Logistics Paper summaries on Tone Reproduction Tone Reproduction - PDF document

Logistics • Paper summaries on Tone Reproduction Tone Reproduction – Any takers? Computer Graphics as Virtual Photography Tone/Color Reproduction (On the home stretch) • Where are we? real camera photo Photographic Photography: scene (captures processing print – Described our scene during modeling light) – Simulated light transport during rendering processing – Captured and projected light from the scene onto a 2D plane during capture camera Computer 3D synthetic – Now we must convert this simulated light tone model Graphics: models image reproduction capture into an image for display (focuses simulated lighting) Tone Reproduction Tone Reproduction • Definition • Definition – Compressing the dynamic range of a scene’s – Dealing with Luminances / radiances luminances/radiances so that it can be displayed – Rendering will be displayed on a given device on a given device in such a way that minimizes – Minimize perceptual difference between real the perceptual difference between viewing the and created. scene and viewing the rendering of the scene.

Tone Reproduction Tone Reproduction • Radiance / Luminance • Luminance levels – Flux arriving at or leaving from a given point or surface in a given direction . – Radiance measured in W / m 2 /sr dA – Luminance measured in cd/m 2 (nit) Sky = 12400 nits Trees = 64 nits Tone Reproduction Tone Reproduction • Using 0 – 1 to indicate light intensity • Simple Linear tone reproduction – What does 1 mean? • CG tends to use intensity space of output device • Images optimized for a given output device. • In typical CG apps, tone reproduction is simply linear scaling. Light source = Searchlight Light source = firefly Tublin-Rushmeier (1993) Tone Reproduction Tone Reproduction • Why bother • Basic pipeline [Tumblin93] – Human response to light is neither simple nor linear. – Most display devices are not linear – Incorrect response modeling results in incorrect perception of results.

Tone Reproduction Tone / Color Reproduction • Basic solution • Response / Observer – How does a system (like the human visual system or photography) respond to the collected light • Display – How do we translate that response using a particular output device (like a CRT or printer) Response Models Human Visual Response • Applying response/observe model will result in the luminances as seen by your display observer. – I.e. Will be in luminance range of your output device. • Observer/Response Models – Human Visual System – Photographic Systems Human Visual Response Human Visual Response • Pupil • Levels of Brightness Response – Regulates the amount of light that gets to the retina – Scotopic • Photoreceptors • 10 -6 to 10 2 cd/m 2 /primarily rods – Rods – Photopic • 75 - 150 million • 0.01 to 10 8 cd/m 2 / Primarily Cones • sensitive to 10 -6 to 10 2 cd/m 2 (low light levels) • Achromatic (detects “brightness”) – Mesopic – Cones • 0.01 to 10 2 cd/m 2 • 6 - 7 million • Both rods and cones • sensitive to 0.01 to 10 8 cd/m 2 (high light levels) • Little known -- active area of research • Responsible for color vision

Human Visual Response Human Visual Response • Spectral Sensitivity • Spectral response – Human Visual System is sensitive to light in the wavelength range of approx. 350 - 700 nm. – Sensitivity changes dependent on illumination level Human Visual System Human Visual System • Acuity • Acuity also changes dependent on – Ability to luminance level resolve spatial detail • Snellen Chart – View from 20 ft away – Line 8 subtends 1 min of visual angle – People who can read this is said to have 20/20 vision Human Visual System Human Visual System • Response at different illumination levels • Adaptation – Our vision system has the ability to adapt to a given luminance level – Light Adaptation - from darkness to light – Dark Adaptation - from brightness to dark – Adaptation is gradual, not immediate

Human Visual System Human Visual System • Threshold Studies • Light adaptation – determines the threshold at which a person can notice the change between a light sample given a certain background luminance. Human Visual System Human Visual System • Light adaptation • Dark adaptation Human Visual System Human Visual System • Ferwerda’s model • Dark Adaptation – Scales luminances as to preserve perceived contrast using psychophysical data as a guide. • L w = mL d – Different models for scotopic and photopic vision with slider to blend the two to simulate mesopic vision. • m will vary dependent upon whether scene is in scotopic, photopic, or mesopic range.

Human Visual System Photographic Response • CG Tone Reproduction Operators • An alternative to modeling visual response directly. – “Tone Reproduction and Physically Based Spectral Rendering” by Devlin, et al. • Models response to photographic materials. EUROGRAPHICS 2002. – Questions? Photographic Response Photographic Response • Why bother with photographic model? • Print Photography – Far better understood than human visual system. Camera Film Process Negative – Optimized for human viewing – Artistic photography – Composition of CG elements with scenes Printer Paper Print Process captured on film. Photographic Response Photographic Response • Brightness Response - high level response • Photographic Materials of an emulsion to light – Comprised of microscopic grains of silver • Spectral Sensitivity - Response of a material halide in a gelatin (emulsion) to different wavelengths of light – Latent image formed when exposed to light • Acuity - Level at which material can – Silver halide converted to metallic silver during reproduce spatial details processing. • Graininess - Observed variation due to grain – Converted silver results in opacity distribution

Photographic Response Photographic Response • Sensitometry exposure density – The science of measuring the sensitivity of E = It D = log (O) photographic materials I = Illuminance (lux) O = opacity = 1 / T – Each characteristic has its own unique t = time (sec) T = transmission sensitometric measure. E = exposure (lux-sec) = I t / I o I t = transmitted light I o = incident light Photographic Response Photographic Response gamma - slope of region II speed - indicates • Brightness Response / Characteristic Curve gives contrast range sensitivity to light 3 3 I - toe III II - straight line y section IV 2 2 t i III - shoulder s n II γ IV - area of e γ D solarization 1 1 I γ - gamma 0 0 -2 S 0 2 Log Exposure -2 0 2 Photographic Response - Speed Photographic Response - Gamma Original Original 100 Speed Film Low Contrast 800 Speed Film Medium Contrast High Contrast 400 Speed Film

Photographic Response Photographic Response Spectral Response Spectral Response 100 panchromatic 0 100 orthochromatic 0 Original Panchromatic Blue Sensitive 100 blue sensitive 0 300 400 500 600 Photographic Response - Acuity Photographic Response - Acuity point spread function modulation transfer function 100 80 60 (%) 40 Without MTF With MTF With MTF & Grain 20 0 0 40 80 120 spatial freq. (cycles/mm) Photographic Response - Grain Photographic Response - Grain Selwyn Granularity: rms deviation: N Σ ( ∆ D i ) 2 2 = 1 (2A) σ σ G = Original w/grain ∆ D i = deviation of Magnified 2x Magnified 4x A = area of scanning sample i from aperture the mean

Photographic Response Modeling Photographic Response • High level description of photographic • Uses sensitometric measures to model response characteristics of photo materials – Can model process at grain level but • Physically based impractical to do so for our purposes. • Built using an imaging pipeline where each – Thankfully, all of these sensiometic module in the pipe represents an image measurements are available for photo materials processing operation. from the manufacturer. Modeling Photographic Response Modeling Photographic Response input image • Must run thru pipeline twice, once for capture on film and once for printing spectral expose resolution sensitivity • Result of model – Image of floats [0, 1] exposure – Represents transmission or reflection values convert to density granularity transmission/ response reflection negative density conversion or print Modeling Photographic Response Modeling Photographic Response • Recall • Virtual Darkroom Applet – Prints are reflective media – Are not visible unless illuminated – values from model must be modified to account for the luminance / color characteristic of the assumed print illumination – Thankfully, luminance / color of CRTs approximates normal interior viewing conditions fairly well.

Tone Reproduction Tone Reproduction • An almost final word on Tone Reproduction • Summary – Things we did not discuss today: – Means of compressing dynamic range of scene • Display model to fit that of display • Color • Viewing conditions also affect perception – Observer / Response Model – TR Operator should also make modifications if viewing • Human Visual System conditions of world observer does not match that of display observer • Photographic Systems – These issues + a complete photographic model, will be – Device Model discussed next time. – Breaktime…