High Dynamic Range Lighting March 24, 2004 Paul Debevec, USC - PDF document

“High Dynamic Range Lighting” March 24, 2004 Paul Debevec, USC Institute for Creative Technologies High Dynamic Range Lighting High Dynamic Range Lighting Paul Debevec Paul Debevec University of Southern California University of Southern California Institute for Creative Technologies Institute for Creative Technologies March 24, 2004 March 24, 2004 5:30 – 6:30 pm 5:30 – 6:30 pm www.debevec.org/IBL2004/ www.debevec.org/IBL2004/ Scenes lit with point light sources lack realism… Real-World HDR Lighting Environments Real-World HDR Lighting Environments Illuminating Objects using Illuminating Objects using Funston Eucalyptus Measurements of Real Light Measurements of Real Light Beach Grove Light Light Environment Environment assigned “glow” assigned “glow” material material property in property in Greg Ward’s Greg Ward’s Object RADIANCE Object RADIANCE system. system. Grace Uffizi Cathedral Gallery http://radsite.lbl.gov/radiance/ http://radsite.lbl.gov/radiance/ Lighting Environments from the Light Probe Image Gallery: Lighting Environments from the Light Probe Image Gallery: http://www.debevec.org/CGAIBL/ http://www.debevec.org/CGAIBL/ http://www.debevec.org/Probes/ http://www.debevec.org/Probes/ Elements of HDRI and IBL Elements of HDRI and IBL High Dynamic Range (HDR) Images High Dynamic Range (HDR) Images Pixels beyond 0-255 Pixels beyond 0-255 Pixel proportional to light levels Pixel proportional to light levels Light Probe Images Light Probe Images Omnidirectional HDR images, or Omnidirectional HDR images, or HDR environment maps HDR environment maps Global Illumination Global Illumination Illuminating CG objects with Illuminating CG objects with images of incident illumination images of incident illumination Lighting with real illumination environments yields greater realism 2004 Game Developer’s Conference 1

“High Dynamic Range Lighting” March 24, 2004 Paul Debevec, USC Institute for Creative Technologies Dynamic Range in the Real Dynamic Range in the Real IBL Tutorial IBL Tutorial World World In Jan/Feb In Jan/Feb Office interior Computer Graphics Computer Graphics Indirect light from window and Applications and Applications 1/60 th sec shutter and the SIGGRAPH and the SIGGRAPH 2002 IBL Course 2002 IBL Course f/5.6 aperture Notes Notes 0 ND filters 0dB gain www.debevec.org www.debevec.org Sony VX2000 video camera Dynamic Range in the Real Dynamic Range in the Real Dynamic Range in the Real Dynamic Range in the Real World World World World Outside in the shade Outside in the sun 1/1000 th sec shutter 1/1000 th sec shutter f/5.6 aperture f/11 aperture 0 ND filters 0 ND filters 0dB gain 0dB gain 16 times the light as inside 64 times the light as inside Dynamic Range in the Real Dynamic Range in the Real Dynamic Range in the Real Dynamic Range in the Real World World World World Straight at the sun Very dim room 1/10,000 th sec shutter 1/4 th sec shutter f/11 aperture f/1.6 aperture 13 stops ND filters 0 stops ND filters 0dB gain 18dB gain 1/1500 th the light than inside 5,000,000 times the light as inside 2004 Game Developer’s Conference 2

“High Dynamic Range Lighting” March 24, 2004 Paul Debevec, USC Institute for Creative Technologies High-Dynamic Range Photography High-Dynamic Range Photography 2,000,000,000 Dynamic Range Dynamic Range in the Real in the Real World World 400,000 25,000 1500 300,000 : 1 300,000 : 1 Debevec and Malik, Recovering High Dynamic Range Radiance Maps from Visualization: Greg Ward 1 Photographs, SIGGRAPH 97 Gamma 2.2 graph Gamma 2.2 graph Implications: A 128 is less than ¼ as 255 m bright as 255 o u 128 is more than 4 n t times as bright as 64 175 is twice as o 175 f bright as 128 L 93 is half as bright 128 i as 128 g 64 www.debevec.org/HDRShop 0 h “128 + 128 = 175” t 0 Pixel Value 255 “128 / 2 = 93” See also Charles Poynton’s Gamma FAQ: Chris Tchou and Paul Debevec. HDR Shop . SIGGRAPH 2001 Technical Sketch http://www.inforamp.net/~poynton/GammaFAQ.html DirectX 9 HDR Data Formats DirectX 9 HDR Data Formats 32-bit floating point textures 32-bit floating point textures • D3DFMT_A32B32G32R32F / D3DFMT_R32F • D3DFMT_A32B32G32R32F / D3DFMT_R32F HDR Image File Formats HDR Image File Formats • IEEE compatible • IEEE compatible 16-bit floating point textures 16-bit floating point textures • D3DFMT_A16B16G16R16F • D3DFMT_A16B16G16R16F • saves memory bandwidth • saves memory bandwidth • often sufficient dynamic range and • often sufficient dynamic range and precision precision 2004 Game Developer’s Conference 3

“High Dynamic Range Lighting” March 24, 2004 Paul Debevec, USC Institute for Creative Technologies HDR Formats: RADIANCE Format HDR Formats: Portable FloatMap (.pfm) HDR Formats: RADIANCE Format HDR Formats: Portable FloatMap (.pfm) (.pic, .hdr) (.pic, .hdr) 12 bytes per pixel, 4 for each channel 12 bytes per pixel, 4 for each channel Greg Ward’s “Real Pixels” format Greg Ward’s “Real Pixels” format 32 bits / pixel 32 bits / pixel Red Green Blue Exponent Red Green Blue Exponent sign exponent mantissa (145, 215, 87, 103) = (145, 215, 87, 149) = (145, 215, 87, 149) = (145, 215, 87, 103) = Text header similar to Jeff Poskanzer’s .ppm (145, 215, 87) * 2^(103-128) = (145, 215, 87) * 2^(149-128) = (145, 215, 87) * 2^(149-128) = (145, 215, 87) * 2^(103-128) = image format: PF (1190000, 1760000, 713000) (1190000, 1760000, 713000) (0.00000432, 0.00000641, 0.00000259) (0.00000432, 0.00000641, 0.00000259) 768 512 1 <binary image data> Floating Point TIFF similar Floating Point TIFF similar Ward, Greg. "Real Pixels," in Graphics Gems IV, edited by James Arvo, Academic Press, 1994 HDR Formats: Ward’s LogLuv TIFF HDR Formats: Ward’s LogLuv TIFF HDR Formats: ILM’s OpenEXR (.exr) HDR Formats: ILM’s OpenEXR (.exr) based on human color perception based on human color perception 6 bytes per pixel, 2 for each channel, compressed 6 bytes per pixel, 2 for each channel, compressed 24 bits:10 for log luminance 14 for chromaticity index 32 bits:15 log luminance 8 u chrominance 8 v chrominance 1 sign sign exponent mantissa • Several lossless compression options, 2:1 typical Larson, G.W., “Overcoming Gamut and Dynamic Range Limitations • Compatible with the “half” datatype in NVidia's Cg in Digital Images,” Proceedings of the Sixth Color Imaging • Supported natively on GeForce FX and Quadro FX Conference, November 1998. http://positron.cs.berkeley.edu/~gwlarson/pixformat/tiffluv.html • Available at: http://www.openexr.net/ Panoramic (Omnidirectional) Panoramic (Omnidirectional) Photography Photography Light Probe Images: Light Probe Images: Capturing Real-World Illumination Capturing Real-World Illumination Other techniques: - Panoramic Stitching (Realviz Stitcher) - Fisheye Images - Scanning Panoramic Cameras (Panoscan, Spheron) 2004 Game Developer’s Conference 4

“High Dynamic Range Lighting” March 24, 2004 Paul Debevec, USC Institute for Creative Technologies HDR Image of a HDR Image of a Comparison: HDRI versus Comparison: HDRI versus Mirrored Ball single image lighting Mirrored Ball single image lighting (60,40,35) (60,40,35) (620,890,1300) (620,890,1300) (5700,8400,11800) (5700,8400,11800) (18,17,19) (11700,7300,2600) (18,17,19) (11700,7300,2600) Assembled from ten digital images, Assembled from ten digital images, ∆ t = 1/4 to 1/10000 sec ∆ t = 1/4 to 1/10000 sec Acquiring the Light Probe Acquiring the Light Probe Image-Based Lighting: Image-Based Lighting: Illuminating Synthetic Objects with Illuminating Synthetic Objects with Real Light Real Light Rendering with Natural Light, SIGGRAPH 98 Assembling the Light Probe Assembling the Light Probe RNL Environment mapped onto interior of large cube 2004 Game Developer’s Conference 5

“High Dynamic Range Lighting” March 24, 2004 Paul Debevec, USC Institute for Creative Technologies RNL Example RNL Example RNL Example RNL Example Renderer Output Renderer Output Defocus & Glare Added Defocus & Glare Added RNL Example RNL Example RNL Example RNL Example Soft Focus Added Light Falloff (Vignetting) Added Soft Focus Added Light Falloff (Vignetting) Added Real-Time RNL Real-Time RNL Jason Mitchell, John Isidoro, Alex Vlachos Jason Mitchell, John Isidoro, Alex Vlachos Rendered in Real Time on ATI RADEON™ 9700 Rendered in Real Time on ATI RADEON™ 9700 www.debevec.org/RNL 2004 Game Developer’s Conference 6