SlideShare a Scribd company logo
➢Introduction: History and Physiology
 Display Taxonomy
 Multi-view Rendering using OpenGL/GLSL
 Designing Content for Glasses-free 3D Displays
 Emerging Technology
Stereo and 3D Displays
Monocular Depth Cues Supported by Conventional Displays
 relative and familiar size
 perspective and occlusion
 texture gradient, shading and lighting, atmospheric effects
Limitations of Conventional Displays
Limitations of Conventional DisplaysLimitations of Conventional Displays
Monocular Depth Cues with Conventional Displays
 relative and familiar size
 perspective and occlusion
 texture gradient, shading and lighting, atmospheric effects
Additional Monocular Depth Cues
 motion parallax [Hermann von Helmholtz, 1866]
 accommodation
What is missing?
Binocular Depth Cues
 retinal disparity [Charles Wheatstone, 1838]
 convergence
“It being thus established that the mind perceives an object of three dimensions by means of the
two dissimilar pictures projected by it on the two retinae, the following question occurs: What
would be the visual effect of simultaneously presenting to each eye, instead of the object itself, its
projection on a plane surface as it appears to that eye?”
Binocular Depth Cues
American Civil War-era stereoscopic photos
• Available from the US library of congress
•http://www.loc.gov/pictures/search - Search for “stereographs civil war prints”
• Lincoln in 3D
• Selection of stereographs converted to red-
cyan anaglyph images
•John J. Richter: ISBN 978-0811872317
Interesting Historical Example
l
The HVS can ignore conflicting or missing depth cues
l
Understand depth in 2D (monocular) video
l
Perceive shape in “noise”
Ponzo Illusion: © Walt Anthony 2006magiceye.com
Conflicting Cues
Stereo and 3D Displays
 Introduction: History and Physiology
➢Display Taxonomy
 Multi-view Rendering using OpenGL/GLSL
 Designing Content for Glasses-free 3D Displays
 Emerging Technology
Taxonomy of 3D Displays:
Glasses-bound vs. Unencumbered Designs
Glasses-bound
Stereoscopic
Immersive
(blocks direct-viewing of real world)
See-through
(superimposes synthetic images onto real world)
Head-mounted
(eyepiece-objective and microdisplay)
Multiplexed
(stereo pair with same display surface)
Spatially-multiplexed (field-concurrent)
(color filters, polarizers, autostereograms, etc.)
Temporally-multiplexed (field-sequential)
(LCD shutter glasses)
Unencumbered
Automultiscopic
Parallax-based
(2D display with light-directing elements)
Volumetric
(directly illuminate points within a volume)
Holographic
(reconstructs wavefront using 2D element)
Parallax Barriers
(uniform array of 1D slits or 2D pinhole arrays)
Integral Imaging
(lenticular sheets or fly’s eye lenslet arrays)
Multi-planar
(time-sequential projection onto swept surfaces)
Transparent Substrates
(intersecting laser beams, fog layers, etc.)
Static
(holographic films)
Dynamic
(holovideo)
Taxonomy adapted from Hong Hua
Taxonomy of 3D Displays:
Immersive Head-mounted Displays (HMDs)
Glasses-bound
Stereoscopic
Immersive
(blocks direct-viewing of real world)
Head-mounted
(eyepiece-objective and microdisplay)
Multiplexed
(stereo pair with same display surface)
Taxonomy of 3D Displays:
See-through Head-mounted Displays (HMDs)
Glasses-bound
Stereoscopic
Immersive
(blocks direct-viewing of real world)
See-through
(superimposes synthetic images onto real world)
Head-mounted
(eyepiece-objective and microdisplay)
Multiplexed
(stereo pair with same display surface)
Taxonomy of 3D Displays:
Spatial Multiplexing (e.g., Anaglyphs)
Glasses-bound
Stereoscopic
Immersive
(blocks direct-viewing of real world)
See-through
(superimposes synthetic images onto real world)
Head-mounted
(eyepiece-objective and microdisplay)
Multiplexed
(stereo pair with same display surface)
Spatially-multiplexed (field-concurrent)
(color filters, polarizers, etc.)
Taxonomy of 3D Displays:
Temporal Multiplexing (e.g., Shutter Glasses)
Glasses-bound
Stereoscopic
Immersive
(blocks direct-viewing of real world)
See-through
(superimposes synthetic images onto real world)
Head-mounted
(eyepiece-objective and microdisplay)
Multiplexed
(stereo pair with same display surface)
Spatially-multiplexed (field-concurrent)
(color filters, polarizers, autostereograms, etc.)
Temporally-multiplexed (field-sequential)
(LCD shutter glasses)
Taxonomy of 3D Displays:
Parallax Barriers
Unencumbered
Automultiscopic
Parallax-based
(2D display with light-directing elements)
Volumetric
(directly illuminate points within a volume)
Holographic
(reconstructs wavefront using 2D element)
Parallax Barriers
(uniform array of 1D slits or 2D pinhole arrays)
NewSight MV-42AD3 42''
(1920x1080, 1x8 views)
Taxonomy of 3D Displays:
Integral Imaging
Unencumbered
Automultiscopic
Parallax-based
(2D display with light-directing elements)
Volumetric
(directly illuminate points within a volume)
Holographic
(reconstructs wavefront using 2D element)
Parallax Barriers
(uniform array of 1D slits or 2D pinhole arrays)
Integral Imaging
(lenticular sheets or fly’s eye lenslet arrays)
Alioscopy 3DHD 42''
(1920x1200, 1x8 views)
Taxonomy of 3D Displays:
Multi-planar Volumetric Displays
Unencumbered
Automultiscopic
Parallax-based
(2D display with light-directing elements)
Volumetric
(directly illuminate points within a volume)
Holographic
(reconstructs wavefront using 2D element)
Parallax Barriers
(uniform array of 1D slits or 2D pinhole arrays)
Integral Imaging
(lenticular sheets or fly’s eye lenslet arrays)
Multi-planar
(time-sequential projection onto swept surfaces)
Taxonomy of 3D Displays:
Transparent-substrate Volumetric Displays
Unencumbered
Automultiscopic
Parallax-based
(2D display with light-directing elements)
Volumetric
(directly illuminate points within a volume)
Holographic
(reconstructs wavefront using 2D element)
Parallax Barriers
(uniform array of 1D slits or 2D pinhole arrays)
Integral Imaging
(lenticular sheets or fly’s eye lenslet arrays)
Multi-planar
(time-sequential projection onto swept surfaces)
Transparent Substrates
(intersecting laser beams, fog layers, etc.)
Taxonomy of 3D Displays:
Static Holograms
Unencumbered
Automultiscopic
Parallax-based
(2D display with light-directing elements)
Volumetric
(directly illuminate points within a volume)
Holographic
(reconstructs wavefront using 2D element)
Parallax Barriers
(uniform array of 1D slits or 2D pinhole arrays)
Integral Imaging
(lenticular sheets or fly’s eye lenslet arrays)
Multi-planar
(time-sequential projection onto swept surfaces)
Transparent Substrates
(intersecting laser beams, fog layers, etc.)
Static
(holographic films)
capture reconstruction
Taxonomy of 3D Displays:
Dynamic Holograms (Holovideo)
Unencumbered
Automultiscopic
Parallax-based
(2D display with light-directing elements)
Volumetric
(directly illuminate points within a volume)
Holographic
(reconstructs wavefront using 2D element)
Parallax Barriers
(uniform array of 1D slits or 2D pinhole arrays)
Integral Imaging
(lenticular sheets or fly’s eye lenslet arrays)
Multi-planar
(time-sequential projection onto swept surfaces)
Transparent Substrates
(intersecting laser beams, fog layers, etc.)
Static
(holographic films)
Dynamic
(holovideo)
Tay et al.
[Nature, 2008]
MIT Media Lab Spatial Imaging Group
[Holovideo, 1989 – present]
Stereo and 3D Displays
 Introduction: History and Physiology
 Display Taxonomy
➢Multi-view Rendering using OpenGL/GLSL
 Designing Content for Glasses-free 3D Displays
 Emerging Technology
Overview:
GLSL: Programmable Pipeline
Fixed Function Pipeline
Simple 1-Slide Explanation!
Drawing APIDrawing API
Process VerticesProcess Vertices
Process PixelsProcess Pixels
FramebufferFramebuffer
Programmable Pipeline
Vertex ProgramVertex Program
Fragment ProgramFragment Program
l
Some graphics cards have support for stereo 3D (Not on mobile)
l
Double buffered stereo = Quad buffered
void
display(void)
{
glDrawBuffer(GL_BACK_LEFT);
<Draw left eye here>
glDrawBuffer(GL_BACK_RIGHT);
<Draw right eye here>
glutSwapBuffers();
}
int
main(int argc, char **argv)
{
glutInit(&argc, argv);
glutInitDisplayMode(
GLUT_DOUBLE | GLUT_RGB | GLUT_STEREO);
glutCreateWindow("stereo example");
glutDisplayFunc(display);
glutMainLoop();
return 0;
}
Anaglyphic Model Viewer:
Stereo 3D in OpenGL
Overview:
Multi-View Rendering in OpenGL
OpenGL
Draw Calls
Render
Standard Pipeline
Output
Multi-View Pipeline
Loop Over
Views
Backbuffer
Framebuffer
Object Array
Render View
Change
Camera
Screen:Memory:
Overview:
Multi-View Interlacing using GLSL Shaders
Framebuffer
Object
Array
Framebuffer
Object
Array
View 1
View 2
View 3
GLSL Program
Translate views
appropriately for
output device
Translate views
appropriately for
output device
BackbufferBackbuffer
Anaglyph
Glasses
Anaglyph
Glasses
LenticularLenticular
Shown in this course…
The model can apply to
many others
Multi-View Rendering in OpenGL:
Off-Axis Perspective Projection with glFrustum()
Output
Anaglyphic Model Viewer:
Anaglyph Compositing Algorithms
LL RR
3x3 Color Transform Matrix Pair3x3 Color Transform Matrix Pair
Full Color
Half Color
Optimized
L= R=
1 0 0
0 0 0
0 0 0
0 0 0
0 1 0
0 0 1
L= R=
0.299
0 0
0.587
0 0
0.114
0 0
0 0 0
0 1 0
0 0 1
L= R=
0 0 0
0.7 0 0
0.3 0 0
0 0 0
0 1 0
0 0 1
=
Source: http://3dtv.at/Knowhow/AnaglyphComparison_en.aspx
% read in images
ImL = imread('l.png');
ImR = imread('r.png');
% define "half color" matrices (see slides)
L = [.299 0 0
.587 0 0
.114 0 0];
R = [0 0 0
0 1 0
0 0 1];
% create a pixel x color array
ImL1d = double(reshape(ImL,prod(size(ImL(:,:,1))),3));
ImR1d = double(reshape(ImR,prod(size(ImR(:,:,1))),3));
% perform per pixel color permutation
ImL1d = ImL1d*L;
ImR1d = ImR1d*R;
Anaglyphic Model Viewer:
Making an Anaglyph Image in MATLAB
% convert back to 2d x color image
ImL = uint8(reshape(ImL1d,size(ImL)));
ImR = uint8(reshape(ImR1d,size(ImR)));
% create output
Iout = ImL + ImR;
anaglyph.m
Anaglyphic Model Viewer:
Demonstration
Stereo and 3D Displays
 Introduction: History and Physiology
 Display Taxonomy
 Multi-view Rendering using OpenGL/GLSL
➢Designing Content for Glasses-free 3D Displays
 Emerging Technology
• Stereo cameras (commercial and
improvised) are common
Source Data
Stereo Cameras
• Many researchers/hobbyists have built their own solutions to
capture light fields
• The PointGrey ProFusion is one of the few commercially available
PointGrey ProFusion
Stanford
Source Data
Light Field Cameras
MIT
Focal Plane
Example in Anaglyph Viewer
Screen
Virtual
Object
Placing objects farm from the
plane of focus is uncomfortable
Displays with limited DOF: objects further from screen plane are blurred
Rendering Tips
Accommodation & Object Placement
Focal Plane
Screen
Kirshnan, V. V., Stark, L. A heuristic model for the human vergence eye movement system, IEEE Trans. BioMed, 1977.
Limit distance of virtual object to viewer
Limit rate of change in scene distance
<1 m/s for
distant objects
Rendering Tips
Comfortable Vergence
Off-axis parallel projection Rotate and translate – Toe-in
• Puts ‘infinity’ at axis of rotation
•Requires user to focus beyond infinity
•Some find diverged eyes uncomfortable
Disadvantages of toe-in • Distortion between views
•Camera distance to most objects change
•Off axis objects will have different perspective
projection
Rendering Tips
Camera Model Choice
Avoid cases that cause a view to differ greatly from its neighbor
Left Right
Pillar pointing at viewer
Left Right
Clipped by edge of screen
Also watch out for
• Far objects clipped by near object
• Edges of hallways, tunnels, tubes,
etc
Also watch out for
• Don’t exit in front of screen plane
• More comfortable behind screen
• Avatar does a good job with this
Rendering Tips
Clipping and Degenerate Cases
Warzone 2100: GL Game Conversion
Stereo and 3D Displays
 Introduction: History and Physiology
 Display Taxonomy
 Multi-view Rendering using OpenGL/GLSL
 Designing Content for Glasses-free 3D Displays
➢Emerging Technology
Tensor Display
Tensor Display
Stereo and 3D Displays Resources
SIGGRAPH 2010/2011 Course: BYO3D
http://web.media.mit.edu/~mhirsch/byo3d/index.html
 Long-form slides
 Code and examples
display
blocks
Display Blocks blog
http://displayblocks.org
 Tutorials
 Building blocks explained
Gordon Wetzstein
gordonw@media.mit.edu
Matt Hirsch
mhirsch@media.mit.edu

More Related Content

PPTX
PPTX
SIGGRAPH 2012 Computational Display Course - 3 Computational Light Field Disp...
PPTX
HR3D: Content Adaptive Parallax Barriers
PDF
Holus - Glassless 3d system and modular apparatus
PPTX
BYO3D 2011: Content
PPTX
BYO3D 2011: Interlacing
PDF
3D television
PPTX
3 D (3-Dimensional) Glasses
SIGGRAPH 2012 Computational Display Course - 3 Computational Light Field Disp...
HR3D: Content Adaptive Parallax Barriers
Holus - Glassless 3d system and modular apparatus
BYO3D 2011: Content
BYO3D 2011: Interlacing
3D television
3 D (3-Dimensional) Glasses

Viewers also liked (20)

PPTX
Google Glass Breakdown
PPTX
Leap Motion Development (Rohan Puri)
PPTX
What is Media in MIT Media Lab, Why 'Camera Culture'
PDF
Raskar UIST Keynote 2015 November
PPTX
Multiview Imaging HW Overview
PDF
What is SIGGRAPH NEXT? Intro by Ramesh Raskar
PPTX
Coded Photography - Ramesh Raskar
PPT
Introduction to Camera Challenges - Ramesh Raskar
PPTX
PPTX
Raskar stanfordextremecompuimagingapr2016
PPTX
Compressed Sensing - Achuta Kadambi
PPTX
Light Field Photography Introduction
PPTX
Introduction to Photography
PPTX
Time of Flight Cameras - Refael Whyte
PDF
Developing "True HDR" for the iPhone
PPTX
Introduction to Light Fields
PPTX
PPTX
Computer Vision Introduction
PDF
Supporting Barack Obama for President
Google Glass Breakdown
Leap Motion Development (Rohan Puri)
What is Media in MIT Media Lab, Why 'Camera Culture'
Raskar UIST Keynote 2015 November
Multiview Imaging HW Overview
What is SIGGRAPH NEXT? Intro by Ramesh Raskar
Coded Photography - Ramesh Raskar
Introduction to Camera Challenges - Ramesh Raskar
Raskar stanfordextremecompuimagingapr2016
Compressed Sensing - Achuta Kadambi
Light Field Photography Introduction
Introduction to Photography
Time of Flight Cameras - Refael Whyte
Developing "True HDR" for the iPhone
Introduction to Light Fields
Computer Vision Introduction
Supporting Barack Obama for President
Ad

Similar to Stereo and 3D Displays - Matt Hirsch (20)

PPTX
Unencumbered 3D Displays
PPTX
BYO3D 2011: Construction
PDF
3D Display Technology: VDC-Whitepaper
PDF
[Paper Presentation] FoveAR: Combining an Optically See-Through Near-Eye Disp...
PPTX
Emerging 3D Display Technologies
DOC
Evolution of the senses
PPTX
BYO3D 2011: Emerging Technology
PDF
Screenless displays seminar report
PPT
Presentatie Willem En Kevin
PDF
Game Engine Overview
PPT
Shadow Mapping with Today's OpenGL Hardware
PPTX
Glasses Based 3D Displays
PPTX
Keynote - SPIE Stereoscopic Displays & Applications 2014
PPTX
Binocular Eye Tracking and Calibration in Head-mounted Displays
PPTX
BYO3D 2011: History
PPT
computer graphics
DOC
Copy of 3 d report
PPTX
3 d display technique
PPTX
Capture and Rendering
RTF
3d television
Unencumbered 3D Displays
BYO3D 2011: Construction
3D Display Technology: VDC-Whitepaper
[Paper Presentation] FoveAR: Combining an Optically See-Through Near-Eye Disp...
Emerging 3D Display Technologies
Evolution of the senses
BYO3D 2011: Emerging Technology
Screenless displays seminar report
Presentatie Willem En Kevin
Game Engine Overview
Shadow Mapping with Today's OpenGL Hardware
Glasses Based 3D Displays
Keynote - SPIE Stereoscopic Displays & Applications 2014
Binocular Eye Tracking and Calibration in Head-mounted Displays
BYO3D 2011: History
computer graphics
Copy of 3 d report
3 d display technique
Capture and Rendering
3d television
Ad

More from Camera Culture Group, MIT Media Lab (11)

PPTX
Raskar Sig2017 Siggraph Achievement Award Talk
PPTX
Lost Decade of Computational Photography
PPTX
God’s Eye View: Will global AI empower us or destroy us? | Ramesh Raskar
PPTX
Dont follow the rainbow: How to avoid career traps that can lead you to fail,...
PDF
Raskar PhD and MS Thesis Guidance
PDF
Making Invisible Visible, Ramesh Raskar Keynote at Embedded Vision 2019
PPTX
Augmented Surgeons: AI AR for Anatome, Raskar Aria 2019
PPTX
Geo-spatial Research: Transition from Analysis to Synthesis
PDF
Split Learning versus Federated Learning for Data Transparent ML, Camera Cult...
PPTX
Unspoken Challenges in AR and XR
Raskar Sig2017 Siggraph Achievement Award Talk
Lost Decade of Computational Photography
God’s Eye View: Will global AI empower us or destroy us? | Ramesh Raskar
Dont follow the rainbow: How to avoid career traps that can lead you to fail,...
Raskar PhD and MS Thesis Guidance
Making Invisible Visible, Ramesh Raskar Keynote at Embedded Vision 2019
Augmented Surgeons: AI AR for Anatome, Raskar Aria 2019
Geo-spatial Research: Transition from Analysis to Synthesis
Split Learning versus Federated Learning for Data Transparent ML, Camera Cult...
Unspoken Challenges in AR and XR

Recently uploaded (20)

PPTX
Belch_12e_PPT_Ch18_Accessible_university.pptx
PDF
SIMNET Inc – 2023’s Most Trusted IT Services & Solution Provider
PDF
Digital Marketing & E-commerce Certificate Glossary.pdf.................
PDF
pdfcoffee.com-opt-b1plus-sb-answers.pdfvi
PDF
IFRS Notes in your pocket for study all the time
PPTX
2025 Product Deck V1.0.pptxCATALOGTCLCIA
PDF
Ôn tập tiếng anh trong kinh doanh nâng cao
PPTX
CkgxkgxydkydyldylydlydyldlyddolydyoyyU2.pptx
PPT
340036916-American-Literature-Literary-Period-Overview.ppt
PDF
Solara Labs: Empowering Health through Innovative Nutraceutical Solutions
PDF
BsN 7th Sem Course GridNNNNNNNN CCN.pdf
PPTX
Probability Distribution, binomial distribution, poisson distribution
PDF
Daniels 2024 Inclusive, Sustainable Development
PDF
Laughter Yoga Basic Learning Workshop Manual
PDF
Cours de Système d'information about ERP.pdf
PDF
Hindu Circuler Economy - Model (Concept)
PDF
Module 2 - Modern Supervison Challenges - Student Resource.pdf
PDF
SBI Securities Weekly Wrap 08-08-2025_250808_205045.pdf
PDF
NISM Series V-A MFD Workbook v December 2024.khhhjtgvwevoypdnew one must use ...
PDF
Nidhal Samdaie CV - International Business Consultant
Belch_12e_PPT_Ch18_Accessible_university.pptx
SIMNET Inc – 2023’s Most Trusted IT Services & Solution Provider
Digital Marketing & E-commerce Certificate Glossary.pdf.................
pdfcoffee.com-opt-b1plus-sb-answers.pdfvi
IFRS Notes in your pocket for study all the time
2025 Product Deck V1.0.pptxCATALOGTCLCIA
Ôn tập tiếng anh trong kinh doanh nâng cao
CkgxkgxydkydyldylydlydyldlyddolydyoyyU2.pptx
340036916-American-Literature-Literary-Period-Overview.ppt
Solara Labs: Empowering Health through Innovative Nutraceutical Solutions
BsN 7th Sem Course GridNNNNNNNN CCN.pdf
Probability Distribution, binomial distribution, poisson distribution
Daniels 2024 Inclusive, Sustainable Development
Laughter Yoga Basic Learning Workshop Manual
Cours de Système d'information about ERP.pdf
Hindu Circuler Economy - Model (Concept)
Module 2 - Modern Supervison Challenges - Student Resource.pdf
SBI Securities Weekly Wrap 08-08-2025_250808_205045.pdf
NISM Series V-A MFD Workbook v December 2024.khhhjtgvwevoypdnew one must use ...
Nidhal Samdaie CV - International Business Consultant

Stereo and 3D Displays - Matt Hirsch

  • 1. ➢Introduction: History and Physiology  Display Taxonomy  Multi-view Rendering using OpenGL/GLSL  Designing Content for Glasses-free 3D Displays  Emerging Technology Stereo and 3D Displays
  • 2. Monocular Depth Cues Supported by Conventional Displays  relative and familiar size  perspective and occlusion  texture gradient, shading and lighting, atmospheric effects Limitations of Conventional Displays
  • 3. Limitations of Conventional DisplaysLimitations of Conventional Displays Monocular Depth Cues with Conventional Displays  relative and familiar size  perspective and occlusion  texture gradient, shading and lighting, atmospheric effects
  • 4. Additional Monocular Depth Cues  motion parallax [Hermann von Helmholtz, 1866]  accommodation What is missing?
  • 5. Binocular Depth Cues  retinal disparity [Charles Wheatstone, 1838]  convergence “It being thus established that the mind perceives an object of three dimensions by means of the two dissimilar pictures projected by it on the two retinae, the following question occurs: What would be the visual effect of simultaneously presenting to each eye, instead of the object itself, its projection on a plane surface as it appears to that eye?” Binocular Depth Cues
  • 6. American Civil War-era stereoscopic photos • Available from the US library of congress •http://www.loc.gov/pictures/search - Search for “stereographs civil war prints” • Lincoln in 3D • Selection of stereographs converted to red- cyan anaglyph images •John J. Richter: ISBN 978-0811872317 Interesting Historical Example
  • 7. l The HVS can ignore conflicting or missing depth cues l Understand depth in 2D (monocular) video l Perceive shape in “noise” Ponzo Illusion: © Walt Anthony 2006magiceye.com Conflicting Cues
  • 8. Stereo and 3D Displays  Introduction: History and Physiology ➢Display Taxonomy  Multi-view Rendering using OpenGL/GLSL  Designing Content for Glasses-free 3D Displays  Emerging Technology
  • 9. Taxonomy of 3D Displays: Glasses-bound vs. Unencumbered Designs Glasses-bound Stereoscopic Immersive (blocks direct-viewing of real world) See-through (superimposes synthetic images onto real world) Head-mounted (eyepiece-objective and microdisplay) Multiplexed (stereo pair with same display surface) Spatially-multiplexed (field-concurrent) (color filters, polarizers, autostereograms, etc.) Temporally-multiplexed (field-sequential) (LCD shutter glasses) Unencumbered Automultiscopic Parallax-based (2D display with light-directing elements) Volumetric (directly illuminate points within a volume) Holographic (reconstructs wavefront using 2D element) Parallax Barriers (uniform array of 1D slits or 2D pinhole arrays) Integral Imaging (lenticular sheets or fly’s eye lenslet arrays) Multi-planar (time-sequential projection onto swept surfaces) Transparent Substrates (intersecting laser beams, fog layers, etc.) Static (holographic films) Dynamic (holovideo) Taxonomy adapted from Hong Hua
  • 10. Taxonomy of 3D Displays: Immersive Head-mounted Displays (HMDs) Glasses-bound Stereoscopic Immersive (blocks direct-viewing of real world) Head-mounted (eyepiece-objective and microdisplay) Multiplexed (stereo pair with same display surface)
  • 11. Taxonomy of 3D Displays: See-through Head-mounted Displays (HMDs) Glasses-bound Stereoscopic Immersive (blocks direct-viewing of real world) See-through (superimposes synthetic images onto real world) Head-mounted (eyepiece-objective and microdisplay) Multiplexed (stereo pair with same display surface)
  • 12. Taxonomy of 3D Displays: Spatial Multiplexing (e.g., Anaglyphs) Glasses-bound Stereoscopic Immersive (blocks direct-viewing of real world) See-through (superimposes synthetic images onto real world) Head-mounted (eyepiece-objective and microdisplay) Multiplexed (stereo pair with same display surface) Spatially-multiplexed (field-concurrent) (color filters, polarizers, etc.)
  • 13. Taxonomy of 3D Displays: Temporal Multiplexing (e.g., Shutter Glasses) Glasses-bound Stereoscopic Immersive (blocks direct-viewing of real world) See-through (superimposes synthetic images onto real world) Head-mounted (eyepiece-objective and microdisplay) Multiplexed (stereo pair with same display surface) Spatially-multiplexed (field-concurrent) (color filters, polarizers, autostereograms, etc.) Temporally-multiplexed (field-sequential) (LCD shutter glasses)
  • 14. Taxonomy of 3D Displays: Parallax Barriers Unencumbered Automultiscopic Parallax-based (2D display with light-directing elements) Volumetric (directly illuminate points within a volume) Holographic (reconstructs wavefront using 2D element) Parallax Barriers (uniform array of 1D slits or 2D pinhole arrays) NewSight MV-42AD3 42'' (1920x1080, 1x8 views)
  • 15. Taxonomy of 3D Displays: Integral Imaging Unencumbered Automultiscopic Parallax-based (2D display with light-directing elements) Volumetric (directly illuminate points within a volume) Holographic (reconstructs wavefront using 2D element) Parallax Barriers (uniform array of 1D slits or 2D pinhole arrays) Integral Imaging (lenticular sheets or fly’s eye lenslet arrays) Alioscopy 3DHD 42'' (1920x1200, 1x8 views)
  • 16. Taxonomy of 3D Displays: Multi-planar Volumetric Displays Unencumbered Automultiscopic Parallax-based (2D display with light-directing elements) Volumetric (directly illuminate points within a volume) Holographic (reconstructs wavefront using 2D element) Parallax Barriers (uniform array of 1D slits or 2D pinhole arrays) Integral Imaging (lenticular sheets or fly’s eye lenslet arrays) Multi-planar (time-sequential projection onto swept surfaces)
  • 17. Taxonomy of 3D Displays: Transparent-substrate Volumetric Displays Unencumbered Automultiscopic Parallax-based (2D display with light-directing elements) Volumetric (directly illuminate points within a volume) Holographic (reconstructs wavefront using 2D element) Parallax Barriers (uniform array of 1D slits or 2D pinhole arrays) Integral Imaging (lenticular sheets or fly’s eye lenslet arrays) Multi-planar (time-sequential projection onto swept surfaces) Transparent Substrates (intersecting laser beams, fog layers, etc.)
  • 18. Taxonomy of 3D Displays: Static Holograms Unencumbered Automultiscopic Parallax-based (2D display with light-directing elements) Volumetric (directly illuminate points within a volume) Holographic (reconstructs wavefront using 2D element) Parallax Barriers (uniform array of 1D slits or 2D pinhole arrays) Integral Imaging (lenticular sheets or fly’s eye lenslet arrays) Multi-planar (time-sequential projection onto swept surfaces) Transparent Substrates (intersecting laser beams, fog layers, etc.) Static (holographic films) capture reconstruction
  • 19. Taxonomy of 3D Displays: Dynamic Holograms (Holovideo) Unencumbered Automultiscopic Parallax-based (2D display with light-directing elements) Volumetric (directly illuminate points within a volume) Holographic (reconstructs wavefront using 2D element) Parallax Barriers (uniform array of 1D slits or 2D pinhole arrays) Integral Imaging (lenticular sheets or fly’s eye lenslet arrays) Multi-planar (time-sequential projection onto swept surfaces) Transparent Substrates (intersecting laser beams, fog layers, etc.) Static (holographic films) Dynamic (holovideo) Tay et al. [Nature, 2008] MIT Media Lab Spatial Imaging Group [Holovideo, 1989 – present]
  • 20. Stereo and 3D Displays  Introduction: History and Physiology  Display Taxonomy ➢Multi-view Rendering using OpenGL/GLSL  Designing Content for Glasses-free 3D Displays  Emerging Technology
  • 21. Overview: GLSL: Programmable Pipeline Fixed Function Pipeline Simple 1-Slide Explanation! Drawing APIDrawing API Process VerticesProcess Vertices Process PixelsProcess Pixels FramebufferFramebuffer Programmable Pipeline Vertex ProgramVertex Program Fragment ProgramFragment Program
  • 22. l Some graphics cards have support for stereo 3D (Not on mobile) l Double buffered stereo = Quad buffered void display(void) { glDrawBuffer(GL_BACK_LEFT); <Draw left eye here> glDrawBuffer(GL_BACK_RIGHT); <Draw right eye here> glutSwapBuffers(); } int main(int argc, char **argv) { glutInit(&argc, argv); glutInitDisplayMode( GLUT_DOUBLE | GLUT_RGB | GLUT_STEREO); glutCreateWindow("stereo example"); glutDisplayFunc(display); glutMainLoop(); return 0; } Anaglyphic Model Viewer: Stereo 3D in OpenGL
  • 23. Overview: Multi-View Rendering in OpenGL OpenGL Draw Calls Render Standard Pipeline Output Multi-View Pipeline Loop Over Views Backbuffer Framebuffer Object Array Render View Change Camera Screen:Memory:
  • 24. Overview: Multi-View Interlacing using GLSL Shaders Framebuffer Object Array Framebuffer Object Array View 1 View 2 View 3 GLSL Program Translate views appropriately for output device Translate views appropriately for output device BackbufferBackbuffer Anaglyph Glasses Anaglyph Glasses LenticularLenticular Shown in this course… The model can apply to many others
  • 25. Multi-View Rendering in OpenGL: Off-Axis Perspective Projection with glFrustum() Output
  • 26. Anaglyphic Model Viewer: Anaglyph Compositing Algorithms LL RR 3x3 Color Transform Matrix Pair3x3 Color Transform Matrix Pair Full Color Half Color Optimized L= R= 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 L= R= 0.299 0 0 0.587 0 0 0.114 0 0 0 0 0 0 1 0 0 0 1 L= R= 0 0 0 0.7 0 0 0.3 0 0 0 0 0 0 1 0 0 0 1 = Source: http://3dtv.at/Knowhow/AnaglyphComparison_en.aspx
  • 27. % read in images ImL = imread('l.png'); ImR = imread('r.png'); % define "half color" matrices (see slides) L = [.299 0 0 .587 0 0 .114 0 0]; R = [0 0 0 0 1 0 0 0 1]; % create a pixel x color array ImL1d = double(reshape(ImL,prod(size(ImL(:,:,1))),3)); ImR1d = double(reshape(ImR,prod(size(ImR(:,:,1))),3)); % perform per pixel color permutation ImL1d = ImL1d*L; ImR1d = ImR1d*R; Anaglyphic Model Viewer: Making an Anaglyph Image in MATLAB % convert back to 2d x color image ImL = uint8(reshape(ImL1d,size(ImL))); ImR = uint8(reshape(ImR1d,size(ImR))); % create output Iout = ImL + ImR; anaglyph.m
  • 29. Stereo and 3D Displays  Introduction: History and Physiology  Display Taxonomy  Multi-view Rendering using OpenGL/GLSL ➢Designing Content for Glasses-free 3D Displays  Emerging Technology
  • 30. • Stereo cameras (commercial and improvised) are common Source Data Stereo Cameras
  • 31. • Many researchers/hobbyists have built their own solutions to capture light fields • The PointGrey ProFusion is one of the few commercially available PointGrey ProFusion Stanford Source Data Light Field Cameras MIT
  • 32. Focal Plane Example in Anaglyph Viewer Screen Virtual Object Placing objects farm from the plane of focus is uncomfortable Displays with limited DOF: objects further from screen plane are blurred Rendering Tips Accommodation & Object Placement
  • 33. Focal Plane Screen Kirshnan, V. V., Stark, L. A heuristic model for the human vergence eye movement system, IEEE Trans. BioMed, 1977. Limit distance of virtual object to viewer Limit rate of change in scene distance <1 m/s for distant objects Rendering Tips Comfortable Vergence
  • 34. Off-axis parallel projection Rotate and translate – Toe-in • Puts ‘infinity’ at axis of rotation •Requires user to focus beyond infinity •Some find diverged eyes uncomfortable Disadvantages of toe-in • Distortion between views •Camera distance to most objects change •Off axis objects will have different perspective projection Rendering Tips Camera Model Choice
  • 35. Avoid cases that cause a view to differ greatly from its neighbor Left Right Pillar pointing at viewer Left Right Clipped by edge of screen Also watch out for • Far objects clipped by near object • Edges of hallways, tunnels, tubes, etc Also watch out for • Don’t exit in front of screen plane • More comfortable behind screen • Avatar does a good job with this Rendering Tips Clipping and Degenerate Cases
  • 36. Warzone 2100: GL Game Conversion
  • 37. Stereo and 3D Displays  Introduction: History and Physiology  Display Taxonomy  Multi-view Rendering using OpenGL/GLSL  Designing Content for Glasses-free 3D Displays ➢Emerging Technology
  • 40. Stereo and 3D Displays Resources SIGGRAPH 2010/2011 Course: BYO3D http://web.media.mit.edu/~mhirsch/byo3d/index.html  Long-form slides  Code and examples display blocks Display Blocks blog http://displayblocks.org  Tutorials  Building blocks explained Gordon Wetzstein gordonw@media.mit.edu Matt Hirsch mhirsch@media.mit.edu