1 of 83

2 of 83

3 of 83

4 of 83

The Next Leap:�How A.I. and automation might change the 3D industry

?

5 of 83

I almost never get asked: �"What's not going to change in the next 10 years?"

I frequently get asked: "What's going to change in the next 10 years?"

...and that’s actually the more important of the two.

--Jeff Bezos, Amazon CEO

6 of 83

Better,

Faster or

Cheaper

Any technology that makes things...

will eventually become standard.

7 of 83

The rising development cost of AAA games

$1B

$100M

$10M

$1M

$100K

$10K

Cost (in 2017 dollars)

1980 1985 1990 1995 2000 2005 2010 2015 2020

Gamasutra: Raph Koster’s “The cost of Games”

8 of 83

Assets: unreasonably expensive

Modelling: 12 hours

Texturing: 10 hours

First pass total: 22 hours

Total of 66 Hours @ $60/hr

$3900

Revisions: x2-3

Building from Turbosquid, user 3dThorium

9 of 83

10 of 83

Buildings

$3900

Buildings

$3900

Buildings

$6,000

Buildings

$6,000

Building

$3900

Building

$3900

Buildings

$3900

Waving flags

$2,250

Bannister

$2,450

Building Front

$3,800

Truck

$5,600

Car

$7,200

Car Damaged

$7,200

Car

$6,000

Traffic Cones

$3,600

Concrete

Divider:

$960

Road Texturing

$10,800

Road Texturing

$11,400

Hero Character

$49,000

Moving Sky

$4,200

Tree

$2,200

Tree

$2,200

Trash

$3,438

Big Trash

$14,400

NPC

$22,500

Car

$6,600

Billboard

$1,313

Bridge

$3,500

Tree

$2,200

Estimated Creation Cost: $200,000

Barrel $940

11 of 83

The problem? Static workflows

Modelling: 12 hours

Texturing: 10 hours

First pass total: 18 hours

Revisions: x2-3

Building from Turbosquid, user 3dThorium

12 of 83

Leap #1: Procedural Workflows

13 of 83

Procedural Modelling: Houdini

Procedural Lake Village by Anastasia Opera

14 of 83

15 of 83

Procedural Materials: Substance Designer

16 of 83

Procedural Texturing: Substance Painter

Bake

Smart Materials

Smart Masks

Now the #1 Texturing Software in the World

Art by 5518 studios

17 of 83

Procedural Level Design: Houdini

Finished ecosystem

Rules determine species growth

Tools for customization

Final game

Watch: Procedural World Generation of Ubisoft’s Far Cry 5

18 of 83

Prediction: Procedural will become standard

Modelling

Texturing

Materials

World Building

Houdini

Substance Painter

Substance Designer

Houdini

19 of 83

Rising game sizes

1TB by 2020

Gamasutra: Raph Koster’s “The cost of Games”

20 of 83

81km²

65GB in size

18,000,000,000,000,000,000 planets

7GB in size

GTA V

1,000+ artists

No Man’s Sky

13 developers

21 of 83

.

Static Content

Procedural

Content

(magnified x1000)

Space on hard drive

22 of 83

Leap #2: Machine Creep

23 of 83

Traditional Software

Machine Learning (example)

The key difference is its ability to learn

Input

Action

Output

Assess

Appropriate Action

Is it good?

No, try again

Input

Output

Compare with others

Yes

24 of 83

Machine learning thrives on

Huge Datasets &

Fast Hardware.

Which will improve over time.

25 of 83

Noisy image (1 sample p/pixel)

Denoised

Reference (4028 samples p/pixel)

Interactive Reconstruction of Monte Carlo Image Sequences using a Recurrent Denoising AutoencoderChakravarty R. Alla Chaitanya (NVIDIA), Anton Kaplanyan (NVIDIA), Christoph Schied (NVIDIA), Marco Salvi, Aaron Lefohn, Derek Nowrouzezahrai (McGill University), Timo Aila

26 of 83

Render

Denoised

Denoising with Kernel Prediction and Asymmetric Loss Functions

Vogels, Thijs and Rousselle, Fabrice and McWilliams, Brian and R\"othlin, Gerhard and Harvill, Alex and Adler, David and Meyer, Mark and Nov\'ak, Jan

27 of 83

Rendered at 50%

Time: 2m 40s

A.I. Upsampled to 100%

Render Time: 2m 40s

Rendered at 100%

Time: 9m 56s

Created using A.I. Gigapixel

28 of 83

A.I Upscaled to 100%

Reference (100%)

380% faster

29 of 83

30 of 83

Mode-Adaptive Neural Networks for Quadruped Motion Control

He Zhang, Sebastian Starke, Taku Komura, Jun Saito

31 of 83

Prediction: Machine Learning will become standard in software

Denoising

Uprezing

Motion capture & Animation

‘If you're not using machine learning in your software, you're doing it wrong,'

-Doug Roble (Digital Domain)

32 of 83

Leap #3: Machine Assisted Creativity

33 of 83

Experimentation takes up 50-70%

34 of 83

Toward Multimodal Image-to-Image Translation

J.Y. Zhu, R. Zhang, D. Pathak, �T. Darrell, A. A. Efros, O. Wang, E. Shechtman.

Input

Ground Truth

Generated Samples

35 of 83

36 of 83

Image-to-Image Demo

affinelayer.com/pixsrv/

37 of 83

38 of 83

Generated Images:

Image Inputs

Progressive Growing of GANs for Improved Quality, Stability, and Variation | NVidia Research�Tero Karras, Timo Aila, Samuli Laine, Jaakko Lehtinen

All generated images

39 of 83

Generated images from text

StackGAN: Text to Photo-realistic Image Synthesis with Stacked Generative Adversarial Networks�Han Zhang, Tao Xu, Hongsheng Li, Shaoting Zhang, Xiaogang Wang, Xiaolei Huang, Dimitris Metaxas�

40 of 83

Style Transfer

Style

Claude Monet:

Input:

A Style-Aware Content Loss for Real-time HD Style Transfer (ECCV), 2018.

Artsiom Sanakoyeu*, Dmytro Kotovenko*, Sabine Lang, Björn Ommer.

Output:

41 of 83

A Style-Aware Content Loss for Real-time HD Style Transfer (ECCV), 2018.

Artsiom Sanakoyeu*, Dmytro Kotovenko*, Sabine Lang, Björn Ommer.

42 of 83

A Style-Aware Content Loss for Real-time HD Style Transfer (ECCV), 2018.

Artsiom Sanakoyeu*, Dmytro Kotovenko*, Sabine Lang, Björn Ommer.

Fooled 39% of art historians

43 of 83

Prediction: Artists will use machine learning to explore new ideas

44 of 83

Changes in the next 5 years:

#1 Procedural workflows become standard

#3 Creative assistance

#2 Machine Learning replaces many tedious technical jobs

45 of 83

Are we being replaced?!

46 of 83

47 of 83

48 of 83

49 of 83

50 of 83

At Risk Jobs

  1. Labour Intensive
  2. Narrow Skilled
  3. Repetitive

Examples:

  • Mocap cleanup
  • Rotoscoping
  • Retopo
  • Mesh Cleanup
  • “Make 100 of these”

Safe Jobs

  • Critical Thinking
  • Wide skilled
  • Niche tasks

Examples:

  • Art Direction
  • Project management
  • Good Generalists
  • Programmers
  • Freelancers

Undesirable grunt tasks

Actual Art

51 of 83

Current

Future

Grunt Work

Creative Decisions

Programming

52 of 83

Box Office Movies

+5.18%

Digital TV

+13.25%

Mobile Gaming

+26.8%

PC & Console Gaming

+2.7%

Sources: Newzoo Global Games Market 2018 | CNBC “Digital TV and video industry to exceed $100 billion, study says” | Statista “Global box office revenue from 2016 to 2020”

53 of 83

3D Rendering and Viz expecting

25.5% growth per year till 2025

or

Double by 2022 &

Quadruple by 2025

54 of 83

Thank you

Special thanks to Cody Winchester for research assistance

55 of 83

Detects surgery problems 1000x faster

465,000 metric tons of carbon emissions saved per plane

Machine Learning Scan Analysis

Could save 1.3M lives per year

Self-Driving Cars

Generative Designed Door

56 of 83

Nvidia OptiX™

57 of 83

58 of 83

Rendering optimizations

#1 Denoising

59 of 83

Nvidia OptiX™

Currently used in Redshift, Arnold, Clarisse, Vray

60 of 83

Octane A.I. Denoiser

61 of 83

62 of 83

Nightmare Machine

From nightmare.mit.edu

63 of 83

A Style-Aware Content Loss for Real-time HD Style Transfer (ECCV), 2018.

Artsiom Sanakoyeu*, Dmytro Kotovenko*, Sabine Lang, Björn Ommer.

64 of 83

65 of 83

Admin &

Legal 40%

Programming 30%

Assets

25%

Concept

Art

5%

Typical AAA game development costs

66 of 83

81km²

65GB in size

67 of 83

18,000,000,000,000,000,000 planets

Only 7GB in size

68 of 83

69 of 83

Mordvintsev, et al., "Differentiable Image Parameterizations", Distill, 2018

70 of 83

Procedural Materials: Substance Designer

71 of 83

Procedural Materials: Substance Designer

By Eric Wiley on ArtStation

72 of 83

A Fully Progressive Approach to Single-Image Super-Resolution

Yifan Wang, Federico Perazzi, Brian McWilliams, Alexander Sorkine-Hornung,Olga Sorkine-Hornung, Christopher Schroers

73 of 83

Low Noise

Low Detail Loss

Low Artifacts

Frame-to-frame consistent

The Ultimate Goal

Cycles

74 of 83

The Division (2016) spent an estimated*

$55M

on assets

*Roughly 40% of total

75 of 83

45% lighter and twice as strong.

Saves 3,180kg of fuel per partition, per year.

76 of 83

Denoising

77 of 83

Noise2Noise: Learning Image Restoration without Clean Data

Jaakko Lehtinen, Jacob Munkberg, Jon Hasselgren, Samuli Laine, Tero Karras, Miika Aittala, Timo Aila

78 of 83

‘If you're not using machine learning in your software, you're doing it wrong,'

-Doug Roble (Digital Domain)

79 of 83

Typical Asset Costs:

20-45%

80 of 83

My Predictions

#1 Procedural Systems become standard

#3 Artists adopt machine assisted creative exploration

#2 Machine Learning replaces many tedious tasks

2018

(Today)

2019

2020

2021

2022

2023

2024

81 of 83

Used to be photographed

Now 3D

Currently photographed

Soon to be 3D

@shudu.gram

82 of 83

83 of 83

Procedural Materials: Substance Designer