> > Thanks, that looks great. I'd been meaning to check it out myself since it went up > on > > Shadertoy and now I don't have to. I really love the aperture grille effect on this > > one. > > Yeah. Jesus. That aperture grille effect looks fucking phenomenal. I'm just slightly > disappointed that it can't use a texture to define the aperture grille pattern. It > damn well puts mine to shame, that's for sure. > > The only real issue with the aperture grille effect is that it's not an accurate > model of how it actually worked on a real CRT. If you look at the shader code, it's > very clear that he has a 1:1 correspondence between the aperture and the pixels, > which is something that was rarely the case on actual CRTs. In reality, the aperture > grille has a fixed physical resolution in the tube, whereas the incoming signal can > be whatever resolution it wants. > > What I would very much like is to eventually make the HLSL code in MAME/MESS generate > the aperture grille effect at the highest resolution possible, then do a > gamma-correct downsample to avoid the horrible colored stripe effect you get when you > crank the aperture grille intensity up. Ideally that would give both a proper > physical modelling of the aperture grille as well as avoid color beating, but that'd > take more time than I'm up for putting in at the moment.
Yeah. On my Surface Pro 3, the resolution is high enough and screen small enough that the aperture grille effect is nearly lost. Something that doesn't map directly to the output pixels would be nice. But like you said, 1:1 mapping avoids the sampling issues for now, so it's a good start. Especially since the grille effect does look so nice.