*(-K).imag - H0(K).imag)*sin(x) ] + i*[(H0(K).real - H0*(-K).real)*sin(x) + (H0(K).imag + H0*(-K).imag)*cos(x)] = [ 2*(H0(K).real)*cos(x) - 2*H0(K).imag)*sin(x) ] It is too simple to be true. Im missing something here. Anyone knows? Thanks.
6. ## Emulating bi-linear filtering in pixel shaders

Thats a good point. Thanks.
7. ## Emulating bi-linear filtering in pixel shaders

Right, i forgot this extension. No way to do it in d3d? What are the advantages of using 0 to 1 tex coords?
8. ## Emulating bi-linear filtering in pixel shaders

Right, i can sample the neighbours, but in order to perform bilinear filtering, i need to sample the texture at the four closest texels, and then interp to find the colour where i want, between them. In order to find these four closest texels i need the tex resolution no? About openGl, i also havn't used it for quiet a while, but i saw a cg program that treats the coords as 0 to width, and not 0 to 1.
9. ## Emulating bi-linear filtering in pixel shaders

Nobody knows? Hmm, it is quiet necessary to do bilinear filtering on floating point textures. Maybe in d3d9 the limitation of no filtering doesn't exist? I looked for some info of it on the docs, but couldn't find.
10. ## OpenGL Emulating bi-linear filtering in pixel shaders

Hi. I want to emulate bi-linear filtering on floating point textures. So i need to write a function to do it for me, right? I think i know how to do it, but the d3d way of handling tex coords seems very non efficient, so i just wanted to make sure what i know is correct: The tex coords are between 0 to 1, so in order to do bi-linear filtering i need to pass to the shader the texel size, and the texture resolution, or at least on of them, right? In openGl the texcoords are 0 to width/height, and thats easier to work with. Am i right?
11. ## rendering to float render targets

Yes i done it too. It is written in "Shader semantic syntax" where ps output semantics are described. but what is true, is that ps2.0 must output float4.
12. ## rendering to float render targets

Hello. I want to render into a floating point render target, but the documantation says that prior to ps3.0 the output of pixel shaders must be clamped to between 0 to 1. I basically wants to use the gpu to compute some floating point operations, so i need to use float textures and render float data into them. Is it not possibile with ps2.0?
13. ## Cloth simulation with an irregular mesh

"Extending the Inmediate Buckling Model to Triangular Meshes for Simulating Complex Clothes" This quiet new paper might be what you are looking for. It is an extension to "Stable but responsive cloth"
14. ## Water Simulation - FFT/ Perlin Noise

I remember a demo with this method. its in nvidia's developer site. I cant find it now, but try to look for it. maybe its in their FX Composer or something. Anyway, maybe i can help you, if you have a more specific question, and if ill find some time. In general, the FFT method just uses an inverse fourier transform to take a lot of random waves, and sum them up to create the height map. each position on the map has a height, and moves only vertically. It looks like water, because the amplitudes are calculated using an empirical equation, philips spectrum, if i recall curectly. Just read a lot, again and again, and read a bit about fourier transforms, you should be fine. leave the navier-stokes equations aside for now.