Original Post
Has anyone been able to successfully implement linear depth projection similar to the one described at this link? I understand the concept and have incorporated the authors suggestion of pre-multiplying by the homogoneous clip space w (to negate the post-transform divide) and dividing by the far plane distance (separately in both the vshader and projection matrix) in a demo program but it doesn't appear to work. I end up getting very weird depth sorting artifacts and even weirder geometry popping issues (as if they were clipped or early z-rejected, due to improper depth). Any insights would be greatly appreciated. Thanks!