Pixel Density vs SuperSampling vs SparseGrid SS?

[Borsch]

I've read time and time again that higher Pixel Density settings that improve clarity of VR are really SuperSampling settings. If that is so, is Sparse Grid Super sampling a better alternative (for Nvidia users at least) to higher Pixel Density? Sparse Grid is significantly faster than SuperSampling in monitor games and gives pretty much the same effect.

 

[FlukeRogi]

Yeah, they're talking about the same thing, though in most tools (Oculus Tray Tool for example), it's generally referred to as super-sampling. I don't know enough about the technical side of how sparse grid works, but maybe "pretty much" might not be good enough when the rendered image is blown up across you field of view.

nVidia's current solutions for speeding things up are more tailored towards actual headsets. There's Lens Matched Shading, which basically only renders pixels that fall within the distorted view that is rendered for the HMD.

Then there's Multi-Res Shading, which breaks up the rendered image into zones and renders the centre of your view at a higher resolution than at the edges. There aren't many games using it atm, though Batman Arkham VR has some excellent MRS options built in that you can play around with.

There's also single pass stereo, which allows all the geometry to be drawn in a single pass, rather than doing so sequentially for the left and right viewports.

 

[Borsch]

THanks for the detailed answer! I'll need to experiment with SparseGridSS then, as other methods need native game support it seems, eg there is no Lens Matched Shading in nvidia options where is SparseGrid is available in all games through nvidia nspector.

EDIT: Yep, SparseGridSS works nicely for VR - for example it cleans up fence shimmer in Assetto Corsa.