c++directx-11vertex-shadernormalspixel-shader

Normal map overrides smooth edges in DirectX 11


My object is a smooth barrel, with only the color and the original normals it looks like this:

barrel with the original normals

When I try to add the normal detail via a texture the smooth normals get overwritten like so:

barrel with the overwritten normals

Is there a better way to merge the new normal texture without overriding the smooth edges?

The code is mostly from the rastertek tutorials found here https://www.rastertek.com/tutdx11.html

The Vertex Shader

cbuffer MatrixBuffer
{
    matrix worldMatrix;
    matrix viewMatrix;
    matrix projectionMatrix;
};

struct VertexInputType
{
    float4 position : POSITION;
    float2 tex : TEXCOORD0;
    float3 normal : NORMAL;
    float3 tangent : TANGENT;
    float3 binormal : BINORMAL;
};

struct PixelInputType
{
    float4 position : SV_POSITION;
    float2 tex : TEXCOORD0;
    float3 normal : NORMAL;
    float3 tangent : TANGENT;
    float3 binormal : BINORMAL;
};

PixelInputType NormalMapVertexShader(VertexInputType input)
{
    PixelInputType output;
    

    // Change the position vector to be 4 units for proper matrix calculations.
    input.position.w = 1.0f;

    // Calculate the position of the vertex against the world, view, and projection matrices.
    output.position = mul(input.position, worldMatrix);
    output.position = mul(output.position, viewMatrix);
    output.position = mul(output.position, projectionMatrix);
    
    // Store the texture coordinates for the pixel shader.
    output.tex = input.tex;
    
    // Calculate the normal vector against the world matrix only and then normalize the final value.
    output.normal = mul(input.normal, (float3x3)worldMatrix);
    output.normal = normalize(output.normal);

    // Calculate the tangent vector against the world matrix only and then normalize the final value.
    output.tangent = mul(input.tangent, (float3x3)worldMatrix);
    output.tangent = normalize(output.tangent);

    // Calculate the binormal vector against the world matrix only and then normalize the final value.
    output.binormal = mul(input.binormal, (float3x3)worldMatrix);
    output.binormal = cross(output.tangent, output.normal);



    return output;
}

The Pixel Shader

Texture2D shaderTextures[2];
SamplerState SampleType;

cbuffer LightBuffer
{
    float4 diffuseColor;
    float3 lightDirection;
};

struct PixelInputType
{
    float4 position : SV_POSITION;
    float2 tex : TEXCOORD0;
    float3 normal : NORMAL;
    float3 tangent : TANGENT;
    float3 binormal : BINORMAL;
};


float4 NormalMapPixelShader(PixelInputType input) : SV_TARGET
{
    float4 textureColor;
    float4 bumpMap;
    float3 bumpNormal;
    float3 lightDir;
    float lightIntensity;
    float4 color;


    // Sample the texture pixel at this location.
    textureColor = shaderTextures[0].Sample(SampleType, input.tex);

    // Sample the pixel in the bump map.
    bumpMap = shaderTextures[1].Sample(SampleType, input.tex);

    // Expand the range of the normal value from (0, +1) to (-1, +1).
    bumpMap = (bumpMap * 2.0f) - 1.0f;


    // Calculate the normal from the data in the bump map.
    bumpNormal = (bumpMap.x * input.tangent) + (bumpMap.y * input.binormal) + (bumpMap.z * input.normal);

    // Normalize the resulting bump normal.
    bumpNormal = normalize(bumpNormal);

    // Invert the light direction for calculations.
    lightDir = -lightDirection;




    // Calculate the amount of light on this pixel based on the bump map normal value.
    lightIntensity = saturate(dot(bumpNormal, lightDir));


    // Determine the final diffuse color based on the diffuse color and the amount of light intensity.
    color = saturate(diffuseColor * lightIntensity);

    // Combine the final bump light color with the texture color.
    color = color * textureColor;



    return color;
}

Solution

  • The issue is this formula:

    // Calculate the normal from the data in the bump map.
    bumpNormal = (bumpMap.x * input.tangent) + (bumpMap.y * input.binormal) + (bumpMap.z * input.normal);
    

    The better approach is to create a matrix for the local tangent space, and then transform the local normal from the texture.

    // Given a local normal, transform it into a tangent space given by surface normal and tangent
    float3 PeturbNormal(float3 localNormal, float3 surfaceNormalWS, float3 surfaceTangentWS)
    {
        float3 normal = normalize(surfaceNormalWS);
        float3 tangent = normalize(surfaceTangentWS);
        float3 binormal = cross(normal, tangent);     // reconstructed from normal & tangent
        float3x3 TBN = { tangent, binormal, normal }; // world "frame" for local normal 
    
        return mul(localNormal, TBN);                // transform to local to world (tangent space)
    }
    
    
    // This function converts the R and G channels from a normal texture map
    // Assumes the input data is "UNORM" (i.e. needs a x2 bias to get back -1..1)
    // Also reconstructs the B channel in case the normal map was compressed as BC5_UNORM
    float3 TwoChannelNormalX2(float2 normal)
    {
        float2 xy = 2.0f * normal - 1.0f;
        float z = sqrt(1 - dot(xy, xy));
        return float3(xy.x, xy.y, z);
    }
    ...
    
        // Before lighting, peturb the surface's normal by the one given in normal map.
        float3 localNormal = TwoChannelNormalX2(NormalTexture.Sample(Sampler, pin.TexCoord).xy);
        
        float3 N = PeturbNormal( localNormal, pin.NormalWS, pin.TangentWS);
    

    Taken from DirectX Tool Kit's NormalMap