Darío

Are you sure you updated your submodules after checking out the branch? ``` git submodule update --init --recursive ```

No worries, common mistake as git's default behavior when checking out branches is fairly silly.

Hey, thanks for the report! I think I've seen this issue in Mario Tennis as well, it seems the other pixels of the multisampling don't seem to accurately sample the...

> There is another problem which is sometimes visible in D3D12, the half transparent shadow which can be seen in Vulkan with MSAA 2x can also be sometimes seen in...

So far I've gathered the following information. - It works fine in NVIDIA D3D12. - It breaks in NVIDIA Vulkan. It also corrupts RenderDoc really badly. - It works fine...

Are you able to test if the issue shows up in Linux?

> I'll spin up a Linux machine and take a look, are you using the open source NVIDIA drivers or the proprietary ones? I'm not using either as my Linux...

Seems pretty consistent across systems and OS at least. I've verified [turning off sample locations](https://github.com/rt64/rt64/blob/051af43bd857fbfa75e174b1f1276c0b5ad869b0/src/render/rt64_raster_shader.cpp#L325) (which introduces other rendering artifacts that it attempts to fix) does fix the issue at...

> Here's some confirmation for that "ghost shadow" issue (this is under D3D12 on Windows): https://github.com/rt64/rt64/assets/169589138/1aa4ec92-1969-4919-b4ae-ab1c7823c928 Thanks, I'll see if I can reproduce it in a different spot than the...

> Here's some confirmation for that "ghost shadow" issue (this is under D3D12 on Windows): https://github.com/rt64/rt64/assets/169589138/1aa4ec92-1969-4919-b4ae-ab1c7823c928 So you're right, when I did go to this spot, I got the exact...