Color Science Fundamentals in Motion Imaging

Transcription

1 Color Science Fundamentals in Motion Imaging Jaclyn Pytlarz Dolby Laboratories Inc. SMPTE Essential Technology Concepts Series of ten 60- to 90-minute online planned for 2019 Designed to present the fundamental principles and concepts that are foundational to the technology and workflows in media & entertainment Interactive webcasts Free to everyone Sessions are recorded for on-demand playback 2019 SMPTE SMPTE SMPTE

2 Your Host Joel E. Welch Director of Education SMPTE 2019 SMPTE Views and opinions expressed during this SMPTE Webcast are those of the presenter(s) and do not necessarily reflect those of SMPTE or SMPTE Members. This webcast is presented for informational purposes only. Any reference to specific companies, products or services does not represent promotion, recommendation, or endorsement by SMPTE 2019 SMPTE SMPTE

3 Today s Guest Speaker Jaclyn Pytlarz Senior Engineer Dolby Laboratories 2019 SMPTE What We ll Cover How do we see color and what is chromaticity? What are color spaces and how are they used? How can we use color science/perception to optimize video quality? Operational practices why does it matter to you? 2019 SMPTE SMPTE

4 Light Spectrum Additive/Subtractive Systems 2019 SMPTE

5 The Human Eye Long, Medium, and Short Wavelength Cones (daylight color vision) Cones in the eye: Rods (night monochrome vision) 2019 SMPTE Goal: Create Model of Color Vision Color Matching Experiment A B C 2019 SMPTE SMPTE

6 ҧ ҧ Cone Spectral Sensitivity r, g, ҧ തb were results of the color matching experiment: cone spectral sensitivities 2019 SMPTE Color Matching Functions (CMF) x, തy, zҧ were derived from linear combinations of the cone spectral sensitivities: color matching functions 2019 SMPTE SMPTE

7 Cone Spectral Sensitivity V(λ) = Luminosity Function 2019 SMPTE Spectral Power Distribution 2019 SMPTE SMPTE

8 Spectral Power Distribution Display SPD X = k න xҧ SPD Color Matching Functions Y = k න തy SPD Z = k න zҧ SPD 2019 SMPTE Spectral Power Distribution Spectroradiometer/Colorimeter X = k න xҧ SPD Y = k න തy SPD Z = k න zҧ SPD 2019 SMPTE SMPTE

9 Chromaticity Coordinates Chromaticity Diagram Tristimulus (3D) X = k න xҧ SPD Y = k න തy SPD Z = k න zҧ SPD Chromaticity (2D) Top Down View x = y = X X + Y + Z Y X + Y + Z 2019 SMPTE Color Gamuts Color gamuts are defined with xy chromaticity coordinates for red, green, blue and white: - ITU-R BT ITU-R BT DCI-P3 - ITU-R BT.2100 Chromaticity Diagram 2019 SMPTE SMPTE

10 Color Gamuts Chromaticity Diagram Color gamuts are defined with xy chromaticity coordinates for red, green, blue and white: - ITU-R BT ITU-R BT DCI-P3 - ITU-R BT SMPTE Color Gamuts Color gamuts are defined with xy chromaticity coordinates for red, green, blue and white: - ITU-R BT ITU-R BT DCI-P3 - ITU-R BT.2100 Chromaticity Diagram 2019 SMPTE SMPTE

11 Color Gamuts Chromaticity Diagram Color gamuts are defined with xy chromaticity coordinates for red, green, blue and white: - ITU-R BT ITU-R BT DCI-P3 - ITU-R BT SMPTE SMPTE Essential Technology Concepts Let s Take a Break for Questions 2019 SMPTE SMPTE

12 White Points White Point = chromaticity of code value [1,1,1] Chromaticity Diagram Not necessarily what you see as white! Each movie/show has it s own adapting white point O DCI O D SMPTE Chromatic Adaptation 2019 SMPTE SMPTE

13 Chromatic Adaptation 2019 SMPTE Chromatic Adaptation 2019 SMPTE SMPTE

14 Color Space 3D Space Describes code value relationship to XYZ BT.709 Color Space [1/5, 3/5, 4/5] in BT.709 means something different than [1/5, 3/5, 4/5] in BT SMPTE Color Volume 3D Space Describes the number of uniquely visible colors SDR: Min=0.01 cd/m 2 Max=100 cd/m 2 BT.1886 EOTF 2019 SMPTE SMPTE

15 Color Volume 3D Space Describes the number of uniquely visible colors HDR: Min cd/m 2 Max 1000 cd/m 2 PQ (ST 2084) or HLG EOTF 2019 SMPTE Color Space Conversion - Remapping to a new coordinate system - Linear transform (3x3 matrix) Chromaticity Diagram R_2020 R_709 G_2020 = M * G_709 B_2020 B_ SMPTE SMPTE

16 Color Encoding System Goal: Separate luma from chroma information: - Yuv - YPbPr - YCbCr - ICtCp Relative to Current Color Space Independent of Current Color Space 2019 SMPTE Color Encoding System Goal: Separate luma from chroma information: Y R Cb = M * G Cr B 2019 SMPTE SMPTE

17 Color Encoding System Original - 20x20 blur filter - Apply in Y CbCr - On Y - On CbCr 2019 SMPTE Color Encoding System Blur on Y - 20x20 blur filter - Apply in Y CbCr - On Y - On CbCr 2019 SMPTE SMPTE

18 Color Encoding System Blur on CbCr - 20x20 blur filter - Apply in Y CbCr - On Y - On CbCr 2019 SMPTE Color Encoding System Blur on CbCr Advantageous for use in subsampling - 4:2:2-4:2: SMPTE SMPTE

19 Let s Take a Break for Questions 2019 SMPTE Enabling Global Education Measuring Color Differences Useful to use an objective metric for evaluating color differences. Tells you how visible differences will be due to errors in your system 2019 SMPTE SMPTE

20 Measuring Color Differences E2000 = SDR Imagery, set adapting white at 100cd/m 2 EITP = HDR Imagery, gives worst case error (no adaptation necessary) SMPTE Operational Practices (CMF) - Everyone has a slightly different CMF - Displays may look different to one person and not to another - Differences more extreme with narrow band primaries 2019 SMPTE SMPTE

21 Operational Practices Spectral Sensitivity in Cameras Cameras also have a Spectral Sensitivity Function. A camera may see an object differently than you SMPTE Operational Practices (Color Management) - Choose a single unified BT.709 RGB Convert to BT.2100 PQ working space - Convert all inputs to this common space BT.2100 PQ BT.2020 Yuv Convert to BT.2100 PQ Combine RGB 2019 SMPTE SMPTE

22 Operational Practices (Color Management) Be Careful to Compensate for: - Different color spaces - Different adapting white points (chromatic adaptation) - Different color encodings - Different EOTF functions 2019 SMPTE Questions? Jaclyn Pytlarz Senior Engineer Dolby Laboratories 2019 SMPTE SMPTE