DirectX Clip space texture coordinates - c++

Okay first up I am using:
DirectX 10
C++
Okay this is a bit of a bizarre one to me, I wouldn't usually ask the question, but I've been forced by circumstance. I have two triangles (not a quad for reasons I wont go into!) full screen, aligned to screen through the fact they are not transformed.
In the DirectX vertex declaration I am passing a 3 component float (Pos x,y,z) and 2 component float (Texcoord x,y). Texcoord z is reserved for texture2d arrays, which I'm currently defaulting to 0 in the the pixel shader.
I wrote this to achieve the simple task:
float fStartX = -1.0f;
float fEndX = 1.0f;
float fStartY = 1.0f;
float fEndY = -1.0f;
float fStartU = 0.0f;
float fEndU = 1.0f;
float fStartV = 0.0f;
float fEndV = 1.0f;
vmvUIVerts.push_back(CreateVertex(fStartX, fStartY, 0, fStartU, fStartV));
vmvUIVerts.push_back(CreateVertex(fEndX, fStartY, 0, fEndU, fStartV));
vmvUIVerts.push_back(CreateVertex(fEndX, fEndY, 0, fEndU, fEndV));
vmvUIVerts.push_back(CreateVertex(fStartX, fStartY, 0, fStartU, fStartV));
vmvUIVerts.push_back(CreateVertex(fEndX, fEndY, 0, fEndU, fEndV));
vmvUIVerts.push_back(CreateVertex(fStartX, fEndY, 0, fStartU, fEndV));
IA Layout: (Update)
D3D10_INPUT_ELEMENT_DESC ieDesc[2] = {
{ "POSITION", 0, DXGI_FORMAT_R32G32B32_FLOAT, 0, 0, D3D10_INPUT_PER_VERTEX_DATA, 0 },
{ "TEXCOORD", 0, DXGI_FORMAT_R32G32B32_FLOAT, 0,12, D3D10_INPUT_PER_VERTEX_DATA, 0 }
};
Data reaches the vertex shader in the following format: (Update)
struct VS_INPUT
{
float3 fPos :POSITION;
float3 fTexcoord :TEXCOORD0;
}
Within my Vertex and Pixel shader not a lot is happening for this particular draw call, the pixel shader does most of the work sampling from a texture using the specified UV coordinates. However, this isn't working quite as expected, it appears that I am getting only 1 pixel of the sampled texture.
The workaround was in the pixel shader to do the following: (Update)
sampler s0 : register(s0);
Texture2DArray<float4> meshTex : register(t0);
float4 psMain(in VS_OUTPUT vOut) : SV_TARGET
{
float4 Color;
vOut.fTexcoord.z = 0;
vOut.fTexcoord.x = vOut.fPosObj.x * 0.5f;
vOut.fTexcoord.y = vOut.fPosObj.y * 0.5f;
vOut.fTexcoord.x += 0.5f;
vOut.fTexcoord.y += 0.5f;
Color = quadTex.Sample(s0, vOut.fTexcoord);
Color.a = 1.0f;
return Color;
}
It was also worth noting that this worked with the following VS out struct defined in the shaders:
struct VS_OUTPUT
{
float4 fPos :POSITION0; // SV_POSITION wont work in this case
float3 fTexcoord :TEXCOORD0;
}
Now I have a texture that's stretched to fit the entire screen, both triangles already cover this, but why did the texture UV's not get used as expected?
To clarify I am using a point sampler and have tried both clamp and wrapping UV.
I was a bit curious and found a solution / workaround mentioned above, however I'd prefer not to have to do it if anyone knows why it's happening?

What semantics are you specifying for your vertex-type? Are they properly aligned with your vertices and also your shader? If you are using a D3DXVECTOR4, D3DXVECTOR3 setup (as shown in your VS code) this could be a problem if your CreateVertex() returns a D3DXVECTOR3, D3DXVECTOR2 struct.
It would be reassuring to see your pixel-shader code too.

Okay well, for one, the texture coordinates outside of 0..1 range get clamped. I made the mistake of assuming that by going to clip space -1 to +1 that the texture coordinates would be too. This is not the case, they still go from 0.0 to 1.0.
The reason why the code in the pixel shader worked, was because it was using the clip space x,y,z coordinates to automatically overwrite these texture coordinates; an oversight on my part. However, the pixel shader code results in texture stretch on a full screen 'quad', so it might be useful to someone somewhere ;)

Related

Cubemap texturing issue (D3D11, C++)

I have a texture problem with the cubemap I'm rendering and can't seem to figure it out. I've generated a cube map with direct x's texture tools and then read it using
D3DX11CreateShaderResourceViewFromFile(device, L"cubemap.dds", 0, 0, &fullcubemap, 0);
The cubemap texture is not high quality at all and it looks really stretched/distorted. I can definitely tell that the images used for the cubemap match correctly, but it's not great at all at the moment
I'm not sure why this is happening. Is it because my textures are too large/small or is it something else? If it's due to the size of the textures, what is a recommended texture size? I am using a sphere for the cubemap not a cube.
Edit:
Shader:
cbuffer SkyboxConstantBuffer {
float4x4 world;
float4x4 view;
float4x4 projection;
};
TextureCube gCubeMap;
SamplerState samTriLinearSam {
Filter = MIN_MAG_MIP_LINEAR;
AddressU = Wrap;
AddressV = Wrap;
};
struct VertexIn {
float4 position : POSITION;
};
struct VertexOut {
float4 position : SV_POSITION;
float4 spherePosition : POSITION;
};
VertexOut VS(VertexIn vin) {
VertexOut vout = (VertexOut)0;
vin.position.w = 1.0f;
vout.position = mul(vin.position, world);
vout.position = mul(vout.position, view);
vout.position = mul(vout.position, projection);
vout.spherePosition = vin.position;
return vout;
}
float4 PS(VertexOut pin) : SV_Target {
return gCubeMap.Sample(samTriLinearSam, pin.spherePosition);//float4(1.0, 0.5, 0.5, 1.0);
}
RasterizerState NoCull {
CullMode = None;
};
DepthStencilState LessEqualDSS {
DepthFunc = LESS_EQUAL;
};
technique11 SkyTech {
pass p0 {
SetVertexShader(CompileShader(vs_4_0, VS()));
SetGeometryShader(NULL);
SetPixelShader(CompileShader(ps_4_0, PS()));
SetRasterizerState(NoCull);
SetDepthStencilState(LessEqualDSS, 0);
}
}
Draw:
immediateContext->OMSetRenderTargets(1, &renderTarget, nullptr);
XMMATRIX sworld, sview, sprojection;
SkyboxConstantBuffer scb;
sview = XMLoadFloat4x4(&_view);
sprojection = XMLoadFloat4x4(&_projection);
sworld = XMLoadFloat4x4(&_world);
scb.world = sworld;
scb.view = sview;
scb.projection = sprojection;
immediateContext->IASetIndexBuffer(cubeMapSphere->getIndexBuffer(), DXGI_FORMAT_R32_UINT, 0);
ID3D11Buffer* vertexBuffer = cubeMapSphere->getVertexBuffer();
//ID3DX11EffectShaderResourceVariable * cMap;
////cMap = skyboxShader->GetVariableByName("gCubeMap")->AsShaderResource();
immediateContext->PSSetShaderResources(0, 1, &fullcubemap);//textures
//cMap->SetResource(fullcubemap);
immediateContext->IASetVertexBuffers(0, 1, &vertexBuffer, &stride, &offset);
immediateContext->VSSetShader(skyboxVertexShader, nullptr, 0);
immediateContext->VSSetConstantBuffers(0, 1, &skyboxConstantBuffer);
immediateContext->PSSetConstantBuffers(0, 1, &skyboxConstantBuffer);
immediateContext->PSSetShader(skyboxPixelShader, nullptr, 0);
immediateContext->UpdateSubresource(skyboxConstantBuffer, 0, nullptr, &scb, 0, 0);
immediateContext->DrawIndexed(cubeMapSphere->getIndexBufferSize(), 0, 0);
Initially I was planning to use this snippet to update the TextureCube variable in the shader
ID3DX11EffectShaderResourceVariable * cMap;
cMap = skyboxShader->GetVariableByName("gCubeMap")->AsShaderResource();
cMap->SetResource(fullcubemap);
But it seems that has no effect, and in fact, without the following line, the sphere I'm using for the cubemap textures with a texture used with another object in the scene, so perhaps there's something going on here? I'm not sure what though.
immediateContext->PSSetShaderResources(0, 1, &fullcubemap);//textures
Edit: Probably not the above, realised that if this wasn't updated, the old texture would be applied as it's never wiped after each draw.
Edit: Tried the cubemap with both a sphere and a cube, still the same texture issue.
Edit: Tried loading the shader resource view differently
D3DX11_IMAGE_LOAD_INFO loadSMInfo;
loadSMInfo.MiscFlags = D3D11_RESOURCE_MISC_TEXTURECUBE;
ID3D11Texture2D* SMTexture = 0;
hr = D3DX11CreateTextureFromFile(device, L"cubemap.dds",
&loadSMInfo, 0, (ID3D11Resource**)&SMTexture, 0);
D3D11_TEXTURE2D_DESC SMTextureDesc;
SMTexture->GetDesc(&SMTextureDesc);
D3D11_SHADER_RESOURCE_VIEW_DESC SMViewDesc;
SMViewDesc.Format = SMTextureDesc.Format;
SMViewDesc.ViewDimension = D3D11_SRV_DIMENSION_TEXTURECUBE;
SMViewDesc.TextureCube.MipLevels = SMTextureDesc.MipLevels;
SMViewDesc.TextureCube.MostDetailedMip = 0;
hr = device->CreateShaderResourceView(SMTexture, &SMViewDesc, &fullcubemap);
Still produces the same output, any ideas?
Edit: Tried increasing the zfar distance and the texture remains the exact same no matter what value I put.
Example with second texture with increased view distance.
This texture is used on another object in my scene and comes out fine.
Edit: I have been trying to mess with the scaling of the texture/object
To achieve this I used
vin.position = vin.position * 50.0f;
This is beginning to look sort of like how it should, however, when I turn my camera angle, the image disappears so I obviously know this isn't correct, but if I could just scale the image per pixel or per vertex properly, I'm sure I could get the end result.
Edit:
I can confirm the cubemap is rendering correctly, I was ignoring the view/projection space and just using world and managed to get this, which is the high quality image i'm after, just not correct. Yes the faces are incorrect, but I'm not fussed about that now, it's easy enough to swap them around, I just need to get it rendering with this quality, in the correct space.
When in camera space does it take into account whether or not it's the outside/inside of the sphere? If my textures were over the outside of the sphere and I have the view from the inside, it's not going to look the same?
Issue is with your texture size, its small, you are applying it on larger surface, Make larger textures with more pixels
Its confirm that zfar and scaling has nothing to do with it.
Finally found the issue, silly mistake.
scb.world = XMMatrixTranspose(sworld);
scb.view = XMMatrixTranspose(sview);
scb.projection = XMMatrixTranspose(sprojection);

HLSL: SV_Position, why/how from float3 to float4?

I'm just at the very very beginning of learning shaders/hlsl etc., so please excuse the probably stupid question.
I'm following Microsoft's DirectX Tutorials (Tutorial (link) , Code (link) ). As far as I understand, they're defining POSITION as 3-element array of float values:
// Define the input layout
D3D11_INPUT_ELEMENT_DESC layout[] =
{
{ "POSITION", 0, DXGI_FORMAT_R32G32B32_FLOAT, 0, 0, D3D11_INPUT_PER_VERTEX_DATA, 0 },
{ "COLOR", 0, DXGI_FORMAT_R32G32B32A32_FLOAT, 0, 12, D3D11_INPUT_PER_VERTEX_DATA, 0 },
};
Makes sense of course, each vertex position has 3 float values: x, y, and z. But now when looking at the Vertex shader, position is suddently of type float4, not of float3:
//--------------------------------------------------------------------------------------
// Vertex Shader
//--------------------------------------------------------------------------------------
VS_OUTPUT VS( float4 Pos : POSITION, float4 Color : COLOR )
{
VS_OUTPUT output = (VS_OUTPUT)0;
output.Pos = mul( Pos, World );
output.Pos = mul( output.Pos, View );
output.Pos = mul( output.Pos, Projection );
output.Color = Color;
return output;
}
I'm aware a float4 is basically a homogenous coordinate and needed for the transformations. As this is a position, I'd expect the fourth value of Pos (Pos.w, if you will) to be 1.
But how exactly does this conversion work? I've just defined POSITION to be 3 floats in C++ code, and now I'm suddently using a float4 in my Vertex shader.
In my naivety, I would have expected one of two things to happen:
Either: Pos is initialized as a float4 array with all zero-elements, and then the first 3 elements are filled with the vertix coordinates. But this would result in the fourth coordinate / w = 0, instead of 1.
Or: Since I've defined "Color" Input with InputSlot=12, i.e. to start with a byte-offset of 12, I could've imagined that Pos[0] = First four bytes (vertex position x), Pos[1] = Next 4 bytes (vertex.y), Pos[2] = Next 4 bytes (vertex.z), and Pos[3] = next 4 bytes - which would be the first element of COLOR.
Why does neither of these alternatives/errors happen? How, and why, does DirectX convert my float3 coordinates automatically to a float4 with w=1?
Thanks!
The default values for vertex attribute components that are missing are (0,0,0,1).
The MSDN confirms that's the case for D3D8 and D3D9 but I can't find an equivalent page that confirms that behaviour continued for 10 and 11. From experience though, I can say that this is still the case. Missing X, Y and Z components are replaced with 0, and missing W components are replaced with 1.

D3D11/C++ Inaccuracies in uv interpolation in pixel shader. How to avoid?

I'm trying to draw a quad with a texture onto the screen such that texels and pixels perfectly align. Sounds pretty easy. I draw 2 triangles (as TRIANGLE_LIST, so 6 vertices) using these shaders:
struct VSOutput
{
float4 position : SV_POSITION;
float2 uv : TEXCOORD0;
};
VSOutput VS_Draw(uint index : SV_VertexId)
{
uint vertexIndex = index % 6;
// compute face in [0,0]-[1,1] space
float2 vertex = 0;
switch (vertexIndex)
{
case 0: vertex = float2(0, 0); break;
case 1: vertex = float2(1, 0); break;
case 2: vertex = float2(0, 1); break;
case 3: vertex = float2(0, 1); break;
case 4: vertex = float2(1, 0); break;
case 5: vertex = float2(1, 1); break;
}
// compute uv
float2 uv = vertex;
// scale to size
vertex = vertex * (float2)outputSize;
vertex = vertex + topLeftPos;
// convert to screen space
VSOutput output;
output.position = float4(vertex / (float2)outputSize * float2(2.0f, -2.0f) + float2(-1.0f, 1.0f), 0, 1);
output.uv = uv;
return output;
}
float4 PS_Draw(VSOutput input) : SV_TARGET
{
uint2 pixelPos = (uint2)(input.uv * (float2)outputSize);
// output checker of 4x4
return (((pixelPos.x >> 2) & 1) ^ ((pixelPos.y >> 2) & 1) != 0) ? float4(0, 1, 1, 0) : float4(1, 1, 0, 0);
}
where outputSize and topLeftPos are constants and expressed in pixel units.
Now for outputSize = (102,12) and topLeftPos=(0,0) I get (what I would expect):
link to image (as i'm not allowed to post images)
But for outputSize = (102,12) and topLeftPos=(0,0.5) I get: Output for x=0, y=0.5
link to image (as i'm not allowed to post images)
As you can see there is a uv-discontinuity where the 2 triangles connect and interpolation of uv is inaccurate). This basically happens (in x and y) only at positions around the .5 (actually below .49 it correctly snaps to texel 0 and above .51 it snaps correctly to texel 1, but in between i get this artifact).
Now for the purpose I need this for it is essential to have pixel perfect mapping. Can anyone enlighten me why this happens ?
There are a two things you need consider to understand what is happening:
Pixel corners in window space have integer coordinates and pixel centers in windows space have half-integer coordinates.
When triangle is rasterized, D3D11 interpolates all attributes to pixel centers
So what is happening is that when topLeftPos=(0,0), the value of input.uv * (float2)outputSize is always half-integer, and it is consistently rounded down to closest integer. However, when topLeftPos=(0,0.5), the (input.uv * (float2)outputSize).y should always be exactly integer. However, due to unpredictable floating-point precision issues, it is sometimes little less than exact integer, and in this case it is rounded down too much. This is where you see your stretched squares.
So if you want perfect mapping, your source square should be aligned with the pixel boundaries and not pixel centers.

Opengl pixel perfect 2D drawing

I'm working on a 2d engine. It already works quite good, but I keep getting pixel-errors.
For example, my window is 960x540 pixels, I draw a line from (0, 0) to (959, 0). I would expect that every pixel on scan-line 0 will be set to a color, but no: the right-most pixel is not drawn. Same problem when I draw vertically to pixel 539. I really need to draw to (960, 0) or (0, 540) to have it drawn.
As I was born in the pixel-era, I am convinced that this is not the correct result. When my screen was 320x200 pixels big, I could draw from 0 to 319 and from 0 to 199, and my screen would be full. Now I end up with a screen with a right/bottom pixel not drawn.
This can be due to different things:
where I expect the opengl line primitive is drawn from a pixel to a pixel inclusive, that last pixel just is actually exclusive? Is that it?
my projection matrix is incorrect?
I am under a false assumption that when I have a backbuffer of 960x540, that is actually has one pixel more?
Something else?
Can someone please help me? I have been looking into this problem for a long time now, and every time when I thought it was ok, I saw after a while that it actually wasn't.
Here is some of my code, I tried to strip it down as much as possible. When I call my line-function, every coordinate is added with 0.375, 0.375 to make it correct on both ATI and nvidia adapters.
int width = resX();
int height = resY();
for (int i = 0; i < height; i += 2)
rm->line(0, i, width - 1, i, vec4f(1, 0, 0, 1));
for (int i = 1; i < height; i += 2)
rm->line(0, i, width - 1, i, vec4f(0, 1, 0, 1));
// when I do this, one pixel to the right remains undrawn
void rendermachine::line(int x1, int y1, int x2, int y2, const vec4f &color)
{
... some code to decide what std::vector the coordinates should be pushed into
// m_z is a z-coordinate, I use z-buffering to preserve correct drawing orders
// vec2f(0, 0) is a texture-coordinate, the line is drawn without texturing
target->push_back(vertex(vec3f((float)x1 + 0.375f, (float)y1 + 0.375f, m_z), color, vec2f(0, 0)));
target->push_back(vertex(vec3f((float)x2 + 0.375f, (float)y2 + 0.375f, m_z), color, vec2f(0, 0)));
}
void rendermachine::update(...)
{
... render target object is queried for width and height, in my test it is just the back buffer so the window client resolution is returned
mat4f mP;
mP.setOrthographic(0, (float)width, (float)height, 0, 0, 8000000);
... all vertices are copied to video memory
... drawing
if (there are lines to draw)
glDrawArrays(GL_LINES, (int)offset, (int)lines.size());
...
}
// And the (very simple) shader to draw these lines
// Vertex shader
#version 120
attribute vec3 aVertexPosition;
attribute vec4 aVertexColor;
uniform mat4 mP;
varying vec4 vColor;
void main(void) {
gl_Position = mP * vec4(aVertexPosition, 1.0);
vColor = aVertexColor;
}
// Fragment shader
#version 120
#ifdef GL_ES
precision highp float;
#endif
varying vec4 vColor;
void main(void) {
gl_FragColor = vColor.rgb;
}
In OpenGL, lines are rasterized using the "Diamond Exit" rule. This is almost the same as saying that the end coordinate is exclusive, but not quite...
This is what the OpenGL spec has to say:
http://www.opengl.org/documentation/specs/version1.1/glspec1.1/node47.html
Also have a look at the OpenGL FAQ, http://www.opengl.org/archives/resources/faq/technical/rasterization.htm, item "14.090 How do I obtain exact pixelization of lines?". It says "The OpenGL specification allows for a wide range of line rendering hardware, so exact pixelization may not be possible at all."
Many will argue that you should not use lines in OpenGL at all. Their behaviour is based on how ancient SGI hardware worked, not on what makes sense. (And lines with widths >1 are nearly impossible to use in a way that looks good!)
Note that OpenGL coordinate space has no notion of integers, everything is a float and the "centre" of an OpenGL pixel is really at the 0.5,0.5 instead of its top-left corner. Therefore, if you want a 1px wide line from 0,0 to 10,10 inclusive, you really had to draw a line from 0.5,0.5 to 10.5,10.5.
This will be especially apparent if you turn on anti-aliasing, if you have anti-aliasing and you try to draw from 50,0 to 50,100 you may see a blurry 2px wide line because the line fell in-between two pixels.

How do I get my textures to bind properly for multitexturing?

I'm trying to render colored text to the screen. I've got a texture containing a black (RGBA 0, 0, 0, 255) representation of the text to display, and I've got another texture containing the color pattern I want to render the text in. This should be a fairly simple multitexturing exercise, but I can't seem to get the second texture to work. Both textures are Rectangle textures, because the integer coordinate values are easier to work with.
Rendering code:
glActiveTextureARB(GL_TEXTURE0_ARB);
glEnable(GL_TEXTURE_RECTANGLE_ARB);
glBindTexture(GL_TEXTURE_RECTANGLE_ARB, TextHandle);
glActiveTextureARB(GL_TEXTURE1_ARB);
glEnable(GL_TEXTURE_RECTANGLE_ARB);
glBindTexture(GL_TEXTURE_RECTANGLE_ARB, ColorsHandle);
glBegin(GL_QUADS);
glMultiTexCoord2iARB(GL_TEXTURE0_ARB, 0, 0);
glMultiTexCoord2iARB(GL_TEXTURE1_ARB, colorRect.Left, colorRect.Top);
glVertex2f(x, y);
glMultiTexCoord2iARB(GL_TEXTURE0_ARB, 0, textRect.Height);
glMultiTexCoord2iARB(GL_TEXTURE1_ARB, colorRect.Left, colorRect.Top + colorRect.Height);
glVertex2f(x, y + textRect.Height);
glMultiTexCoord2iARB(GL_TEXTURE0_ARB, textRect.Width, textRect.Height);
glMultiTexCoord2iARB(GL_TEXTURE1_ARB, colorRect.Left + colorRect.Width, colorRect.Top + colorRect.Height);
glVertex2f(x + textRect.Width, y + textRect.Height);
glMultiTexCoord2iARB(GL_TEXTURE0_ARB, textRect.Width, 0);
glMultiTexCoord2iARB(GL_TEXTURE1_ARB, colorRect.Left + colorRect.Width, colorRect.Top);
glVertex2f(x + textRect.Width, y);
glEnd;
Vertex shader:
void main()
{
gl_Position = gl_ModelViewProjectionMatrix * gl_Vertex;
gl_TexCoord[0] = gl_MultiTexCoord0;
gl_TexCoord[1] = gl_MultiTexCoord1;
}
Fragment shader:
uniform sampler2DRect texAlpha;
uniform sampler2DRect texRGB;
void main()
{
float alpha = texture2DRect(texAlpha, gl_TexCoord[0].st).a;
vec3 rgb = texture2DRect(texRGB, gl_TexCoord[1].st).rgb;
gl_FragColor = vec4(rgb, alpha);
}
This seems really straightforward, but it ends up rendering solid black text instead of colored text. I get the exact same result if the last line of the fragment shader reads gl_FragColor = texture2DRect(texAlpha, gl_TexCoord[0].st);. Changing the last line to gl_FragColor = texture2DRect(texRGB, gl_TexCoord[1].st); causes it to render nothing at all.
Based on this, it appears that calling texture2DRect on texRGB always returns (0, 0, 0, 0). I've made sure that GL_MULTISAMPLE is enabled, and bound the texture on unit 1, but for whatever reason I don't seem to actually get access to it inside my fragment shader. What am I doing wrong?
The overalls look fine. It is possible that your texcoords for unit 1 are messed up, causing sampling outside the colored portion of your texture.
Is your color texture fully filled with color ?
What do you mean by "causes it to render nothing at all." ? This should not happen except if your alpha channel in color texture is set to 0.
Did you try with the following code, to override the alpha channel ?
gl_FragColor = vec4( texture2DRect(texRGB, gl_TexCoord[1].st).rgb, 1.0 );
Are you sure the the font outline texture contains a valid alpha values? You said that the texture is black and white, but you are using the alpha value! Instead of using the a component, try to use the r one.
Blending affects fragment shader output: it blends ths fragment color with the corresponding one.