# DX11 DX11 - SSAO

## Recommended Posts

Migi0027    4628

I'm so terribly sorry to re-re-re open this one, but I never fully succeeded at this one, SSAO.

Texture2D t_depthmap : register(t0);
Texture2D t_normalmap : register(t1);
Texture2D t_random : register(t2);
SamplerState ss;

cbuffer SSAOBuffer : register(c0)
{
float g_scale;
float g_bias;
float g_intensity;
float ssaoIterations;
float3 pppspace;

matrix view;
};

struct VS_Output
{
float4 Pos : SV_POSITION;
float2 Tex : TEXCOORD0;
};

{
VS_Output Output;
Output.Tex = float2((id << 1) & 2, id & 2);
Output.Pos = float4(Output.Tex * float2(2,-2) + float2(-1,1), 0, 1);
return Output;
}

// Helper for modifying the saturation of a color.
{
// The constants 0.3, 0.59, and 0.11 are chosen because the
// human eye is more sensitive to green light, and less to blue.
float grey = dot(color, float3(0.3, 0.59, 0.11));

return lerp(grey, color, saturation);
}

// Ambient Occlusion Stuff --------------------------------------------------

float3 getPosition(in float2 uv)
{
return t_depthmap.Sample(ss, uv).xyz;
}

float3 getNormal(in float2 uv)
{
return normalize(t_normalmap.Sample(ss, uv).xyz * 2.0f - 1.0f);
}

float2 getRandom(in float2 uv)
{
//return normalize(t_random.Sample(ss, uv ).xy * 2.0f - 1.0f); // ~100FPS
return normalize(t_random.Sample(ss, float2(600, 800) * uv / float2(60, 60)).xy * 2.0f - 1.0f);
}

float doAmbientOcclusion(in float2 tcoord,in float2 uv, in float3 p, in float3 cnorm)
{
float3 diff = getPosition(tcoord + uv) - p;
const float3 v = normalize(diff);
const float d = length(diff)*g_scale;
return max(0.0,dot(cnorm,v)-g_bias)*(1.0/(1.0+d))*g_intensity;
}

// End

{
const float2 vec[4] = {float2(1,0),float2(-1,0),
float2(0,1),float2(0,-1)};

float3 p = getPosition(input.Tex);
float3 n = getNormal(input.Tex);
float2 rand = getRandom(input.Tex);

float ao = 0.0f;

//**SSAO Calculation**//
int iterations = 4;
for (int j = 0; j < iterations; ++j)
{
float2 coord2 = float2(coord1.x*0.707 - coord1.y*0.707,
coord1.x*0.707 + coord1.y*0.707);

ao += doAmbientOcclusion(input.Tex, coord1*0.25, p, n);
ao += doAmbientOcclusion(input.Tex, coord2*0.5, p, n);
ao += doAmbientOcclusion(input.Tex, coord1*0.75, p, n);
ao += doAmbientOcclusion(input.Tex, coord2, p, n);
}
ao /= (float)iterations*4.0;

ao = saturate(ao);

return float4(ao, ao, ao, 1.0f);
}


How I write my depth map to the ssao:

VS
output.depth = output.position.z / output.position.w;
PS
float depth = input.depth;
output.Depth = float4(depth, depth, depth, 1);


How i write my normal map to the ssao:

VS -- I think this is wrong
output.NormalW = mul(mul(worldMatrix, viewMatrix), float4(normal.xyz,0) );
PS
output.Normals = float4(input.NormalW, 1);


Render Result:

Ehh.

Again, I'm so sorry for re-posting this, it's just, i really want this feature completed (well, working at least).

Thank You, like always...

##### Share on other sites

Why not use an existing implementation like the one from nvidia at https://developer.nvidia.com/nvidia-graphics-sdk-11-direct3d ?

Looking at your shader the first thing that strikes me is that your depth texture has more than one channel, where they are usually single channel.

##### Share on other sites
Migi0027    4628

So how would the depth shader be written correctly to the SSAO shader?

Thank you for your interest by the way, appreciate it!

##### Share on other sites
belfegor    2834

I think that that ssao shader needs view space position (whole three components x,y and z), as i can see you store normals that way, although:

output.NormalW = mul(mul(worldMatrix, viewMatrix), float4(normal.xyz,0) );


this mul order is right if shader packs matrices as column major, i don't know dx11  shader systems so i am not sure, but in dx9 i can force row major with this statement:

#pragma pack_matrix(row_major)



beacuse it matches my system in cpp code and i always mul vectors with matrices this way:

result = mul( vec, matrix );

Edited by belfegor

##### Share on other sites
Migi0027    4628

Ok, it is a bit better now in the sense that it doesn't just become randomly black: (SSAO Map)

But this isn't right yet, well I don't think it is, I mean look at the left of the cube, where the left side is rendered, it's black, why?

Ohh, and belfegor, about the matrices, you were right!

How I write my depth map to the ssao:

VS:

output.position = mul(position, worldMatrix);
output.position = mul(output.position, viewMatrix);
output.position = mul(output.position, projectionMatrix);

output.NormalW = mul(float4(-normal.xyz,0), mul(worldMatrix, viewMatrix));

// Store the position value in a second input value for depth value calculations.
output.depthPosition = output.position;

PS:

// Depth
float depth = 1.0f - (input.depthPosition.z / input.depthPosition.w);
output.Depth = float4(depth, depth, depth, 1.0f);

// Normals
output.Normals = float4(input.NormalW, 1); 

Now, is this the correct way to map the depth and the normal maps?

##### Share on other sites
belfegor    2834

I said output view space position, so it is like this:

VS
output.depthPosition.xyz = mul(float4(position.xyz,1), mul(worldMatrix, viewMatrix)).xyz;
PS
output.Depth = float4(input.depthPosition.xyz, 1.0f);


although you must have proper render target (or "view" whatever they call it in dx11) format that can hold at least 3 floating point components.

And why do you invert your normals now?

output.NormalW = mul(float4(-normal.xyz,0), mul(worldMatrix, viewMatrix));



Then when you fix those issues, play with ssao "sampling radius" to match your world scale (in case you don't get expected results).

Edited by belfegor

##### Share on other sites
Migi0027    4628

Sorry about the normals, just testing, and forgot it there...

The result, is this correct?

##### Share on other sites
belfegor    2834

I would say no. Try decreasing "g_sample_rad" and invert result like this:

ao = 1 - saturate(ao);


##### Share on other sites
Migi0027    4628

Now looks like this, look at the variables in the upper left corner:

##### Share on other sites
belfegor    2834

Let me try this one with my renderer and i get back to you.

EDIT: Here is it (on and off). All parameters are same as yours except radius which i set to 0.1f (10cm in my world) since it feels more natural for me.

Edited by belfegor

##### Share on other sites
Migi0027    4628

So are my values correct, taking away the radius?

##### Share on other sites
Migi0027    4628

I have absolutely no idea of where i snapped these values.

##### Share on other sites
belfegor    2834

I played a little more with this and done some changes that look better then before (at least for me), so you might want to try:

1. I have normalized view space normals in texture, and i don't get why do we need to scale-bias here? Also, this fixes some ugly artifacts.

float3 getNormal(in float2 uv)
{
//return normalize(tex2D(gbNormal_samp, uv).xyz * 2.0f - 1.0f);
return tex2D(gbNormal_samp, uv).xyz;
}


So you might need to change your gbuffer normal to be normalized also

output.Normals = float4(normalize(input.NormalW), 1);


2. I had getRandom before which looks nice for me:

float3 getRandom(in float2 uv)
{
return tex2D(rand_samp, uv * ScreenParams.xy / 4.0f).xyz * 2.0f - 1.0f;
}


3. Setting scale to 0 looks better for me, so i can even remove some calculations:

float doAmbientOcclusion(in float2 tcoord,in float2 uv, in float3 p, in float3 cnorm)
{
float g_scale = SSAO_params.x;
float g_bias = SSAO_params.y;
float g_intensity = SSAO_params.z;

float3 diff = getPosition(tcoord + uv) - p;
const float3 v = normalize(diff);
//const float d = length(diff)*g_scale;
return max(0.0,dot(cnorm,v)-g_bias)*g_intensity;//(1.0/(1.0+d))*g_intensity;
}


4. I get strange "haloing" artifacts on screen corners, so i set position and normal texture UV addressing mode to mirror and it seems that fixes that issue.

5. Changed radius to 0.15 and intensity to ~2.5

Edited by belfegor

##### Share on other sites
Migi0027    4628

The reason for the false occlusions at the corners is simply because there is nothing in the scene there,

But as you see in the bottom, when I go too close to the mesh, these false occlusions start to appear, and when really close, they're everywhere!

## Create an account

Register a new account

• ### Similar Content

• 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?

• 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!

• 15
• 14
• 16
• 10
• 18