Though it said that it is "1:1 implementation/copy of the GLSL algorithm".Looks like some calculation is different from original paper,like the weight calculation part.And in the renderloop,at the first time is enter loop,"glUniform1f(filter_c_phiUniform, 1.0f / i * c_phi0);" should produce a divide zero problem.I am not sure if that is by design.The kernel part is also different from the paper.In the paper is says it should based on B3 spine.I am not sure if I misunderstand the paper or it is bug.But anyway,thank you for sharing the code.
Though it said that it is "1:1 implementation/copy of the GLSL algorithm".Looks like some calculation is different from original paper,like the weight calculation part.And in the renderloop,at the first time is enter loop,"glUniform1f(filter_c_phiUniform, 1.0f / i * c_phi0);" should produce a divide zero problem.I am not sure if that is by design.The kernel part is also different from the paper.In the paper is says it should based on B3 spine.I am not sure if I misunderstand the paper or it is bug.But anyway,thank you for sharing the code.