DX11 plugin

[GoogleCodeExporter]

What steps will reproduce the problem?
Notes to self:

1: Mudlord needs to make a DX11 port of the DX9 plugin.
2: Mudlord needs to port the shaders and do fixed function emulation.
3: Do code on private SVN *before* committing to main nulldc tree.

What is the expected output? What do you see instead?

The expected output is: Proper Z with antialiasing, because of increased depth 
precision.

But instead I see: Improper Z sorting due to DX9 not supporting 32bit floating 
Z.

What version of the product are you using? What build? What plugins?

I use version: Latest nullDC Debug SVN with latest debug plugins

On what kind of system?
My system specifications are as follows:

Operating System: Windows 7 Ultimate x64
CPU: AMD Phenom II 550 @ 3.1Ghz
Video Card: ATI Radeon 5770 with 1GB RAM
Sound Card: Onboard 7.1 Via based.

Original issue reported on code.google.com by mudl...@tpg.com.au on 16 Aug 2010 at 1:57

[GoogleCodeExporter]

"Issue owner must be a project member" ... =S

Mail ain't enough it seems.

Original comment by KrossX3 on 16 Aug 2010 at 5:55

• Changed state: Accepted
• Added labels: Priority-Critical, Type-Enhancement
• Removed labels: Priority-Medium, Type-Defect

[GoogleCodeExporter]

will this D3D11 plugin have D3D10 fallbacks? D3D10 supports 32FPZ as well

Original comment by danialho...@gmail.com on 16 Aug 2010 at 9:02

[GoogleCodeExporter]

You can use Dx10 HW with D3D11. The float depth support should get rid of the 
z-fighting, and the Out of Order Transparency should get rid of the sorting 
problems. But the latter does require Dx11 HW I think.

Original comment by KrossX3 on 17 Aug 2010 at 12:57

[GoogleCodeExporter]

Actually, the current dx9 code is a mess-- you'l probably have more luck 
writing something form the beginning :p. Been wanting to do this for some time, 
but never really got around more than dx10 init code

Original comment by drkIIRaziel on 17 Aug 2010 at 10:58

[GoogleCodeExporter]

out of order transparency works on D3D10 as well.

Original comment by danialho...@gmail.com on 17 Aug 2010 at 6:03

[GoogleCodeExporter]

nvidia's depth peeling has been nominated to be a form of out of order 
transparency.

Original comment by danialho...@gmail.com on 18 Aug 2010 at 12:01

[GoogleCodeExporter]

heck, apparently Nvidia has supported this since the Geforce 3
http://developer.nvidia.com/object/order_independent_transparency.html

I expect it may be a D3D11 requirement, which im sure the 4xx cards will 
support from the go.

Original comment by danialho...@gmail.com on 18 Aug 2010 at 12:09

[GoogleCodeExporter]

http://www.slideshare.net/ohyecloudy/shaderx6-37-robust-orderindependent-transpa
rency-via-reverse-depth-peeling-in-directx-10

Original comment by danialho...@gmail.com on 18 Aug 2010 at 12:14

[GoogleCodeExporter]

Using depth peeling to achieve OIT? Might as well use software rendering 
instead. Same effect, same speed, but with s/w rendering being more accurate. :p

Original comment by Mr.PsyMan on 18 Aug 2010 at 1:36

[GoogleCodeExporter]

Reverse OIT is apparently nearly hitless.

Original comment by danialho...@gmail.com on 18 Aug 2010 at 5:01

[GoogleCodeExporter]

@drkRaz: Fine, doing it with MSVC2008. :P

Original comment by mudl...@tpg.com.au on 18 Aug 2010 at 5:50

[GoogleCodeExporter]

The pvr hardware implements a version of depth peeling

Original comment by drkIIRaziel on 18 Aug 2010 at 11:04

[GoogleCodeExporter]

Original comment by KrossX3 on 18 Aug 2010 at 11:47

• Changed state: Started

[GoogleCodeExporter]

indeed it does drk.

surely we can write a custom shader to emulate PVR GPU Depth peeling?

Original comment by danialho...@gmail.com on 19 Aug 2010 at 2:17

[GoogleCodeExporter]

We are salvaging what we can from the old D3D9 plugin, as some is still useful. 
Really, the shaders, rasterization etc, needs to be redone.

Original comment by mudl...@mudlord.info on 19 Aug 2010 at 2:26

[GoogleCodeExporter]

Original comment by mudl...@mudlord.info on 19 Aug 2010 at 2:30

[GoogleCodeExporter]

First initial commits.
Added code base.

Original comment by mudl...@mudlord.info on 7 Sep 2010 at 9:46

[GoogleCodeExporter]

Just saw the first inicial commit, really good!!!!! Keep the good work, one 
more to add their name to the scene :D

Original comment by telmo.qu...@gmail.com on 8 Sep 2010 at 2:12