Jump to content

  • Log In with Google      Sign In   
  • Create Account

Husbjörn

Member Since 27 Jan 2014
Offline Last Active Sep 26 2016 12:11 PM

Posts I've Made

In Topic: Clarifications on sharing a 2d texture between D3D9Ex and D3D11

28 August 2016 - 01:21 PM

I see... indeed, IDXGIKeyedMutex snuck into my D3D9 project through the Windows 7 API rather than D3D9 itself, leading me to assume it was available. 

Thanks for pointing that out, as well as the links! Now all that is left is to wait for the next weekend so that I can dig into it. :)


In Topic: Clarifications on sharing a 2d texture between D3D9Ex and D3D11

28 August 2016 - 10:49 AM

What is DXGI_FORMAT_A8G8B8R8_UNORM? As far as I can tell, it does not exist. The D3D11 equivalent of D3DFMT_A8R8G8B8 is DXGI_FORMAT_B8G8R8A8_UNORM.

Ah that's a typo haha.

I cut the relevant part out of a function, it was originally an argument identifier, so I thought I'd replace it with DXGI_FORMAT_B8G8R8A8_UNORM (which is indeed what it should have been) for clarity.


In Topic: Periodic shadow mapping artifacts

13 June 2016 - 08:48 AM

I'm going to bump this on the off chance that somebody who haven't already seen it would have any ideas.

Sorry if this isn't allowed.


In Topic: Are there any ways to eliminate / reduce the seam artifacts where dual parabo...

06 June 2016 - 05:20 PM

Hm I don't think that is what's causing the seams though. It isn't that things line up incorrectly between the hemispheres, but rather that a line of "nothingness" (the depth clear value) end up around it. The most obvious cause I can imagine is that the texels at the very edge of the projected circle blend together with the cleared outside. Since the shadow map is circular there will always be this "invalid" outline, so that's where I was thinking that maybe you could try to project a bit beyond a hemisphere and then only sample the middle 95% or something, such that you get a more proper "clamping" behaviour at the edge texels.

Another thing I thought of just now is that perhaps it would be better (if more expensive) to perform the paraboloid space transform in the pixel shader instead of the vertex shader? I imagine there will be more edges in the vertex shader approach, especially if the geometry isn't that detailed. I'm not sure if it would fix this particular issue though but it could be worth a try.


In Topic: SampleLevel not honouring integer texel offset

06 June 2016 - 01:08 AM

One way to make sure that you use the right version is to not use an import lib at all, and instead manually call LoadLibrary/GetProcAddress to get a pointer to the function you want to use from d3dcompiler_47.dll.

Ah yes, that should work; there are only two functions being imported from there so should be an easy thing, thanks.

 

I probably should see about making the switch and look further into what that will entail in regards to older OS support one of these days though.


PARTNERS