GRAPHICS GEMS

Tiago Sousa, R&D Principal Graphics Engineer

AGENDA
Anti-aliasing

Practical Deferred MSAA Temporal Antialiasing: SMAA 1TX

Camera Post-Processing

Depth of Field Motion Blur

Sharing results from ongoing research

Results not used in a shipped game yet

Advances in Real-Time Rendering course, Siggraph 2013

ANTIALIASING\DEFERRED MSAA REVIEW

The problem: Multiple passes + r/w from Multisampled RTs

DX 10.1 introduced SV_SampleIndex / SV_Coverage system value semantics. Allows to solve via multipass for pixel/sample frequency passes [Thibieroz11] Forces pixel shader execution for each sub-sample and provides index of the sub-sample currently executed Index can be used to fetch sub-sample from a Multisampled RT. E.g. FooMS.Load( UnnormScreenCoord, nSampleIndex) Indicates to pixel shader which sub-samples covered during raster stage. Can modify also sub-sample coverage for custom coverage mask
Loop through MSAA tagged sub-samples
Advances in Real-Time Rendering course, Siggraph 2013 3

SV_SampleIndex

SV_Coverage

DX 11.0 Compute Tiled based deferred shading/lighting MSAA is simpler

DEFERRED MSAA\HEADS UP !
Simple theory, troublesome practice

At least with complex deferred renderers

Breaks regularly, as new techniques added without MSAA consideration Even if still works.. Very often youll need to pinpoint and fix non-msaa friendly techniques, as these introduce visual artifacts.
E.g. white/dark outlines, or no AA at all

Non-MSAA friendly code accumulates fast.

Do it upfront. Retrofitting a renderer to support Deferred MSAA is some work

And it is very finiky

Advances in Real-Time Rendering course, Siggraph 2013 4

DEFERRED MSAA\CUSTOM RESOLVE & PER-SAMPLE MASK

Post G-Buffer, perform a custom msaa resolve

Pre-resolves sample 0, for pixel frequency passes such as lighting/other MSAA dependent passes In same pass create sub-sample mask (compare samples similarity, mark if mismatching) Avoid default SV_COVERAGE, since it results in redundant processing on regions not requiring MSAA

SV_Coverage

Custom Per-Sample Mask

Advances in Real-Time Rendering course, Siggraph 2013 5

DEFERRED MSAA\STENCIL BATCHING [SOUSA13]

Batching per-sample stencil mask with regular stencil buffer usage

Reserve 1 bit from stencil buffer Update with sub-sample mask Tag entire pixel-quad instead of just single pixel -> improves stencil culling efficiency Make usage of stencil read/write bitmask to avoid per-sample bit override
StencilWriteMask = 0x7F

Restore whenever a stencil clear occurs

Not possible due to extreme stencil usage?

Use clip/discard Extra overhead also from additional texture read for per-sample mask
Advances in Real-Time Rendering course, Siggraph 2013 6

DEFERRED MSAA\PIXEL AND SAMPLE FREQUENCY PASSES

Pixel Frequency Passes

Set stencil read mask to reserved bits for per-pixel regions (~0x80) Bind pre-resolved (non-multisampled) targets SRVs Render pass as usual

Sample Frequency Passes

Set stencil read mask to reserved bit for per-sample regions (0x80) Bind multisampled targets SRVs Index current sub-sample via SV_SAMPLEINDEX Render pass as usual
Advances in Real-Time Rendering course, Siggraph 2013 7

DEFERRED MSAA\ALPHA TEST SSAA

Alpha testing requires ad hoc solution

Default SV_Coverage only applies to triangle edges E.g. check if current sub-sample uses AT or not and set bit
Alpha Test SSAA Disabled

Create your own sub-sample coverage mask

static const float2 vMSAAOffsets[2] = {float2(0.25, 0.25),float2(-0.25,-0.25)}; const float2 vDDX = ddx(vTexCoord.xy); const float2 vDDY = ddy(vTexCoord.xy); [unroll] for(int s = 0; s < nSampleCount; ++s) { float2 vTexOffset = vMSAAOffsets[s].x * vDDX + (vMSAAOffsets[s].y * vDDY); float fAlpha = tex2D(DiffuseSmp, vTexCoord + vTexOffset).w; uCoverageMask |= ((fAlpha-fAlphaRef) >= 0)? (uint(0x1)<<i) : 0; }

Alpha Test SSAA Enabled

Advances in Real-Time Rendering course, Siggraph 2013 8

DEFERRED MSAA\PERFORMANCE SHORTCUTS

Deferred cascades sun shadow maps

Render shadows as usual at pixel frequency Bilateral upscale during deferred shading composite pass

Advances in Real-Time Rendering course, Siggraph 2013

DEFERRED MSAA\PERFORMANCE SHORTCUTS (2)

Non-opaque techniques accessing depth (e.g. Soft-Particles)

Recommendation to tackle via per-sample frequency is fairly slow on real world scenarios Using Max Depth works ok for most cases and N-times faster

Advances in Real-Time Rendering course, Siggraph 2013

10

MSAA\PERFORMANCE SHORTCUTS (3)

Many games, also doing: Skipping Alpha Test Super Sampling

Use alpha to coverage instead, or even no alpha test AA (let morphological AA tackle that)

Render only opaque with MSAA

Then render transparents withouth MSAA Assuming HDR rendering: note that tone mapping is implicitly done post-resolve resulting is loss of detail on high contrast regions

Advances in Real-Time Rendering course, Siggraph 2013

11

DEFERRED MSAA\MSAA FRIENDLINESS

Look out for these:

No MSAA noticeably working, or noticeable bright/dark silhouettes.

Incorrect

Incorrect
Advances in Real-Time Rendering course, Siggraph 2013 12

DEFERRED MSAA\MSAA FRIENDLINESS

Look out for these:

No MSAA noticeably working, or noticeable bright/dark silhouettes.

Fixed

Fixed
Advances in Real-Time Rendering course, Siggraph 2013 13

DEFERRED MSAA\RECAP
Accessing and/or rendering to Multisampled RTs?

Then you need to care about accessing and outputting correct sub-sample Avoid vanilla deferred lighting
Prefer fully deferred, hybrids, or just skip deferred altogether.

In general always strive to minimize BW

If deferred, prefer thin g-buffers

Each additional target on g-buffer incurs in export rate overhead [Thibieroz11] NV/AMD (GCN): Export Cost = Cost(RT0)+Cost(RT1)..., AMD (older hw): Export Cost = (Num RTs) * (Slowest RT) Fat formats are half rate sampling cost for bilinear filtering modes on GCN [Thibieroz13] For lighting/some hdr post processes: 32 bit R11G11B10F fmt suffices for most cases
Advances in Real-Time Rendering course, Siggraph 2013 14

ANTIALIASING + 4K RESOLUTIONS\WILL WE NEED MSAA AT ALL?

Likely can start getting creative here

Highly Compressed 4k

Source 4k
Advances in Real-Time Rendering course, Siggraph 2013 15

ANTIALIASING \THE QUEST FOR BETTER (AND FAST) AA

2011: the boom year of alternative AA modes (and naming combos)

FXAA, MLAA, SMAA, SRAA, DEAA, GBAA, DLAA, ETC AA Filtering Approaches for Real-Time Anti-Aliasing [Jimenez et all 11]

Shading Anti-aliasing

Mip mapping normal maps [Toksvig04] Spectacular Specular: LEAN and CLEAN Specular Highlights [Baker11] Rock-Solid Shading: Image Stability withouth Sacrificing Detail [Hill12]

Advances in Real-Time Rendering course, Siggraph 2013

16

TEMPORAL SSAA\SMAA 2TX/4X REVIEW [JIMENEZ11][SOUSA11]

Morphological AA + MSAA + Temporal SSAA combo
Balanced cost/quality tradeoff, techniques complement each other. Temporal component uses 2 sub-pixel buffers. Each frame adds a sub-pixel jitter for 2x SSAA. Naive Blending Reproject previous frame and blend between current and previous frames, via velocity length weighting. Preserves image sharpness + reasonable temporal stability

Reprojection

|V| weighting

Advances in Real-Time Rendering course, Siggraph 2013

17

TEMPORAL AA\COMMON ROBUSTNESS FLAWS

Relying on opaque geometry information

Cant handle signal (color) changes nor transparency. For correct result, all opaque geometry must output velocity
Alpha blended surfaces (e.g. particles), lighting/shadow/reflections/uv animation/etc Any scatter and alike post processes, before the AA resolve E.g. ghosting on transparency, lighting, shadows and such Silhouettes might appear, from scatter and alike post processes (e.g. bloom)
Incorrect Blending Simplest solution: force resource sync NVIDIA exposes driver hint to force sync resource , via NVAPI. This is solution used by NVIDIAs TXAA
Note to hw vendors: would be great if all vendors exposed such (even better if Multi-GPU API functionality generalized)

Pathological cases

Can result in distracting errors

Multi-GPU

Advances in Real-Time Rendering course, Siggraph 2013

18

SMAA 1TX\A MORE ROBUST TEMPORAL AA

Concept: Only track signal changes, dont rely on geometry information

For higher temporal stability: accumulate multiple frames in an accumulation buffer, alike TXAA [Lottes12] Re-project accumulation buffer Weighting: Map acc. buffer colors into the range of curr. frame neighborhood color extents [Malan2012]; different weight for hi/low frequency regions (for sharpness preservation).

TL TR M BL BR
Current Frame (t0)

M
Accumulation Buffer (tN)
19

Advances in Real-Time Rendering course, Siggraph 2013

SMAA 1TX\A MORE ROBUST TEMPORAL AA (2)

Concept: Only track signal changes, dont rely on geometry information

For higher temporal stability: accumulate multiple frames in an accumulation buffer, alike TXAA [Lottes12] Re-project accumulation buffer Weighting: Map acc. buffer colors into the range of curr. frame neighborhood color extents [Malan2012]; different weight for hi/low frequency regions (for sharpness preservation).

Regular TAA
Advances in Real-Time Rendering course, Siggraph 2013

Smaa 1tx
20

SMAA 1TX\A MORE ROBUST TEMPORAL AA (3)

Sample code
float3 cM = tex2D(tex0, tc.xy); float3 cAcc = tex2D(tex0, reproj_tc.xy); float3 float3 float3 float3 cTL cTR cBL cBR = = = = tex2D(tex0, tex2D(tex0, tex2D(tex0, tex2D(tex0, tc0.xy); tc0.zw); tc1.xy); tc1.zw);

float3 cMax = max(cTL, max(cTR, max(cBL, cBR))); float3 cMin = min(cTL, min(cTR, min(cBL, cBR))); float3 wk = abs((cTL+cTR+cBL+cBR)*0.25-cM); return lerp(cM, clamp(cAcc, cMin, cMax), saturate(rcp(lerp(kl, kh, wk)));

Advances in Real-Time Rendering course, Siggraph 2013

21

DEPTH OF FIELD

Advances in Real-Time Rendering course, Siggraph 2013

22

DEPTH OF FIELD\PLAUSIBLE DOF: PARAMETERIZATION

Artist friendly parameters is one reason why games DOF tends to look wrong

Typical controls such as focus range + blur amount and others have not much physical meaning CoC depends mainly on f-stops, focal length and focal distance. These last 2 directly affect FOV. If you want more Bokeh, you need to max your focal length + widen aperture. This means also getting closer or further from subject for proper framing.
Not the typical way a game artist/programmer thinks about DOF.

Wider fov, less bokeh

Shallow fov, more bokeh

Advances in Real-Time Rendering course, Siggraph 2013 23

DEPTH OF FIELD\FOCAL LENGTH

50 mm

200 mm
Advances in Real-Time Rendering course, Siggraph 2013 24

DEPTH OF FIELD\F-STOPS

2 f-stops

8 f-stops

22 f-stops

Advances in Real-Time Rendering course, Siggraph 2013

25

DEPTH OF FIELD\F-STOPS (2)

2 f-stops

2.8 f-stops

4 f-stops

5.6 f-stops

Advances in Real-Time Rendering course, Siggraph 2013

26

DEPTH OF FIELD\FOCAL DISTANCE

0.5 m
Advances in Real-Time Rendering course, Siggraph 2013 27

DEPTH OF FIELD\FOCAL DISTANCE

0.75 m
Advances in Real-Time Rendering course, Siggraph 2013 28

DEPTH OF FIELD\FOCAL DISTANCE

1.0 m
Advances in Real-Time Rendering course, Siggraph 2013 29

DEPTH OF FIELD\PLAUSIBLE DOF: BOKEH

Out of focus region is commonly referred in photography as Bokeh (Japanese word for blur) Bokeh shape has direct relation to camera aperture size (aka f-stops) and diaphragm blades count

Bigger aperture = more circular bokeh, smaller aperture = more polygonal bokeh
Polygonal bokeh look depends on diaphragm blades count Blades count varies on lens characteristics

Bigger aperture = more light enters, smaller aperture = less light

On night shots, you might notice often more circular bokeh and more motion blur

Bokeh kernel is flat

Almost same amount of light enters camera iris from all directions
Edges might be in shadow, this is commonly known as Vignetting Poor lenses manufacturing may introduce a vast array of optical aberrations [Wiki01]

This is main reason why gaussian blur, diffusion dof, and derivative techniques look wrong/visually unpleasant
Advances in Real-Time Rendering course, Siggraph 2013 30

DEPTH OF FIELD\STATE OF THE ART OVERVIEW

Scatter based techniques [Cyril05][Sawada07][3DMark11][Mittring11][Sousa11]

Render 1 quad or tri per-pixel, scale based on CoC

Simple implementation and nice results. Downside: performance, particularly on shallow DOF
Variable/inconsistent fillrate hit, depending on near/far layers resolution and aperture size might reach >5 ms Quad generation phase has fixed cost attached.
Advances in Real-Time Rendering course, Siggraph 2013 31

DEPTH OF FIELD\STATE OF THE ART OVERVIEW (2)

Gather based: separable (inflexible kernel) vs. kernel flexibility

[Kawase09]

[Gotanda09]

[White11]

[Andreev 12]
Advances in Real-Time Rendering course, Siggraph 2013

[Macintosh12]
32

DEPTH OF FIELD\A PLAUSIBLE AND EFFICIENT DOF RECONSTRUCTION FILTER

Separable flexible filter: low bandwidth requirement + different bokeh shape possible

1st pass N^2 taps (e.g: 7x7). 2nd pass N^2 taps (e.g: 3x3) for flood filling shape R11G11B10F: downscaled HDR scene; R8G8: CoC Done at half resolution Far/Near fields processed in same pass Limit offset range to minimize undersampling Higher specs hw can have higher tap count

Diaphragm and optical aberrations sim Physically based CoC

Advances in Real-Time Rendering course, Siggraph 2013

33

DEPTH OF FIELD\LENS REVIEW

Pinhole Lens

A camera withouth lens Light has to pass through single small aperture before hitting image plane Tipical realtime rendering

Thin lens

Camera lenses have finite dimension Light refracts through lens until hitting image plane. F = Focal lenght P = Plane in focus I = Image distance
P
Advances in Real-Time Rendering course, Siggraph 2013

Image Plane

Circle of Confusion

I
34

DEPTH OF FIELD\LENS REVIEW (2)

The thin lens equation gives relation between:

F = Focal length (where light starts getting in focus) P = Plane in focus (camera focal distance) I = Image distance (where image is projected in focus)
f = f-stops (aka as the f-number or focal ratio) D = Object distance A = Aperture diameter

Circle of Confusion [Potmesil81]

Simplifies to:

Note: f and F are known variables from camera setup Folds down into a single mad in shader
Typical film formats (or sensor), 35mm/70mm Can alternatively derive focal length from FOV
Advances in Real-Time Rendering course, Siggraph 2013 35

Camera FOV:

DEPTH OF FIELD\SAMPLING
Concentric Mapping [Shirley97] used for uniform sample distribution

Maps unit square to unit circle Square mapped to (a,b) [-1,1]2 and divided into 4 regions by lines a=b, a=-b First region given by:

Diaphragm simulation by morphing samples to n-gons

Via a modified equation for the regular polygon.

Advances in Real-Time Rendering course, Siggraph 2013

36

DEPTH OF FIELD\

ND SAMPLING: 2 ITERATION

To floodfill final shape, composite via boolean union, similarly to [McIntosh12]

+...=

U...=
37

Advances in Real-Time Rendering course, Siggraph 2013

DEPTH OF FIELD\SEPARABLE FILTER PASSES

1st iteration: 49 taps (0.426ms)

2nd iteration: 9 taps (0.094 ms; 441 taps accumulated)

Advances in Real-Time Rendering course, Siggraph 2013 38

DEPTH OF FIELD\REFERENCE VS SEPARABLE FILTER

144 taps (1.31ms)

58 taps (0.52ms)
Advances in Real-Time Rendering course, Siggraph 2013 39

DEPTH OF FIELD\DIAPHRAGM SIMULATION IN ACTION

2f-stops

4f-stops
Advances in Real-Time Rendering course, Siggraph 2013 40

DEPTH OF FIELD\TILE MIN/MAX COC

Tile Min/Max CoC

Downscale CoC target k times (k = tile count) Take min fragment for far field, max fragment for near field R8G8 storage
Dynamic branching using Tile Min/Max CoC for both fields Balances cost between far/near Also used for scatter as gather approximation for near field

Used to process near/far fields in same pass

Tile min-CoC (far field)

Can fold cost with other post-processes

Initial downscale cost folded with HDR scene downscale for bloom, also pack near/far fields HDR input into R11G11B10F - all in 1 pass

Tile max-CoC (near field)

Advances in Real-Time Rendering course, Siggraph 2013

41

DEPTH OF FIELD\FAR + NEAR FIELD PROCESSING

Both fields use half resolution input

Careful: downscale is source of error due to bilinear filtering Use custom bilinear (bilateral) filter for downscaling
Scale kernel size and weight samples with far CoC [Scheumerman05] Pre-multiply layer with far CoC [Gotanda09]
Prevents bleeding artifacts from bilinear/separable filter

Far Field

No weighting

CoC weighting

CoC weighting + CoC pre-multiply

Advances in Real-Time Rendering course, Siggraph 2013

42

DEPTH OF FIELD\FAR + NEAR FIELD PROCESSING

Both fields use half resolution input

Careful: downscale is source of error due to bilinear filtering Use custom bilinear (bilateral) filter for downscaling
Scale kernel size and weight samples with far CoC [Scheumerman05] Pre-multiply layer with far CoC [Gotanda09]
Prevents bleeding artifacts from bilinear/separable filter

Far Field

Near Field
Scatter as gather aproximation Scale kernel size + weight fragments with Tile Max CoC against near CoC Pre-multiply with near CoC

Only want to blur near field fragments (cheap partial occlusion approximation)

Far Field

Near Field

Advances in Real-Time Rendering course, Siggraph 2013

43

DEPTH OF FIELD\FINAL COMPOSITE

Far field: upscale via bilateral filter

Take 4 taps from half res CoC, compare against full res CoC Weighted using bicubic filtering for quality [Sigg05] Far field CoC used for blending
Half resolution near field CoC used for blending Can bleed as much as possible Also using bicubic filtering Linear blending doesnt look good (signal frequency soup)
Can be seen in many games, including all Crysis series (puts hat of shame)

Near field: upscale carelessly

Linear blend

Carefull with blending

Simple to address: use non-linear blend factor instead.

Non-linear blend (better)

Advances in Real-Time Rendering course, Siggraph 2013

44

MOTION BLUR

Advances in Real-Time Rendering course, Siggraph 2013

45

MOTION BLUR\SHUTTER SPEED AND F-STOPS REVIEW

Amount of motion blur is relative to camera shutter speed and f-stops usage

The longer the exposure (slower shutter), the more light received (and the bigger amount of motion blur), and vice-versa The lower f-stops the faster the exposure can be (and have less motion blur), and vice versa

2 f-stops, shutter 1/20 sec

4 f-stops, shutter 1/6 sec

Advances in Real-Time Rendering course, Siggraph 2013 46

MOTION BLUR\STATE OF THE ART OVERVIEW

Scatter via geometry expansion [Green03][Sawada07]

Require additional geometry pass + gs shader usage *

Advances in Real-Time Rendering course, Siggraph 2013

47

MOTION BLUR\STATE OF THE ART OVERVIEW (2)

Scatter as gather [Sousa08][Gotanda09][Sousa11][Maguire12]

E.g.velocity dilation, velocity blur, tile max velocity; single vs. multiple pass composite; depth/v/obj ID masking; single pass DOF+MB

Advances in Real-Time Rendering course, Siggraph 2013

48

MOTION BLUR\RECONSTRUCTION FILTER FOR PLAUSIBLE MB [MCGUIRE12]

Tile Max Velocity and Tile Neighbor Max Velocity

Downscale Velocity buffer by k times (k is tile count) Take max length velocity at each step

Motion Blur Pass

Tile Neighbor Max for early out Tile Max Velocity as center velocity tap At each iteration step weight against full resolution ||V|| and Depth

Advances in Real-Time Rendering course, Siggraph 2013

49

MOTION BLUR\AN IMPROVED RECONSTRUCTION FILTER

Performant Quality

Simplify and vectorize inner loop weight computation (ends up in couple mads) Fat buffers sampling are half rate on GCN hw with bilinear (point filtering is fullrate, but doesnt look good due to aliasing) Inputs: R11G11B10F for scene , bake ||V|| and 8 bit depth into a R8G8 target Make it separable, 2 passes [Sousa08]

1 iteration: 6 taps (0.236 ms)

2nd iter: 6 taps (+0.236 ms; 36 taps acc.)

50

Advances in Real-Time Rendering course, Siggraph 2013

MOTION BLUR\AN IMPROVED RECONSTRUCTION FILTER (2)

Inner loop sample
const float2 tc = min_tc + blur_step * s; const float lensq_xy = abs(min_len_xy + len_xy_step * s); const float2 vy = tex2Dlod(tex1, float4(tc.xy, 0, 0)); // x =||v||, y=depth const float2 cmp_z = DepthCmp(float2(vx.y, vy.y), float2(vy.y, vx.y), 1); const float4 cmp_v = VelCmp(lensq_xy, float2(vy.x, lensq_vx)); const float w = (dot(cmp_z.xy, cmp_v.xy) + (cmp_v.z * cmp_v.w) * 2); acc.rgb += tex2Dlod(tex0, float4(tc.xy, 0, 0)) * w; wacc += w;

float2 DepthCmp(float2 z0, float2 z1, float2 fSoftZ) { return saturate( (1.0f + z0* fSoftZ) - z1* fSoftZ ); } float4 VelCmp(float lensq_xy, float2 vxy) { return saturate((1.0f - lensq_xy.xxxx *rcp(vxy.xyxy)) + float4(0.0f, 0.0f, 0.95f, 0.95f)); }

Advances in Real-Time Rendering course, Siggraph 2013

51

MOTION BLUR\AN IMPROVED RECONSTRUCTION FILTER (3)

Output object velocity in G-Buffer (only when required)

Avoids separate geometry passes. Rigid geometry: object distance < distance threshold Deformable geometry: if amount of movement > movement threshold Moving geometry rendered last R8G8 fmt
Velocity encoded in gamma 2.0 space Precision still insufficient, but not much noticeable in practice
Encode

Composite with camera velocity

Object velocity

Decode
Advances in Real-Time Rendering course, Siggraph 2013

Composite with camera velocity

52

MOTION BLUR\MB OR DOF FIRST?

In real world MB/DOF occur simultaneously

A dream implementation: big N^2 kernel + batched DOF/MB

Or sprite based with MB quad stretching Full resolution! 1 Billion taps! FP16! Multiple layers!

But... performance still matters (consoles):

DOF before MB introduces less error when MB happening on focus

This is due MB is a scatter as gather op relying on geometry data. Any other similar op after will introduce error. And vice-versa. Error from MB after DOF is less noticeable.

Order swap makes DOF harder to fold with other posts

Additional overhead

MB before DOF
Advances in Real-Time Rendering course, Siggraph 2013

MB after DOF
53

FINAL REMARKS
Practical MSAA details

Dos and Donts

For just 4 extra texture ops and couple alu Separable flexible filter, any bokeh kernel shape doable 1st pass: 0.426ms, 2nd pass: 0.094ms. Sum: 0.52ms for reconstruction filter *

SMAA 1TX: A More Robust Temporal AA

A Plausible and Performant DOF Reconstruction Filter

An Improved Reconstruction Filter for Plausible Motion Blur

Separable, 1st pass: 0.236 ms, 2nd pass: 0.236ms. Sum: 0.472ms for reconstruction filter *

* 1080p + AMD 7970

Advances in Real-Time Rendering course, Siggraph 2013 54

SPECIAL THANKS
Natalya Tartachuck Michael Kopietz, Nicolas Schulz, Christopher Evans, Carsten Wenzel, Christopher Raine, Nick Kasyan, Magnus Larbrant, Pierre Donzallaz

Advances in Real-Time Rendering course, Siggraph 2013

55

WE ARE HIRING !

Advances in Real-Time Rendering course, Siggraph 2013

56

QUESTIONS ?

[email protected] / Twitter: Crytek_Tiago

Advances in Real-Time Rendering course, Siggraph 2013

57

REFERENCES
Potmesil M., Chakravarty I. Synthetic Image Generation with a Lens and Aperture Camera Model, 1981 Shirley P., Chiu K., A Low Distortion Map Between Disk and Square, 1997 Green S. Stupid OpenGL Tricks, 2003 Toksvig M., Mip Mapping Normal Maps, 2004 Scheuermann T., Tatarchuk N. Improved Depth of Field Rendering, Shader X3, 2005 Sigg C., Hadwiger M., Fast Third-Order Texture Filtgering, 2005 Cyril, P et al. Photographic Depth of Field Rendering, 2005 Gritz L., Eon E. The Importance of Being Linear, 2007 Sawada Y., Talk at Game Developers Conference, http://www.beyond3d.com/content/news/499 , 2007 Sousa T. Crysis Next Gen Effects, 2008 Gotanda Y. Star Ocean 4: Flexible Shader Management and Post Processing, 2009 Yang L. et al, Amortized SuperSampling, 2009 Kawase M., Anti-Downsized Buffer Artifacts, 2011 Sousa T., CryENGINE 3 Rendering Techniques, 2011 Binks D., Dynamic Resolution Rendering, 2011 Sousa T., Schulz N. , Kasyan N., Secrets of CryENGINE 3 Graphics Technology, 2011 Sousa T., Anti-Aliasing Methods in CryENGINE 3, 2011 Jimenez J. et.al, Filtering Approaches for Real-Time Anti-Aliasing, 2011 Thibieroz N., Deferred Shading Optimizations, 2011
Advances in Real-Time Rendering course, Siggraph 2013 58

REFERENCES
Cloward B., Otstott A., Cinematic Character Lighting in Star Wars: The Old Republic, 2011 Mittring M., Dudash B., The Technology Behind the DirectX11 Unreal Engine Samaritan Demo, 2011 Futuremark, 3DMark11 White Paper, 2011 Baker D., Spectacular Specular: LEAN and CLEAN Specular Highlights, 2011 White J., Brisebois C., More Performance! Five Rendering Ideas from Batlefield 3 and Need for Speed: The Run, 2011 Lottes T., Nvidia TXAA, http://www.nvidia.in/object/txaa-anti-aliasing-technology-in.html#gameContent=1 , 2012 McGuire M. et al, A Reconstruction Filter for Plausible Motion Blur, 2012 McIntosh L. Et al, Efficiently Simulating Bokeh, 2012 Malan H., Real-Time Global Illumination and Reflections in Dust 514, 2012 Hill S., Baker D., Rock-Solid Shading: Image Stability withouth Sacrificing Detail, 2012 Sousa T., Wenzel C., Raine C., Rendering Technologies of Crysis 3, 2013 Thibieroz N., Gruen H., DirectX11 Performance Reloaded, 2013 Andreev D., Rendering Tricks in Dead Space 3, 2013 http://en.wikipedia.org/wiki/Lens_%28optics%29 http://en.wikipedia.org/wiki/Pinhole_camera http://en.wikipedia.org/wiki/F-number http://en.wikipedia.org/wiki/Angle_of_view

Advances in Real-Time Rendering course, Siggraph 2013

59