Issues with MSAA - why set the render window to have MSAA!??

[al2950]

I am actually digging around an issue with using open VR which spits out the following 'hidden' error, which actually gets caught by Ogre at a different point;

Code: Select all

ERROR: ID3D11DeviceContext::CopySubresourceRegion: If CopySubresourceRegion is used with Multisampled or D3D11_BIND_DEPTH_STENCIL Resources, then the source and destination Subresource must have the same dimensions.

Anyway this got me thinking about how I set up my MSAA render targets. In short I always resolve my MSAA render targets early, before any post processing, however in Ogre MSAA settings are always based on the final render target settings. This seems, well wrong. Am I missing something?? I appreciated this may be a hang over from old school FFP.

Am i wrong in saying that in my setup the final render target should NOT be MSAA, but i should be able to setup scene render targets to have MSAA which I explicitly resolve later before the final render target?

PS. I am using Ogre 2.1, not sure if Ogre 2.2 may help with this issue...

[dark_sylinc]

From the December 2017 report:

Another change for example that in Ogre 2.1 all render textures in the compositor default to using MSAA if the main target is using MSAA. In Ogre 2.2; we default to not using MSAA unless told otherwise. We found out most textures do not need MSAA because they’re postprocessing FXs or similar; thus the MSAA is only a waste of memory and performance by doing useless resolves. Therefore only a few render textures actually need MSAA. This saves a lot of GPU RAM and some performance.

So in short, you're right. You're also right that 2.1's handling of MSAA is very old school (FFP era) and treats it like an esoteric magic black box.

[al2950]

Thanks for the confirmation.... I guess its about time I get 2.2!

Quick question re 2.2, I shall assume cubemap arrays are supported, but does this mean that we have per pixel reflection probes, or is it still per object?

[dark_sylinc]

Quick question re 2.2, I shall assume cubemap arrays are supported, but does this mean that we have per pixel reflection probes, or is it still per object?

That's one of the things I want to tackle next, in the short term (for the moment it's still per object, like 2.1)

[xrgo]

Quick question re 2.2, I shall assume cubemap arrays are supported, but does this mean that we have per pixel reflection probes, or is it still per object?

That's one of the things I want to tackle next, in the short term (for the moment it's still per object, like 2.1)

I will extremely love that, please please