3 replies to this topic

[afraidofdark]

Hello, I am trying to implement lighting as close as possible to lighting done in 3ds max 2013 (realistic mode)
Here are the things that I make similar with max's lights and material in order to get same results
 
attenuation is inverse square in both max and my lighting
material diffuse, ambient & specular colors. Glossiness & specular level are same
both materials have phong shading model
point light effective area (where lights attenuation is 1) and light radius are same
 
Effective area start from lights origin and and has a radius "e", every pixel fall in to that region has an attenuation factor of 1
After effective region every light shows inverse square attenuation characteristic. I come to this solution after investigating and examining how lighting and attenuation is handled in max. But I still cant figure out how exactly lighting is handled in max.

here is what I've got so far.
 

 
Each light in max dominates its region with its color more clearly, in my program colors are mixed and green dominates the scene and blue is almost lost. This might be a gamma problem I am not sure, so I ask your opinions. 
 
To see the difference more clearly I changed the contrast of the image


also here is my lighting shader

void phongOmni(out float3 diffCont, out float3 specCont, in light lght, in material mat, in VS_OUT inp)
{
	float3 ep = mul(eyePos, invWorld).xyz; // object space eye position
	float4 l4p;
	l4p.xyz = lght.position;
	l4p.w = 1;
	float3 lp = mul(l4p, invWorld).xyz; // object space light position
	
	// lighting
	float3 n = normalize(inp.n);
	float3 v = normalize(ep - inp.p);
	float3 l = normalize(lp - inp.p);
	float3 h = normalize(v + l);
	
	float diffMag = max(dot(n, l), 0);
	diffCont = lght.color * diffMag; // diffuse contribution
	float specMag = pow(max(dot(n, h), 0), mat.glossiness * 100);
	specCont = mat.specularLevel * lght.color * specMag;  // specular contribution
	

        // Attenuation   
	float d = distance(lp, inp.p);
	float r = lght.attenuationEnd;
	float e = lght.attenuationStart;

        // if distance of the pixel to the light is less than effective region "e"
        // pixel's attenuation is 1
	float att = 1; 
	if (d > e)
	{
		d = d - e;
		r = r - e;
		float denom = d / r + 1;
		att = 1 / (denom * denom);
	}
	
        // if attenuation is less than 0.005 than it has a 0 attenuation
	att = (att -  0.005) / (1 -  0.005);
        att = max(att, 0);
	
	specCont *= att;
	diffCont *= att;
}

float3 PS(VS_OUT inp) : SV_Target
{	
	float3 dacum = float3(0, 0, 0); // specular light accumulation
	float3 sacum = float3(0, 0, 0); // diffuse light accumulation

	for (int i = 0; i < lightCount; i++) {
		float3 diffCont;
		float3 specCont;
		phongOmni(diffCont, specCont, lights[i], mat, inp);
		dacum += diffCont;
		sacum += specCont;
	}
	
	float3 color = (sacum + ((dacum + mat.ambientColor) * mat.diffuseColor));
	return color;
}

 

 

[ScottZhang]

maybe, gamma correction?

[afraidofdark]

Applying gamma correction really helped a lot. However in max, light stop smoothly when it reaches its radius. I am trying to implement it cheaply. Here is the progress

 

[afraidofdark]

yep, I did it. Now my lights behave exactly like max lights. The only difference than standard phong - point light shading is that; they apply a gamma correction, and inverse linearly (based on distance) scale the attenuation of the light. also the effective area thing that I explained above.

 

here is the final result. It is good enough for me.