Blinn's approximation sucks according to me...
Phong
if( (*nearestObject)->material().shininess() > 0.0 ){ const cVector3<> viewDirection = ray.direction(); const cVector3<> reflection = directionToLight - Real(2.0) * dotLN * surfaceNormal; const Real dotVR = Dot( viewDirection, reflection ); if( dotVR > 0 ) { // Apply specular (phong) lightiing const Real specular = pow( dotVR, (*nearestObject)->material().shininess() ) * (*nearestObject)->material().specular(); color += specular * light->color() * shade; }}Blinn
if( (*nearestObject)->material().shininess() > 0.0 ){ const cVector3<> viewDirection = -ray.direction(); // should be directionToViewer? const cVector3<> sumLV = directionToLight + viewDirection; const cVector3<> halfway = sumLV / sumLV.length(); const Real dotHN = Dot( halfway, surfaceNormal ); if( dotHN > 0 ) { // Apply specular (Blinn's Phong) lightiing const Real specular = pow( dotHN, (*nearestObject)->material().shininess() ) * (*nearestObject)->material().specular(); color += specular * light->color() * shade; // NOTE: this is subject to attune }}Image compare, Phong to the left and Blinn to the right
VS
And the judge:
