Re: Using setNumSamples() doesn't work
Posted by Demoscene Passivist on Nov 26, 2010; 11:38am
URL: https://forum.jogamp.org/Using-setNumSamples-doesn-t-work-tp1427769p1972286.html
>how about performance ?
It's ok'isch. I would estimate 2xSSAA is performance wise roughly equivalent to 16xMSAA and look like 8xMSAA+.
Taking into account that SSAA isn't just smoothing the polygon edges but also the textures (thereby reducing texture aliasing effects), I guess its a valid tradeoff wich could be considered.
Other than that if u do some kind of deferred shading stuff, SSAA is the only way to go anyway.
>looks like a glsl toolbox candidate.
Was thinking about that too. As u can see I already have encapsulated it a bit into some kind of helper class, but it's still not usable as simple opt-in out-of-the-box. Maybe if I should thow some more brainz at this I could come up with a smoother solution API wise. Or some else has a nice idea about this ?
URL: https://forum.jogamp.org/Using-setNumSamples-doesn-t-work-tp1427769p1972286.html
>how about performance ?
It's ok'isch. I would estimate 2xSSAA is performance wise roughly equivalent to 16xMSAA and look like 8xMSAA+.
Taking into account that SSAA isn't just smoothing the polygon edges but also the textures (thereby reducing texture aliasing effects), I guess its a valid tradeoff wich could be considered.
Other than that if u do some kind of deferred shading stuff, SSAA is the only way to go anyway.
>looks like a glsl toolbox candidate.
Was thinking about that too. As u can see I already have encapsulated it a bit into some kind of helper class, but it's still not usable as simple opt-in out-of-the-box. Maybe if I should thow some more brainz at this I could come up with a smoother solution API wise. Or some else has a nice idea about this ?
| Free forum by Nabble | Edit this page |