Progressive Transient Photon Beams Julio Marco 1 Ibn Guilln 1 - - PowerPoint PPT Presentation

▶
progressive transient photon beams
SMART_READER_LITE
LIVE PREVIEW

Progressive Transient Photon Beams Julio Marco 1 Ibn Guilln 1 - - PowerPoint PPT Presentation

Jun 06, 2023 470 likes •1.1k views

Progressive Transient Photon Beams Julio Marco 1 Ibn Guilln 1 Wojciech Jarosz 2 Diego Gutierrez 1 Adrian Jarabo 1 1 Universidad de Zaragoza, I3A 2 Dartmouth College 1 300,000 km/s LIGHT TRANSPORT 2 TRANSIENT LIGHT TRANSPORT 3

slide-1
SLIDE 1

Progressive Transient Photon Beams

1

Julio Marco1 Wojciech Jarosz2 Diego Gutierrez1 Adrian Jarabo1

1 Universidad de Zaragoza, I3A 2 Dartmouth College

Ibón Guillén1

slide-2
SLIDE 2

300,000 km/s LIGHT TRANSPORT

2

slide-3
SLIDE 3

TRANSIENT LIGHT TRANSPORT

3

slide-4
SLIDE 4

4

Femto-photography [Velten et al. 2013]

slide-5
SLIDE 5

5

Femto-photography [Velten et al. 2013]

slide-6
SLIDE 6

Transient Light Transport- What for?

6

  • Light in motion [Velten13, Heide13, Peters15…]
  • Visible geometry [Wu14, OToole14, Marco17…]
  • Transparent Objects [Kadambi13]
  • Hidden geometry [Velten12, Buttafava15, OToole18, Liu19,…]
  • Reflectance estimation [Naik11, Naik13]
  • GI Components Separation [Wu14, OToole14]
  • Vision through media [Heide14, Wu18…]
  • …
slide-7
SLIDE 7

SIMULATION

7

Forward model for inverse problems Benchmarking algorithms Prototyping Machine learning

slide-8
SLIDE 8

8

Forward model for inverse problems Benchmarking algorithms Prototyping Machine learning

Transient rendering

slide-9
SLIDE 9

9

Forward model for inverse problems Benchmarking algorithms Prototyping Machine learning

OUR GOAL Robust time-resolved participating media

slide-10
SLIDE 10

10

Steady state

Transient Rendering vs. Steady-state

slide-11
SLIDE 11

11

Radiance Steady state

Transient Rendering vs. Steady-state

slide-12
SLIDE 12

Finite speed of light è Temporal dimension

12

time Radiance Steady state

Transient Rendering

slide-13
SLIDE 13

Finite speed of light è Temporal dimension

13

time Radiance Steady state

Transient Rendering

slide-14
SLIDE 14
  • [Meister et al. 2013, Ament et al 2014, Hullin 2014]

èApplication-specific, approximations, point samples

  • [Jarabo et al. 2014]

è Time-resolved path integral formulation è Temporal progressive density estimations è Time-based importance sampling èPoint samples: Bidirectional path tracing, photon mapping

14

Transient Rendering

slide-15
SLIDE 15
  • Monte Carlo methods è Variance is aggravated in time

15

time Radiance

Transient Rendering

Challenges

slide-16
SLIDE 16
  • Monte Carlo methods è Variance is aggravated in time

16

time Radiance

Transient Rendering

Challenges

slide-17
SLIDE 17
  • Monte Carlo methods è Variance is aggravated in time

17

Transient Rendering

Challenges

Slow convergence

time

slide-18
SLIDE 18

18

Transient Rendering

Participating media

Classic RTE in rendering TIME-INDEPENDENT

slide-19
SLIDE 19

19

Transient Rendering

Participating media

Classic RTE in rendering TIME-INDEPENDENT

slide-20
SLIDE 20

20

Transient Rendering

Participating media

Transient RTE NEED TO ACCOUNT FOR TIME

slide-21
SLIDE 21

21

Transient Rendering

Participating media

Transient RTE NEED TO ACCOUNT FOR TIME

slide-22
SLIDE 22

22

Transient Rendering

Participating media

Optical path IOR Scattering events

Transient RTE NEED TO ACCOUNT FOR TIME

slide-23
SLIDE 23

23

time Radiance

Transient Rendering

Participating media

[Jarabo 2014]à Point samples (BDPT, photon mapping)

à SPARSE SAMPLES IN TIME

slide-24
SLIDE 24

[Jarabo 2014]à Point samples (BDPT, photon mapping)

à SPARSE SAMPLES IN TIME

24

Transient Rendering

Participating media

time Radiance

NEED DENSER TEMPORAL SAMPLING

slide-25
SLIDE 25

[Jarosz et al. 2011a, 2011b]: Steady-state media rendering

25

Steady-state - Photon Beams

slide-26
SLIDE 26
  • 1. Stores photon trajectories on a BEAMS MAP

26

Steady-state - Photon Beams

slide-27
SLIDE 27
  • 1. Stores photon trajectories on a BEAMS MAP
  • 2. Performs ray-beam density estimations

27

Steady-state - Photon Beams

slide-28
SLIDE 28

Why photon beams for transient rendering?

28

Transient Photon Beams

Full photon trajectories

Denser sampling the temporal domain

time Radiance

slide-29
SLIDE 29

Why photon beams for transient rendering?

29

Transient Photon Beams

Full photon trajectories

Denser sampling the temporal domain

time Radiance

Closed form density estimations

Arbitrary temporal resolution

slide-30
SLIDE 30

30

Transient Photon Beams

Optical path IOR Scattering events

  • 1. Tracing: Sample Transient RTE

èStore beam starting time

slide-31
SLIDE 31

31

  • 2. Rendering: Spatio-temporal

density estimations

Transient Photon Beams

Optical path IOR Scattering events

  • 1. Tracing: Sample Transient RTE

èStore beam starting time

Spatial KDE (time-resolved)

slide-32
SLIDE 32

32

  • 2. Rendering: Spatio-temporal

density estimations

Transient Photon Beams

Optical path IOR Scattering events

time Radiance

  • 1. Tracing: Sample Transient RTE

èStore beam starting time

Spatial KDE Temporal KDE

slide-33
SLIDE 33

Transient Photon Beams

slide-34
SLIDE 34

Transient Photon Beams

time Spatio-temporal slice

slide-35
SLIDE 35

Transient Photon Beams

time Spatio-temporal slice

slide-36
SLIDE 36

Transient Photon Beams

time Spatio-temporal slice

slide-37
SLIDE 37

Transient Photon Beams

time Spatio-temporal slice

BIAS

slide-38
SLIDE 38

PROGRESSIVE APPROACH

38

slide-39
SLIDE 39

39

Spatial density estimations Temporal density estimations

Progressive Transient Photon Beams

time Radiance

slide-40
SLIDE 40

40

Spatial density estimations Temporal density estimations

Progressive Transient Photon Beams

time Radiance

slide-41
SLIDE 41

41

Spatial density estimations Temporal density estimations

Progressive Transient Photon Beams

time Radiance

slide-42
SLIDE 42

42

Spatial density estimations Temporal density estimations

1D spatial kernel

Progressive Transient Photon Beams

time Radiance

1D temporal kernel

slide-43
SLIDE 43

43

Spatial density estimations Temporal density estimations

1D spatial kernel

Progressive Transient Photon Beams

time Radiance

1D temporal kernel

slide-44
SLIDE 44

Progressive Transient Photon Beams

time Spatio-temporal slice 24 iterations

slide-45
SLIDE 45

Progressive Transient Photon Beams

time Spatio-temporal slice 24 iterations

slide-46
SLIDE 46

Progressive Transient Photon Beams

time Spatio-temporal slice 2000 iterations

slide-47
SLIDE 47

RESULTS

47

slide-48
SLIDE 48

Soccer

40M beams (2000 iterations x 20k beams/iteration)

48

slide-49
SLIDE 49
slide-50
SLIDE 50

Pumpkin

Progressive transient PT [Jarabo 2014] vs. Our method (equal–time comparsion)

50

slide-51
SLIDE 51

Steady state

Pumpkin – Equal-time comparison

slide-52
SLIDE 52

[Jarabo et al. 2014]

Pumpkin – Equal-time comparison

Our method

slide-53
SLIDE 53

Pumpkin – Equal-time comparison

Steady state Transient state Radiance time (ns)

slide-54
SLIDE 54

Pumpkin – Equal-time comparison

Transient state Radiance time (ns) [Jarabo 2014] Ours Steady state

slide-55
SLIDE 55

Pumpkin – Equal-time comparison

Transient state Radiance time (ns) [Jarabo 2014] Ours Steady state

slide-56
SLIDE 56

Pumpkin – Equal-time comparison

Transient state Radiance time (ns) [Jarabo 2014] Ours Steady state

slide-57
SLIDE 57

Juice

24M beams (1200 iterations x 20k beams/iteration)

58

slide-58
SLIDE 58
slide-59
SLIDE 59

Conclusion

  • Robust method for low-variance time-resolved participating media
  • Render complex time-resolved effects
  • Consistent approach
  • Optimal 1D x 1D spatio-temporal kernel reduction ratios

60

slide-60
SLIDE 60

What next?

  • Introduce time-based importance sampling [Jarabo et al. 2014]
  • Extend to hybrid methods, all volumetric estimators
  • Improve temporal reconstruction

61

slide-61
SLIDE 61

Thanks!

62