Introduction to Color Management
for Film and TV
Igor Riđanović
Why color manage in food industry?
Why color manage in retail?
common
source
multiple
display
devices
Why color manage in media?
correct color interpretation
incorrect color interpretation
Why color manage in post?
Why color manage in post?
What is color management?

Perceptually consistent color across a range of devices

Consistent color within a pipeline

Interoperability
Why is color management important in film and TV?
Brand identity
 Protects integrity of creative intent


Maintains suspension of disbelief
right
wrong
The nature of light and color
The nature of light and color
Objects and living beings have no color
 Light energy absorption, reflection and transmission

3D model Copyright © 2003-2013 Andrew Kator & Jennifer Legaz
How do we see light and color?
Rods see low light and no color
 Three type of cones can see color


Near logarithmic response to luminance
HansWerner
Hunzik
er
How do we see light and color?
S, M and L cones
How do we see light and color?
perceived
brightness
perceived
brightness
linear light intensity
Wolphram|Alpha
doubling light intensity
Additive and subtractive color mixing
Pko
Cantus
Color model and color space
Color model—an abstract primary color mixing principle
 Color space—specific values for the primaries


Color gamut—unambiguous color set defined by the primaries
CIE XYZ(1931) color space
CIE xy
chromaticity diagram
sRGB color space
PAR
Color space classification
Device independent (example: CIE XYZ)
 Device dependent (example: Rec. 709)

Device dependent color
Examples:
 Traditional CMYK press

Analog and Digital TV
Device dependent color—numerical representation
In RGB color model each pixel is represented by
the additive mix of its three channels (R + B + G)
R 205
G 012
B 005
R 000
G 000
B 000
R 255
G 255
B 255
Device dependent color—numerical representation
Bit depth determines color precision
1 bit: fax
8 bit: grayscale
8 bit per channel:
truecolor
Device dependent color—numerical representation
A bit is a binary numeral representing either “0” or “1”
1 bit: 2 levels
2 bit: 4 levels
10 bit: 1024 levels
Device dependent color—numerical representation
Values are intrinsically tied to the device
R 255
G 255
B 000
R 255
G 255
B 000
sRGB color space
(computer monitor)
mystery color space
(some handheld device)
Device dependent color
R 205
G 012
B 005
R 205
G 012
B 005
device 1
source device
R 205
G 012
B 005
device 2
Device dependent color
R 205
G 012
B 005
source device
R?
G?
B?
device 2
Device dependent color
CMS
R 205
G 012
B 005
source device
R 212
G 008
B 010
device 2
Device dependent color
Color management:
Typically translates color values from one device dependent color
space to another device dependent color space with respect to
gamut, gamma and dynamic range
Device dependent color—gamut
Rec. 2020 vs.
Rec. 709 color gamut
GrandDrake
Device dependent color—gamut
Rec. 709 vs.
DCI-P3
3D color gamut
Device dependent color—gamma
linear gamma
2.2 gamma
Wolfram|Alpha
Device dependent color—dynamic range
higher dynamic range
lower dynamic range
original
gamma mismatch
gamut mismatch
reduced dynamic range
Case Study: Black and White TV (simplified)
subject
camera
video
TV set
OTA broadcast
Traditional TV color management
Single input, single output
TV camera
editorial
TV monitor
γ = 2.5
TV camera
γ = 0.45
Contemporary color management requirements
Alexa
5D
Vimeo
editorial
HD monitor
graphics
DVD
film scan
QT MOV
...
VFX
...
delivery
Color management system
Color management system
Rendering intent defines how values are translated
from one color gamut to another:

Perceptual

Saturation

Relative colorimetric

Absolute colorimetric
Color management system
Calibration and profiling:

Calibration brings device to factory recommended specification

Profiling builds a fingerprint of the specific device's differential from a
standard color space specification
Good news!
Color management in most NLEs is much simpler.
Color management can be more sophisticated in picture finishing and
color correction systems.
Sometimes aided by the operating system
 Can be dependent on user selected LUTs or matrices

Transforming the numbers
input values
1D LUT
 3D LUT


3X3 matrix
0
1
2
3
4
5
output values
R
G
B
0.034439 0.034439 0.034439
0.034712 0.034712 0.034712
0.034988 0.034988 0.034988
0.035266 0.035266 0.035266
0.035546 0.035546 0.035546
0.035828 0.035828 0.035828
...
If input R or G or B = 0
Output R or G or B = 0.034439
Transforming the numbers
1D LUT
 3D LUT


3X3 matrix
Rec. 709 vs.
DCI-P3
3D color gamut
Transforming the numbers
1D LUT
 3D LUT


3X3 matrix
Mathematical transformation.
Not as precise as 3D LUT but adequate
for many applications.
How do popular programs color manage?
Media Composer
 Creative Cloud


FCP X
How do popular programs color manage?
tape 709
files
601/709
MC 6.x
Rec. 709
TV monitor
Rec. 709
sRGB
files
user choices
How do popular programs color manage?
0
0
255
16
0 mV
superblack
235 255
700 mV
superwhite
How do popular programs color manage?
without dithering
with dithering
Recent developments
CDL
 Scene referred linear


ACES
Special Thanks
James Mathers, Digital Cinema Society
Matt Feury, Avid Technologies
Further reading
Fraser, Murphy and Bunting (2005). Real World Color Management.
Berkeley: Peachpit Press.
Graham, S. (2002). The Science of Imaging: An Introduction. Bristol: IOP Publishing.
Bloch, C. (2007). The HDRI Handbook. Santa Barbara: Rocky Nook.
Kennel, G. (2007). Color and Mastering for Digital Cinema. Burlington: Focal Press.
www.poynton.com
www.lightillusion.com
This presentation is available for download at HDhead.com.