Fix T44398: Compositing displace node makes image fuzzy with zero displacement
EWA filtering with zero derivatives is introducing some fuzzyness into the image. Currently solved by using regular sampling for cases when derivatives are zero, which should also make compo faster in that areas. Still need to look into checking if EWA filter can be tweaked in a way so no fuzzyness is introduced.
This commit is contained in:
@@ -55,9 +55,13 @@ void DisplaceOperation::executePixelSampled(float output[4], float x, float y, P
|
||||
float uv[2], deriv[2][2];
|
||||
|
||||
pixelTransform(xy, uv, deriv);
|
||||
|
||||
/* EWA filtering (without nearest it gets blurry with NO distortion) */
|
||||
this->m_inputColorProgram->readFiltered(output, uv[0], uv[1], deriv[0], deriv[1], COM_PS_BILINEAR);
|
||||
if(is_zero_v2(deriv[0]) && is_zero_v2(deriv[1])) {
|
||||
this->m_inputColorProgram->readSampled(output, uv[0], uv[1], COM_PS_BILINEAR);
|
||||
}
|
||||
else {
|
||||
/* EWA filtering (without nearest it gets blurry with NO distortion) */
|
||||
this->m_inputColorProgram->readFiltered(output, uv[0], uv[1], deriv[0], deriv[1], COM_PS_BILINEAR);
|
||||
}
|
||||
}
|
||||
|
||||
bool DisplaceOperation::read_displacement(float x, float y, float xscale, float yscale, const float origin[2], float &r_u, float &r_v)
|
||||
|
Reference in New Issue
Block a user