i was trying to render to a texture to use it for post-processing. I've read a lot of tutorials about this stuff und have tried it for myself now, but apparently i am doing something wrong.
So first of all these are the error message im getting over and over:
D3D11: ERROR: ID3D11DeviceContext::OMSetRenderTargets: The RenderTargetView at slot 0 is not compatable with the DepthStencilView. DepthStencilViews may only be used with RenderTargetViews if the effective dimensions of the Views are equal, as well as the Resource types, multisample count, and multisample quality. The RenderTargetView at slot 0 has (w:400,h:300,as:1), while the Resource is a Texture2D with (mc:1,mq:0). The DepthStencilView has (w:1280,h:720,as:1), while the Resource is a Texture2D with (mc:1,mq:0). D3D11_RESOURCE_MISC_TEXTURECUBE factors into the Resource type, unless GetFeatureLevel() returns D3D_FEATURE_LEVEL_10_1 or greater. [ STATE_SETTING ERROR #388: OMSETRENDERTARGETS_INVALIDVIEW ]
D3D11: ERROR: ID3D11DeviceContext::Draw: Rasterization Unit is enabled (PixelShader is not NULL or Depth/Stencil test is enabled and RasterizedStream is not D3D11_SO_NO_RASTERIZED_STREAM) but position is not provided by the last shader before the Rasterization Unit. [ EXECUTION ERROR #362: DEVICE_DRAW_POSITION_NOT_PRESENT ]
And:
D3D11: ERROR: ID3D11DeviceContext::Draw: Vertex Shader - Geometry Shader linkage error: Signatures between stages are incompatible. The input stage requires Semantic/Index (RADIUS,0) as input, but it is not provided by the output stage. [ EXECUTION ERROR #342: DEVICE_SHADER_LINKAGE_SEMANTICNAME_NOT_FOUND ]
D3D11: ERROR: ID3D11DeviceContext::Draw: Vertex Shader - Geometry Shader linkage error: Signatures between stages are incompatible. The input stage requires Semantic/Index (TEXTUREINDEX,0) as input, but it is not provided by the output stage. [ EXECUTION ERROR #342: DEVICE_SHADER_LINKAGE_SEMANTICNAME_NOT_FOUND ]
D3D11: ERROR: ID3D11DeviceContext::Draw: Geometry Shader - Pixel Shader linkage error: Signatures between stages are incompatible. The input stage requires Semantic/Index (POSITION,0) as input, but it is not provided by the output stage. [ EXECUTION ERROR #342: DEVICE_SHADER_LINKAGE_SEMANTICNAME_NOT_FOUND ]
D3D11: ERROR: ID3D11DeviceContext::OMSetRenderTargets: The RenderTargetView at slot 0 is not compatable with the DepthStencilView. DepthStencilViews may only be used with RenderTargetViews if the effective dimensions of the Views are equal, as well as the Resource types, multisample count, and multisample quality. The RenderTargetView at slot 0 has (w:400,h:300,as:1), while the Resource is a Texture2D with (mc:1,mq:0). The DepthStencilView has (w:1280,h:720,as:1), while the Resource is a Texture2D with (mc:1,mq:0). D3D11_RESOURCE_MISC_TEXTURECUBE factors into the Resource type, unless GetFeatureLevel() returns D3D_FEATURE_LEVEL_10_1 or greater. [ STATE_SETTING ERROR #388: OMSETRENDERTARGETS_INVALIDVIEW ]
I am pretty sure that i am misunterstanding a major aspect of this rendertarget switching and stuff.
When my DirectX Device is created i am also creating this:
So i am creating a ShaderResourceView and a RenderTargetView aswell as a cube with a vertex buffer to draw this cube where the texture will be drawn on.
Now in every frame update after ALL the other stuff was drawn im using this:
I have not explicitly created a depthStencilView which i would guess is my main error here. But i have no idea how to create something like this right.
I am pretty much confused about how this all is working in the background...maybe someone cares to explain this.
Am i doing anything right here or is this just wrong?
I see several problems with posted snipsets.
1) If you render to texture, that is smaller than your standard render target, you must change viewport to texture dimensions. (maybe last error is due to this)
2) Before you render to texture, save your targets. After you render to texture, dont forget to unbind render targets from device and restore original ones (aka. classic "screen")
3) If you render "classic way" to the screen, can you see some errors ? Because errors with shaders are not render-to-texture relevant. In DX11 rendering geometry to texture is exactly the same as classic rendering, except for render targets changes.
Hey, thanks for your advices.
I've reworked my whole Render to Texture process based on this tutorial and it gives me no errors and it seems like that i can work with the texture.
The only problem left now ist, that it seems like i cannot draw a simple square to draw my texture on anymore. This is how it looks like at the moment (with a black/ white shader): http://imgur.com/HrKXR
this is the code AFTER RenderToTexture() and RenderScene():
so basically it splits the whole thing in 4 parts. There are some things i've noticed aswell.
1. The Texture on the square seems to be exaclty mirrored (this could be fixed through the shader though)
2. The 4 tiles are not present at the beginning - i have to move my camera just a little bit then it appears.
3. Its disappearing in random intervals.
thanks for your help. I really appreciate it. Your hint with the TexCoords was a hit. I was using an old calculation for them. Now i can render the whole texture over my scene with a black and white shader. I have a few questions / problems left though:
1. I understand that for the coordinate system for rendering 0,0 is the center and -1,-1 is the lower left corner, isnt it? But what about the texture coordinates. If i understand the numbers correctly there 0,0 is the upper left corner? Why is there such a difference?
2. My shader seems not to be active at first startup , i have to move my camera a bit before it comes active.
3. The shader effect seems to disappear from time to time (random)
I have tried to disable the depth for the Stencil with something like this:
now i have researched some stuff and there are some open questions in my mind which i hope some of you can answer me. I am now trying to downsample the rendered texture to the 4th of the original size. After that i want to use a bloom effect or something and (upsample?) it again to merge it with the original rendered scene.
Now my question is fairly simple: how can i let my original texture (full screen size) walk through some shader passes so at the end i will have a downsampled texture with effects and the original texture to work with?
i dont understand how to work with multiple passes. How do i pass the texture from one pass to another and at the end merge the original one with the edited one?
Thanks in advance
You basically create two rendertargets. Size doesn't matter - they just need to be big enough for the largest (typically your screen dimensions).
So, in pass 1 you draw from target 0 to target 1 as a half-sized quad (the size of the quad doesn't have to be the same as your render target size).
In pass 2 you draw from 1 to 0 as a half-sized quad (halving again the half size from above).
Repeat until you get it down to the size you want, flipping the render targets each pass.
After the final pass you set back your default render target, then draw from the target set in the final pass to this.
Edit: silly me, I forgot that with D3D11 you can autogen mipmaps from a render target. I've never used this, but it does present another option for you that may be faster.
Direct3D has need of instancing, but we do not. We have plenty of glVertexAttrib calls.
I think i understood this. But i cant really image how to draw from one target to another? is it a simple draw() call with a fullscreen quad textured by the texture i want to pass and the render target i want to use next?
