Jump to content

  • Log In with Google      Sign In   
  • Create Account

We need your feedback on a survey! Each completed response supports our community and gives you a chance to win a $25 Amazon gift card!


Member Since 29 May 2010
Offline Last Active Oct 08 2014 07:35 AM

Topics I've Started

Specular Highlights with Directional Lightmaps

12 August 2014 - 06:00 AM

Hi, i've been adding directional lightmaps to an old DX9 engine, which is working fine, but i wanted to have a go at adding specular highlights for static lighting. I've heard that other engines such as Unity and UE3 have this but my google-fu is failing me as I can't find any solid implementation details. So, i've had a blind stab at it myself but there's a divide by zero issue blocking me at the moment.
The theory is that the directional lightmaps have an approximation of the light hitting a surface, so rather than getting the diffuse light by using the surface normal directly (with normal map) I get the reflected light by using the reflected eye vector. I then use the luminance of the reflected light as it's attenuation, on which the specular exponent and coefficient is applied before multiplying the reflected light to get the final specular value.
This doesn't seem to work too badly, even if it is totally innacurate and bodged together tongue.png.
Diffuse light
Specular Light
Specular attenuated
Baseline (No directional lightmaps)
Final Diffuse
Final Diffuse + Specular

static const float3 BumpBasis[3] =
	float3(	sqrt(2.0) / sqrt(3.0),				0.0,	1.0 / sqrt(3.0)),
	float3(	-1.0 / sqrt(6.0),		1.0 / sqrt(2.0),	1.0 / sqrt(3.0)),
	float3(	-1.0 / sqrt(6.0),		-1.0 / sqrt(2.0),	1.0 / sqrt(3.0))

float3 CalculateDLMContribution(float3 TangentNormal)
	TangentNormal.y = -TangentNormal.y;
	float3 lightmapContribution = 0;
	lightmapContribution.x = dot(BumpBasis[0], TangentNormal);
	lightmapContribution.y = dot(BumpBasis[1], TangentNormal);
	lightmapContribution.z = dot(BumpBasis[2], TangentNormal);
	lightmapContribution = saturate(lightmapContribution);
	lightmapContribution *= lightmapContribution;
	return lightmapContribution;
float3 CalculateDLMDiff(float3 TangentNormal,
	float3 DLightmap0,
	float3 DLightmap1,
	float3 DLightmap2)
	float3 lightmapContribution = CalculateDLMContribution(TangentNormal);

	float3 light = 0;
	float sum = dot(lightmapContribution, float3(1.0f, 1.0f, 1.0f));
	light += DLightmap0 * lightmapContribution.x;
	light += DLightmap1 * lightmapContribution.y;
	light += DLightmap2 * lightmapContribution.z;
	light /= sum;
	return light;
float3 CalculateDLMSpec(float3 TangentNormal,
	float3 EyeVectorTS,
	float3 DLightmap0,
	float3 DLightmap1,
	float3 DLightmap2)
	float3 reflectedEye = -reflect(normalize(EyeVectorTS), TangentNormal);	
	float3 lightmapContribution = CalculateDLMContribution(reflectedEye);
	float3 light = 0;	
	light += DLightmap0 * lightmapContribution.x;
	light += DLightmap1 * lightmapContribution.y;
	light += DLightmap2 * lightmapContribution.z;
	light /= 1.33f; // Arbitrary value :(
	float luminance = dot(light, float3(0.30f, 0.59f, 0.11f));
	luminance = pow(luminance, 6.0f) * 7.0f; // Hardcoded test values
	return saturate(light * luminance);
However, the issue I mentioned is with the light averaging that occurs in the diffuse light calculation. I'd expect to have to do the same in the specular light calculation but if I do I end up with a divide by zero since the reflected vector can point below the surface normal resulting in zero light contributions from any of the 3 basis directions. This causes a harsh black spot in the resulting specular light, and i'm not sure how to counter it.
Black Spots
At the moment i'm dividing by an arbitrary value to get a reasonable result but i'm hoping there's a *proper* way to handle this?

Directional Lightmap issues

13 July 2013 - 11:15 AM

Hi, a short while ago I wrote a Maxscript tool for 3ds MAX to bake out directional lightmaps but I wasn't getting the results I had expected when previewing the output. However, after getting someone with Maya and Turtle to bake out some example radiosity normal maps to compare with, it turns out my script is actually working correctly. So the issues i'm having must be shader side.


The problem I have is naturally best shown with pictures smile.png:



The test scene is a cube with a point light dead centre in it, so as you can see the lighting is being taken from the wrong directions. Looking at it I thought I might just need to flip the Y value of the normal map which partially works:



...giving me working walls, but the floor is still wrong and I don't know why sad.png.


The shader function i'm using is as follows:

static const float3 BumpBasis[3] =
	float3(	sqrt(2.0) / sqrt(3.0),				0.0,	1.0 / sqrt(3.0)),
	float3(	-1.0 / sqrt(6.0),		1.0 / sqrt(2.0),	1.0 / sqrt(3.0)),
	float3(	-1.0 / sqrt(6.0),		-1.0 / sqrt(2.0),	1.0 / sqrt(3.0))
float3 CalculateDLM(float3 TangentNormal,
	float3 DLightmap0,
	float3 DLightmap1,
	float3 DLightmap2)
	float3 LightmapContribution = 0;
	LightmapContribution.x = saturate(dot(TangentNormal, BumpBasis[0]));
	LightmapContribution.y = saturate(dot(TangentNormal, BumpBasis[1]));
	LightmapContribution.z = saturate(dot(TangentNormal, BumpBasis[2]));
	LightmapContribution *= LightmapContribution;

	float3 diffuseLight = 0;
	float diffuseSum = dot(LightmapContribution, float3(1.0f, 1.0f, 1.0f));
	diffuseLight += DLightmap0 * LightmapContribution.x;
	diffuseLight += DLightmap1 * LightmapContribution.y;
	diffuseLight += DLightmap2 * LightmapContribution.z;
	diffuseLight /= diffuseSum;
	return diffuseLight;

The passed in TangentNormal value is the unbiased normal map texel and the ordering of the bump basis values is from here as apparently the ordering in the valve paper is misleading.


For good measure this is how the TangentNormal values are coming out:



I'm wondering whether I need to transform the basis in some way in relation to the mesh normals but It's not clear what i'm missing from the papers/examples I've seen.

[3ds Max] Directional Lightmap baking script problems

16 May 2013 - 05:38 AM

Hi, i've been working on a script to bake directional lightmaps (or Radiosity Normal maps as Valve called it) using 3ds Max but i'm having issues with the lightmaps it produces.

The script modifies the normals of the objects to face the 3 basis vectors before baking the lightmaps, but I think there is something amiss when doing this as the lightmaps aren't producing the required result.

The 3 lightmaps for an unwrapped box end up coming out like this:


And the final lightmap with shader looks like this:


The code that modifies the normals is as follows:


-- computeTangentSpace function taken from HalfVector on cgsociety
-- The face position has been removed from the resulting matrices
-- http://forums.cgsociety.org/showpost.php?p=3881243&postcount=11
function computeTangentSpace obj =
    local theMesh = snapshotAsMesh obj
    local tSpace = #()
    -- Do we have to flip faces?
    local flip = false
    local indices = #(1, 2, 3)
    if dot (cross obj.transform.row1 obj.transform.row2) obj.transform.row3 <= 0 do (
        indices[2] = 3
        indices[3] = 2
        flip = true
    for nFace = 1 to theMesh.numFaces do (
        local face = getFace theMesh nFace
        local tface = getTVFace theMesh nFace
        local v1 = getVert theMesh face[indices[1]]
        local v2 = getVert theMesh face[indices[2]]
        local v3 = getVert theMesh face[indices[3]]
        local uv1 = getTVert theMesh tface[indices[1]]
        local uv2 = getTVert theMesh tface[indices[2]]
        local uv3 = getTVert theMesh tface[indices[3]]
        local dV1 = v1 - v2
        local dV2 = v1 - v3
        local dUV1 = uv1 - uv2
        local dUV2 = uv1 - uv3
        local area = dUV1.x * dUV2.y - dUV1.y * dUV2.x
        local sign = if area < 0 then -1 else 1
        local tangent = [0,0,1]
        tangent.x = dV1.x * dUV2.y - dUV1.y * dV2.x
        tangent.y = dV1.y * dUV2.y - dUV1.y * dV2.y
        tangent.z = dV1.z * dUV2.y - dUV1.y * dV2.z
        tangent = (normalize tangent) * sign
        local normal = normalize (getFaceNormal theMesh nFace)
        if flip do normal = -normal
        local binormal = (normalize (cross normal tangent)) * sign
        append tSpace (Matrix3 tangent binormal normal [0,0,0])
    delete theMesh
    return tSpace
function SetNormalsModifier nodeList &editNormalsMod lightmapType =
    normalBasis = [0, 0, 1]
    if (lightmapType == 1) then
        return 0
    else if (lightmapType == 2) then
        normalBasis = [sqrt (2.0 / 3.0), 0, 1.0 / (sqrt 3.0)]
    else if (lightmapType == 3) then
        normalBasis = [-(1.0 / (sqrt 6.0)), 1.0f / (sqrt 2.0), 1.0f / (sqrt 3.0)]
    else if (lightmapType == 4) then
        normalBasis = [-(1.0 / (sqrt 6.0)), -(1.0f / (sqrt 2.0)), 1.0f / (sqrt 3.0)]
    for j = 1 to nodeList.count do
        obj = nodeList[j]
        -- Get the face TBN's
        meshFaceTBNs = computeTangentSpace obj
        normalFaceCount = editNormalsMod.getNumFaces node:obj
        normalVertexCount = editNormalsMod.getNumNormals node:obj
        -- Create an array of arrays to contain the vertex normals for averaging later
        normalsListArray = #()
        for i = 1 to normalVertexCount do
            append normalsListArray #()
        for i = 1 to normalFaceCount do
            faceSelectionArray = #{i}
            normalSelection = #{}
            -- Get the normal indices of the selected face
            editNormalsMod.ConvertFaceSelection &faceSelectionArray &normalSelection node:obj
            -- Get the face TBN
            TBN = meshFaceTBNs[i]        
            for k in normalSelection do
                -- Transform the normal basis into tangent space?
                normal = normalize(normalBasis * TBN)
                -- Append the new normal to a list for this vertex
                append normalsListArray[k] normal
        for i = 1 to normalVertexCount do
            normal = [0, 0, 0]
            -- Average the normals from the list for this vertex
            for k = 1 to normalsListArray[i].count do
                normal += normalsListArray[i][k]
            normal /= normalsListArray[i].count
            -- Set the vertex normal
            editNormalsMod.SetNormal i (normalize normal) node:obj


I've also attached the script as-is and a DirectX shader to test with. The script is set up as a macroscript so will need to be added as a button to your UI. The shader requires lightmap coords and normal texture coords in channels 1 and 2, and has a checkbox to state which is which.

I have a feeling there is a matrix transform i'm missing to get the normal bases from Direct3D's coordinate system to 3ds' but nothing i've tried has magically fixed the problem and I don't understand everything here well enough to know the exact reason why it doesn't work.

Apologies if this is the wrong forum, was toss up between graphics theory and tools :).

Problems converting vertex lighting assembly to per pixel lighting HLSL

29 May 2010 - 05:06 AM

Hi, i'm doing some modifications to Halo Custom Edition (PC) to add some new things to the graphics engine, and currently i'm trying to get normals maps onto the game objects such as vehicles and scenery. But i'm having a few problems. The game uses vertex lighting for its ingame models and I need to move that lighting into the pixel shader, but the vertex shaders are written in assembly and my attempts to convert them haven't been giving the correct results. The light attenuation code in the vertex shader is like so:
add r4.xyz, c[15], -r0
dp3 r4.w, r4, r4
rsq r10.w, r4.w
mul r4.xyz, r4, r10.w
mad r8.x, r4.w, -c[15].w, v4.w
dp3 r9.x, r4, r1 
dp3 r10.w, r4, -c[16]
mul r10.w, r10.w, c[16].w
add r8.z, r10.w, c[17].w
r0 has the vertex world position, r1 has the vertex normal and v4 is texture coordinate 0. I don't know alot about lighting so i'm unsure as to what all of the constants contain, but my current conversion looks like this:
sampler TexSampler4 : register(s4);

struct PS_INPUT{
	float3 PositionWorld		: COLOR0;
 	float4 DiffuseMultiUV		: TEXCOORD0;
 	float3 Normal	 		: TEXCOORD2;
 	float3 BiNormal 		: TEXCOORD3;
 	float3 Tangent	 		: TEXCOORD4;

float4 c_object_point_light0_position : register(c[15]);
float4 c_object_point_light0_rotation : register(c[16]);
float4 c_object_point_light0_colour : register(c[17]);

void Lighting(in PS_INPUT IN, out float4 Lighting)
	Lighting = 0;

	float3x3 WorldToTangentSpace;
	WorldToTangentSpace[0] = normalize(IN.Tangent);
	WorldToTangentSpace[1] = normalize(IN.BiNormal);
	WorldToTangentSpace[2] = normalize(IN.Normal);

	float4 PointLight0Vector = float4(mul(WorldToTangentSpace, c_object_point_light0_position.xyz - IN.PositionWorld), c_object_point_light0_position.w);
	float4 PointLight0Rotation = float4(mul(WorldToTangentSpace, c_object_point_light0_rotation.xyz), c_object_point_light0_rotation.w);

	float3 NormalBump = (2.0 * (tex2D( TexSampler4, IN.DiffuseMultiUV.xy ).xyz)) - 1.0;

	float3 VectorNorm;
	float2 PointAttenuation;

	//--VS Conversion--//
	//dp3 r4.w, r4, r4
	//rsq r10.w, r4.w
	//mul r4.xyz, r4, r10.w
	VectorNorm = normalize(PointLight0Vector.xyz);

	//Distance attenuation?
	//mad r8.x, r4.w, -c[15+0].w, v4.w
	PointAttenuation.x = (dot(PointLight0Vector.xyz, PointLight0Vector.xyz) * -PointLight0Vector.w) + IN.DetailUV.w;
	//dp3 r9.x, r4, r1
	PointAttenuation.x *= dot(NormalBump, VectorNorm);

	//Spotlight falloff
	//dp3 r10.w, r4, -c[15+1]
	//mul r10.w, r10.w, c[15+1].w
	//add r8.z, r10.w, c[15+2].w
	PointAttenuation.x *= (dot(VectorNorm, PointLight0Rotation.xyz) * PointLight0Rotation.w) + c_object_point_light0_colour.w;
	//--VS Conversion--//

	PointAttenuation = clamp(PointAttenuation, 0, 1);

	Lighting.xyz += (PointAttenuation.x * c_object_point_light0_colour.xyz);
I've cut out some code from this for multiple lights/reflections and such but I think the posted code is valid. The problems i've been having are better seen in pictures :). Spotlight pointing from the players left shoulder to the vehicle: Photobucket Point Light from the end of the weapon: Photobucket So i'm hoping someone can see where i've messed up and set me on the right track :).