# DX11 Texture close view looks pixelated

## Recommended Posts

dxdude    110
Good day,
i have some problem with my textures in my directx 11 app.
In the distance the texture filtering seems to work like it should.
But they look very pixelated when iam very close to it.

See this picture:

And this one:

Iam using "D3D11_FILTER_MIN_MAG_MIP_LINEAR" to create the texture sampler.

//Code for sampler creation
D3D11_SAMPLER_DESC SamDesc;
ZeroMemory(&SamDesc, sizeof(D3D11_SAMPLER_DESC));
SamDesc.Filter		= D3D11_FILTER_MIN_MAG_MIP_LINEAR;
SamDesc.MipLODBias	= 0.0f;
SamDesc.MaxAnisotropy	= 1;
SamDesc.ComparisonFunc	= D3D11_COMPARISON_NEVER;
SamDesc.BorderColor[0]	= SamDesc.BorderColor[1] = SamDesc.BorderColor[2] = SamDesc.BorderColor[3] = 0;
SamDesc.MinLOD		= 0;
SamDesc.MaxLOD		= D3D11_FLOAT32_MAX;


I tested it by using the filter "D3D11_FILTER_ANISOTROPIC" with MaxAnisotropy of 4 or 8 and the texture still looks pixelated at very close view.

//Code for texture loading
imageInfo.Width		= D3DX11_DEFAULT;
imageInfo.Height	= D3DX11_DEFAULT;
imageInfo.Depth		= D3DX11_DEFAULT;
imageInfo.FirstMipLevel	= D3DX11_DEFAULT;
imageInfo.MipLevels	= D3DX11_DEFAULT;
imageInfo.Usage		= D3D11_USAGE_DEFAULT;
imageInfo.Format	= DXGI_FORMAT_R8G8B8A8_UNORM;
imageInfo.MipFilter	= D3DX11_FILTER_LINEAR;
imageInfo.Filter	= D3DX11_FILTER_LINEAR;



It doesnt matter what kind of Filter i set or using default values for the 'D3DX11_IMAGE_LOAD_INFO' structure it still pixelated at close view.

First of all i tested if swapchain, backbuffer, depth texture are the same size as the window creation.
They where all matching.

//Code window creation
RECT rc = { 0, 0, 1600, 960 };

hWnd = CreateWindow(L"myclass", L"myapp", WS_OVERLAPPEDWINDOW,
CW_USEDEFAULT, CW_USEDEFAULT, rc.right - rc.left, rc.bottom - rc.top, NULL, NULL, hInstance, NULL);

if (!hWnd)
return FALSE;

g_WindowHWND = hWnd;
ShowWindow(g_WindowHWND, nCmdShow);

//Code buffer creation after window is created
m_hWndMainRenderTarget = hwnd;

// Create a Direct2D render target
RECT rcRenderTarget;
GetClientRect( hwnd, &rcRenderTarget);
m_uiRenderTargetWidth	= rcRenderTarget.right-rcRenderTarget.left;
m_uiRenderTargetHeight	= rcRenderTarget.bottom-rcRenderTarget.top;

// Create swapchain settings
DXGI_SWAP_CHAIN_DESC sSwapChainDesc;
ZeroMemory( &sSwapChainDesc, sizeof( sSwapChainDesc ) );

sSwapChainDesc.BufferCount		= 1;
sSwapChainDesc.BufferDesc.Width		= m_uiRenderTargetWidth;
sSwapChainDesc.BufferDesc.Height	= m_uiRenderTargetHeight;
sSwapChainDesc.BufferDesc.Format	= DXGI_FORMAT_R8G8B8A8_UNORM;
sSwapChainDesc.BufferUsage		= DXGI_USAGE_RENDER_TARGET_OUTPUT;
sSwapChainDesc.OutputWindow		= m_hWndMainRenderTarget;
sSwapChainDesc.SampleDesc.Count		= 1;
sSwapChainDesc.SampleDesc.Quality	= 0;
sSwapChainDesc.Windowed			= m_bWindowed;

// Retrive device, adapter and factory that was created with the device
IDXGIDevice * pDXGIDevice;
hr = m_pDevice->QueryInterface(__uuidof(IDXGIDevice), (void **)&pDXGIDevice);

IDXGIFactory * pIDXGIFactory;

// Create swap chain seperate
if(FAILED( hr = pIDXGIFactory->CreateSwapChain( pDXGIDevice, &sSwapChainDesc, &m_pSwapChain ) ))
{
return hr;
}
// Get a pointer to the back buffer
if(FAILED( hr = m_pSwapChain->GetBuffer( 0, __uuidof( ID3D11Texture2D ), ( LPVOID* )&m_pBackBuffer )))
{
return hr;
}

// Create a render-target view
if(FAILED( hr = m_pDevice->CreateRenderTargetView( m_pBackBuffer, NULL, &m_pRenderTargetView )))
{
return hr;
}

D3D11_TEXTURE2D_DESC	sDepthStencilTextureDesc;

// Create depth stencil texture etc..
sDepthStencilTextureDesc.Width			= m_uiRenderTargetWidth;
sDepthStencilTextureDesc.Height			= m_uiRenderTargetHeight;
sDepthStencilTextureDesc.MipLevels		= 1;
sDepthStencilTextureDesc.ArraySize		= 1;
sDepthStencilTextureDesc.Format			= DXGI_FORMAT_D24_UNORM_S8_UINT;
sDepthStencilTextureDesc.SampleDesc.Count	= 1;
sDepthStencilTextureDesc.SampleDesc.Quality	= 0;
sDepthStencilTextureDesc.Usage			= D3D11_USAGE_DEFAULT;
sDepthStencilTextureDesc.BindFlags		= D3D11_BIND_DEPTH_STENCIL;
sDepthStencilTextureDesc.CPUAccessFlags		= 0;
sDepthStencilTextureDesc.MiscFlags		= 0;

if(FAILED( hr = m_pDevice->CreateTexture2D( &sDepthStencilTextureDesc, NULL, &m_pDepthStencilTexture )))
{
return hr;
}

D3D11_DEPTH_STENCIL_VIEW_DESC	sDepthStencilViewDesc;

// Depth stencil view desc...
ZeroMemory( &sDepthStencilViewDesc, sizeof( sDepthStencilViewDesc ) );
sDepthStencilViewDesc.Format			= DXGI_FORMAT_D24_UNORM_S8_UINT;
sDepthStencilViewDesc.ViewDimension		= D3D11_DSV_DIMENSION_TEXTURE2D;
sDepthStencilViewDesc.Texture2D.MipSlice	= 0;
sDepthStencilViewDesc.Flags			= 0;

if(FAILED( hr = m_pDevice->CreateDepthStencilView( m_pDepthStencilTexture, &sDepthStencilViewDesc,&m_pDepthStencilView )))
{
return hr;
}

// Bind the view
m_pDeviceContext->OMSetRenderTargets( 1, &m_pRenderTargetView, m_pDepthStencilView );

D3D11_VIEWPORT sViewPort;

// Setup the viewport
sViewPort.Width		= (FLOAT)m_uiRenderTargetWidth;
sViewPort.Height	= (FLOAT)m_uiRenderTargetHeight;
sViewPort.MinDepth	= 0.0f;
sViewPort.MaxDepth	= 1.0f;
sViewPort.TopLeftX	= 0;
sViewPort.TopLeftY	= 0;



During some debug session i tested if all the dimension of the created buffers match up with the window client rect.
Every created buffer has the same size as the client area of the created window.
The window istself is very basic. No menu, bars or anything. Just a simply titlebar.
Text displayed on the screen done with my font engine are pixel perfect so i think the buffers size match up with the window client area size.
At this point i have to say that other application like the samples from the Microsoft sdk works like accepted.

Since iam running out of ideas i can post here some more code that could be relevant.

//Code projection matrx creation. zNear = 1.0f and zFar = 12000.0f
int width	= g_CDevice.GetRenderTargetWidth();
int height	= g_CDevice.GetRenderTargetHeight();
float fov	= 0.785398163f;
float aspectRatio = width / (float)height;
D3DXMatrixPerspectiveFovLH(&g_mProjection, fov, aspectRatio, g_fZNear, g_fZFar);



The pixelated effect normaly starts when iam comming close to the zNear value.
Something like 2 units above the zNear settings.

Here is the pixelshader iam using.
It is a combination of some detail textures with slope based texturing and some blending of mixed uv values for the detail textures.
They are also combined via a noise blending map for mixing the slope based texture values to get ride of the repeat effect of detail textures.
All textures used here are 512x512.
The texture are loaded all the same way like described at the beginning of this post and using the same texture sampler.

//Code pixelshader
Texture2D txColorMap_1 : register( t0 ); // grass
Texture2D txColorMap_2 : register( t1 ); // dirt
Texture2D txColorMap_3 : register( t2 ); // rock
Texture2D txColorMap_4 : register( t3 ); // rgba random blend map
Texture2D txColorMap_5 : register( t4 ); // noisy normal map

SamplerState samLinear2D_1 : register( s0 );

struct PixelInputType
{
float4 position : SV_POSITION;
float3 normal   : TEXCOORD0;
float2 tex_1    : TEXCOORD1; // uv for detail
float2 tex_2    : TEXCOORD2; // uv for noise map
};

float4 PSMain(PixelInputType input) : SV_Target
{
const float uvDetail = 32.0f; // detail uv in the range of 0.125f - 0.25f
const float4 vLightDir = float4(-0.5f, 1.0f, 1.0f, 1.0f); // Directional light for testing
const float4 vLightColor = float4(0.8f, 0.8f, 0.8f, 1.0f);
const float4 vAmbientColor = float4(0.2f, 0.2f, 0.2f, 1.0f);

float4 finalColor;
float blendAmount;

float4 t_1 =  txColorMap_4.Sample( samLinear2D_1, input.tex_2 * 0.125f);// random map blend values
float4 col;

float4 c1 = txColorMap_1.Sample( samLinear2D_1, input.tex_1 * uvDetail);
float4 c2 = txColorMap_2.Sample( samLinear2D_1, input.tex_1 * uvDetail);
float4 c3 = txColorMap_3.Sample( samLinear2D_1, input.tex_1 * uvDetail);

// Read detail texture with lower uv values for mixing
float4 c4 = txColorMap_1.Sample( samLinear2D_1, input.tex_1 * uvDetail * 0.25f);
float4 c5 = txColorMap_2.Sample( samLinear2D_1, input.tex_1 * uvDetail * 0.125f);
float4 c6 = txColorMap_3.Sample( samLinear2D_1, input.tex_1 * uvDetail * 0.125f);

// lerp the values with the blend values of the
c1 = lerp(c1*c4, c2, t_1.r);
c2 = lerp(c2*c5, c3, t_1.g);
c3 = lerp(c3,c6, t_1.b);

//  Slope calculation based on rastertek tutorial
float slope = 1.0f - input.normal.y;

if(slope < 0.2)
{
blendAmount = slope / 0.2f;
col = lerp(c1, c2, blendAmount);
}

if((slope < 0.7) && (slope >= 0.2f))
{
blendAmount = (slope - 0.2f) * (1.0f / (0.7f - 0.2f));
col = lerp(c2, c3, blendAmount);
}

if(slope >= 0.7)
{
col = c3;
}

// add normal map noise values for some cheap bump effect
float3 n1 = txColorMap_5.Sample(samLinear2D_1, input.tex_1 * uvDetail);
float3 n2 = txColorMap_5.Sample(samLinear2D_1, input.tex_1 * uvDetail * 0.25f);
n1 = lerp(n1, n2, t_1.r);
float  d = dot((float3)vLightDir,n1);
col *= d;

// calculate directional lighting
finalColor = saturate( dot( (float3)vLightDir,input.normal) * vLightColor) * col;
finalColor += col * vAmbientColor;

return finalColor;
}


Thats all the code that could be relevant to guess what kind of bug i have.
Thanks in advance to anyone here.
Edited by dxdude

##### Share on other sites
AmzBee    485

The one thing I can't see mentioned here is the UV coordinates you are using, which may mean you are trying to debug code that doesn't actually need to be debugged. Have a look (if you haven't already) to see if the UV coords you are using use the full resolution of the texture.

Aimee

##### Share on other sites
dxdude    110

Thank you for the answer AmzBee.

I cant test any code here at work atm so will do it when iam back home later.

The UV are generated inside my vertex shader.

Can very low uv coodinates make problems?

So when mapping from 0 to 1 and the uv for it are like 0,0078125 - 0,01..... ?

But ill check my vertex shader later.

Thank you very much.

##### Share on other sites
dxdude    110

The one thing I can't see mentioned here is the UV coordinates you are using, which may mean you are trying to debug code that doesn't actually need to be debugged. Have a look (if you haven't already) to see if the UV coords you are using use the full resolution of the texture.

Aimee

Woot i solved it now.

There was a bug in the uv generation in the vertex shader process.

Thank you very much.

And sorry for the big post :)

##### Share on other sites
Norman Barrows    7179

Can very low uv coodinates make problems?
So when mapping from 0 to 1 and the uv for it are like 0,0078125 - 0,01..... ?

I don't do shaders, but i follow pretty much everything you're doing.

looks like you might have a classic case of "5 foot texture, 10 foot rock" as i call it.   IE your texture is not high rez enough for the size of the mesh its mapped onto.

when this occurs, you can get very small u,v coords for a given tri, such as you mention.

that and your first image (a classic case of low rez texture) are what make me suspect "5 foot texture, 10 foot rock", or "1 meter texture, 2 meter rock" if you prefer.

if thats the case, then the image in your texture is a picture of a piece of land thats not as big as the area it gets mapped onto in your world. so it gets stretched. instead of vertex coords 0 to 1 mapping to UVs  0 to 1, you get UVs 0 to 0.1 or .01, etc.

so, you can do one or a combo of things:
1. increase texture rez.   i run 256x256 for speed, but have tested up to 4096x4096 on the exact same case you're working (ground textures). i saw that first image of yours, and i was like "yeah, been there, done that, i remember that, walking up to rocks and ground that sticks up, and figuring out how to get high enough rez with the lowest rez textures possible, while tweaking texture wrap, quad size, and seamless textures.   increasing texture size costs memory.   bigger textures run slower (at least in fixed function pipeline).

2. decrease quad/triangle size.  1/2 as big with same mapping means twice the resolution from the same texture.  seamless textures may be required.

3. repeat the texture more than once across a quad - usually requires seamless texture. this gets you the high rez of a high rez texture without the memory hit, and doesn't increase your triangle count. the downside is possible moire' patterns from repeating the same texture 2 or more times across a surface.

here's the load code for the texture:

// load texture - creates mipmaps. no image filtering, box filtering for mipmaps.
{
HRESULT h;
char s2[100];
h=D3DXCreateTextureFromFileExA(Zd3d_device_ptr,s,D3DX_DEFAULT,  // widtth
D3DX_DEFAULT,   // height
D3DX_DEFAULT,      // miplvls ( default = complete chain )
0,                 // usage (0=not render & not dynamic)
D3DFMT_A8R8G8B8,
mempool,                         // memory pool of your choice - managed, most likely.
D3DX_FILTER_NONE,      //   image filter   ( default = tri + dither )
D3DX_DEFAULT,      // mip filter ( default = box )
0,     // color key (0=none)
NULL,NULL,&(Ztex[numtextures].tex));
if (h != D3D_OK) { strcpy_s(s2,100,"Error loading "); strcat_s(s2,100,s); Zmsg2(s2); exit(1); }
strcpy_s(Ztex[numtextures].name,s);
numtextures++;
}

and here are the states of the pipeline when drawing:

// no blending, turn on gouraud shading
Zd3d_device_ptr->SetTextureStageState( 0, D3DTSS_COLOROP, D3DTOP_MODULATE );
Zd3d_device_ptr->SetTextureStageState( 0, D3DTSS_COLORARG1, D3DTA_TEXTURE );
Zd3d_device_ptr->SetTextureStageState( 0, D3DTSS_COLORARG2, D3DTA_DIFFUSE );
Zmipmaps(1);         //  Zd3d_device_ptr->SetSamplerState(0,D3DSAMP_MIPFILTER,D3DTEXF_LINEAR);
Zminmagfilter(2);   // 0=point, 1=linear, 2=aniso min and magnification
Zambient(255,255,255);  //  i control ambient with materials
Znormalize(1);                // normalize normals on.   i do scaling of meshes on the fly.
Zspecular(1);                  // turn on specular
Zalphablend(0);             // alpha blend off

and heres the results with 256x256 texture on a 10x10 quad.    when i get up close and personal like in your first image, i get about 4x the rez you are getting (IE just slightly pixelated).

well, i was going to post an image, but the image button doesn't work. i have a couple screenshots in my gallery if you want to check them out. not sure how well they show the ground though.

It is a combination of some detail textures with slope based texturing and some blending of mixed uv values for the detail textures.
They are also combined via a noise blending map for mixing the slope based texture values to get ride of the repeat effect of detail textures.

i believe this is the only place where we're doing things differently. i do a simple aniso tmapping of a seamless texture onto a height mapped quad.

as i said, i'm using 256x256 tex on a 10x10 quad. based on you first image, if you're using a 512x512 texture, it looks like you're mapping it onto a quad size of about 80x80, whereas 512x512 mapped onto 20x20 would be the equivalent of what i'm doing (4 times the rez).

from my testing, i found i couldn't go bigger than a 10x10 with a 256x256 texture, without unacceptable levels of pixelation at closest ranges. this would be the equivalent of a 20x20 quad using 512x512 textures.

don't forget that the size of the image in the texture can be a factor too.

if your mapping a 512x512 tex onto a 20x20 quad, at 1 meter per d3d or Ogl unit, and your image on your texture is only a section of land 5 meters across, you'll also get similar results to your first image.

##### Share on other sites
Norman Barrows    7179

There was a bug in the uv generation in the vertex shader process.

uv coords too small , eh?

Edited by Norman Barrows

##### Share on other sites
AmzBee    485

The one thing I can't see mentioned here is the UV coordinates you are using, which may mean you are trying to debug code that doesn't actually need to be debugged. Have a look (if you haven't already) to see if the UV coords you are using use the full resolution of the texture.

Aimee

Woot i solved it now.

There was a bug in the uv generation in the vertex shader process.

Thank you very much.

And sorry for the big post

no worries, sometimes the biggest problem can be solved with the smallest fix lol :P

##### Share on other sites
dxdude    110

There was a bug in the uv generation in the vertex shader process.

uv coords too small , eh?

Nice info Norman.

There where not to small but i did a small mistake.

Since i update a small constant buffer for each patch that get rendered i also supplied sector data from the patch.

The sector data was used to generate continuous uv coordinates.

But i got a small bug in the update process for the constant buffer.

Nice info you supplied. Thank you very much.

Edited by dxdude

##### Share on other sites
Norman Barrows    7179

But i got a small bug in the update process for the constant buffer.

its always the little stuff, isn't it?

i was working on blending height map edges the other day.   it helps if you calculate the average height at an edge as (h1+h2)/2 as opposed to h1 + h2/2. i forgot the parentheses. that one took a few hours to find, tracing up and down the call stack (so to speak, purely by code inspection).

## Create an account

Register a new account

• ### Similar Content

• By Endurion
I have a gaming framework with an renderer interface. Those support DX8, DX9 and latest, DX11. Both DX8 and DX9 use fixed function pipeline, while DX11 obviously uses shaders. I've got most of the parts working fine, as in I can switch renderers and notice almost no difference. The most advanced features are 2 directional lights with a single texture
My last problem is lighting; albeit there's documentation on the D3D lighting model I still can't get the behaviour right. My mistake shows most prominently in the dark side opposite the lights. I'm pretty sure the ambient calculation is off, but that one's supposed to be the most simple one and should be hard to get wrong.
Interestingly I've been searching high and low, and have yet to find a resource that shows how to build a HLSL shader where diffuse, ambient and specular are used together with material properties. I've got various shaders for all the variations I'm supporting. I stepped through the shader with the graphics debugger, but the calculation seems to do what I want. I'm just not sure the formula is correct.
This one should suffice though, it's doing two directional lights, texture modulated with vertex color and a normal. Maybe someone can spot one (or more mistakes). And yes, this is in the vertex shader and I'm aware lighting will be as "bad" as in fixed function; that's my goal currently.
• By Mercesa
Hey folks. So I'm having this problem in which if my camera is close to a surface, the SSAO pass suddenly spikes up to around taking 16 milliseconds.
When still looking towards the same surface, but less close. The framerate resolves itself and becomes regular again.
This happens with ANY surface of my model, I am a bit clueless in regards to what could cause this. Any ideas?
In attached image: y axis is time in ms, x axis is current frame. The dips in SSAO milliseconds are when I moved away from the surface, the peaks happen when I am very close to the surface.

Edit: So I've done some more in-depth profiling with Nvidia nsight. So these are the facts from my results
Count of command buffers goes from 4 (far away from surface) to ~20(close to surface).
The command buffer duration in % goes from around ~30% to ~99%
Sometimes the CPU duration takes up to 0.03 to 0.016 milliseconds per frame while comparatively usually it takes around 0.002 milliseconds.
I am using a vertex shader which generates my full-screen quad and afterwards I do my SSAO calculations in my pixel shader, could this be a GPU driver bug? I'm a bit lost myself. It seems there could be a CPU/GPU resource stall. But why would the amount of command buffers be variable depending on distance from a surface?

Edit n2: Any resolution above 720p starts to have this issue, and I am fairly certain my SSAO is not that performance heavy it would crap itself at a bit higher resolutions.

• In DirectX 11 we have a 24 bit integer depth + 8bit stencil format for depth-stencil resources ( DXGI_FORMAT_D24_UNORM_S8_UINT ). However, in an AMD GPU documentation for consoles I have seen they mentioned, that internally this format is implemented as a 64 bit resource with 32 bits for depth (but just truncated for 24 bits) and 32 bits for stencil (truncated to 8 bits). AMD recommends using a 32 bit floating point depth buffer instead with 8 bit stencil which is this format: DXGI_FORMAT_D32_FLOAT_S8X24_UINT.
Does anyone know why this is? What is the usual way of doing this, just follow the recommendation and use a 64 bit depthstencil? Are there performance considerations or is it just recommended to not waste memory? What about Nvidia and Intel, is using a 24 bit depthbuffer relevant on their hardware?
Cheers!

• By gsc
Hi! I am trying to implement simple SSAO postprocess. The main source of my knowledge on this topic is that awesome tutorial.
But unfortunately something doesn't work... And after a few long hours I need some help. Here is my hlsl shader:
float3 randVec = _noise * 2.0f - 1.0f; // noise: vec: {[0;1], [0;1], 0} float3 tangent = normalize(randVec - normalVS * dot(randVec, normalVS)); float3 bitangent = cross(tangent, normalVS); float3x3 TBN = float3x3(tangent, bitangent, normalVS); float occlusion = 0.0; for (int i = 0; i < kernelSize; ++i) { float3 samplePos = samples[i].xyz; // samples: {[-1;1], [-1;1], [0;1]} samplePos = mul(samplePos, TBN); samplePos = positionVS.xyz + samplePos * ssaoRadius; float4 offset = float4(samplePos, 1.0f); offset = mul(offset, projectionMatrix); offset.xy /= offset.w; offset.y = -offset.y; offset.xy = offset.xy * 0.5f + 0.5f; float sampleDepth = tex_4.Sample(textureSampler, offset.xy).a; sampleDepth = vsPosFromDepth(sampleDepth, offset.xy).z; const float threshold = 0.025f; float rangeCheck = abs(positionVS.z - sampleDepth) < ssaoRadius ? 1.0 : 0.0; occlusion += (sampleDepth <= samplePos.z + threshold ? 1.0 : 0.0) * rangeCheck; } occlusion = saturate(1 - (occlusion / kernelSize)); And current result: http://imgur.com/UX2X1fc
I will really appreciate for any advice!
• By isu diss
I'm trying to code Rayleigh part of Nishita's model (Display Method of the Sky Color Taking into Account Multiple Scattering). I get black screen no colors. Can anyone find the issue for me?