Recent content by shuipi

It's true the irradiance ratio between the 2 spheres should be inversely proportional to their areas, but the original problem gave the radiance (watt/sq meter/steradian) L at the light source's surface sphere, and asked for the irradiance(watt/sq meter) at some other point outside it.

I understand if there were an arbitrarily oriented plane passing through that point, to calculate the irradiance into that plane, I need to use the cosine factor. But this problem is without that plane, just the irradiance into that point.

Given a spherical light source with radius r and uniformly emits radiance L from every point on its surface in all possible directions, what is the irradiance at a point outside the light source that's distance R away from the center of the sphere? My idea is, every point on the surface of...

Ah never mind. I just realized that as long as the enclosed mesh doesn't have vertices coinciding with the AABB's vertices then this method returns bigger than necessary bounding sphere.

Everyone says the method of obtaining a bounding sphere by using the AABB's center as the sphere center and half diagnal length as the radius will not give an optimal bounding sphere. But I can't think of a case where it's not optimal. It looks pretty optimal to me: the radius can't be any...

Thanks for the explanation. I think my mistake was using dPhi * dTheta as the differential solid angle, which in fact should be sin(Theta) * dPhi * dTheta. Phi being the azimuth.

Say I have a point with a normal direction N, and a hemisphere of outgoing directions Lo around N leaving this point. Both N and Lo are just normalized directions. From the books it says the integral of N dot Lo for all possible Lo over the hemisphere is Pi, but I always get 2Pi, can you help...

Thanks for the explanation. On the correctness of this method, I think not only do you need a 180 degree fov or some sort of cube map for each texel, when down sampling them it also can't be a plaint down sample because the irradiance filter needs to be a cosine weighted filter

GI is traditionally done by tracing rays and bouncing them around on the cpu. But recently I've seen some people doing it by rendering a small image from each light map texel's perspective and filtering it to one value to represent irradiance, thus completely getting rid of the expensive ray...

note cone-plane intersection may not work because the box could completely enclose the cone without any instercetion yet still shouldn't be rejected

When doing lighting we usually want to on the cpu side do a rough culling between lights and objects, in order to save shader cost. Spot light is a common type of light used in many games, to determine if an object's AABB overlaps the spot light's cone however doesn't seem to be trivial. What's...

I'm in the process of my first implementation of a pre-computed GI solution. So I have a basic question: when a ray hits a diffuse surface, what is the common practice to retrieve the diffuse reflectance of that point? I assume returning some sort of "average diffuse color" of the surface would...

I need something similar to this for a shader running on pc http://msdn.microsoft.com/en-us/library/bb313952.aspx I think I can do it maually in shader, but want to know if there's an intrinsic or anything that can do it

halo 3 uses cook-torrence from their 2008 gdc and siggraph papers

if just let the hardware do the filtering, it will be linear filter, and the lerped normal won't be normalized anymore right? Most games just ignores this?