light spectrum and color...
TRANSCRIPT
1
Yoshi Ohno1 and Cameron Miller2
National Institute of Standards and Technology
Gaithersburg, Maryland USA
1 NIST Fellow, IES Fellow, CIE VP Technical 2 NIST Photometry Project Leader, IES/TPC Chair
Light Spectrum and Color Quality
Department of Energy, Solid State Lighting R&D Workshop Jan. 29-31, 2013, Long Beach, CA
3
Light Color vs. Object Color
Chromaticity diagrams such as (x,y), (u’,v’) are two-dimensional and are only for light color. These are not for object color.
No black, grey, or brown
Object color needs another axis: black—white
so, 3-dimensional color space.
}
4
Object Color Space
n Three attributes of object color are hue, chroma (saturation), and lightness, and are expressed in a three dimensional space.
n CIE standard object color spaces:
CIELAB CIELUV published in 1976. 3D color space
Ligh
tnes
s
Hue
Chroma
white
black
5
CIE 1976 (L*a* b*) color space (CIELAB color space)
: for object surface
: white reference (perfect diffuser)
Ref. CIE 15:2004
Color difference
a*
b*
L*=100
L*=0
L*
6
CIELAB (a*,b*) plots: Hue-Chroma plot
Ref. D65
1200 Munsell color samples 15 saturated Munsell color samples
7
Color Rendering Index (CRI)
Reference source Test source
Planckian (CCT<5000 K)
CIE Dxx
Standard Daylight (CCT > 5000 K)
Same CCT [K]
#1 #2 #3 #4 #5 #6 #7 #8
#9 #10 #11 #12 #13 #14
Ra
400 500 600 700
CIE 13.3
R9
8
R9 for red object rendering
Why R9 scores go negative? The red region in color space (W*U*V*) used in CRI is extremely stretched (by a factor of 3 or 4). Color difference (∆E) in red is calculated to be much larger than other colors and its score gets low: R9= 100 - 4.6 x ∆E9.
R9 : Good indication for rendering of red objects and skin tone (very important). Useful information to supplement Ra.
To convert R9 value to the similar scale as Ra (for 2700 to 3000 K), R9’=100 - (100 – R9)/4 So, R9>0 R9’>75 (acceptable) R9>50 R9’>88 (very good)
(2856 K)
(6500 K)
9
n 15 saturated test color samples n Update the old formulae in CRI - CIELAB color space - CMCCAT2000 chromatic adaptation - 0 to 100 scale - RMS averaging of color differences n Saturation factor (address the Hunt effect) n Discussed in CIE TC1-69. Next action ..
Fixes the problems of CRI for SSL sources Produce one number score that correlates well with visual
perception of color rendering for real objects
W. Davis and Y. Ohno, Color Quality Scale, Optical Engineering 49 (3), 033602 March 2010
Color Quality Scale (CQS)
10
Color Rendering Simulation (CQS 9.0 EXCEL sheet)
Y. Ohno, Spectral Design Considerations for Color Rendering of White LED Light Sources, Opt. Eng. 44, 111302 (2005).
11
Color Rendering
Trade-off
Optimize spectrum for intended applications.
Color quality and energy efficiency
Light spectrum Chromaticity, CCT Color Quality
Luminous Efficacy (lm/W)
Energy Efficiency
12
Low pressure sodium lamp
White light source
520 lm/W 300 ~ 400 lm/W ~ 250 lm/W LER (theoretical maximum)
Trade-off between efficacy and color rendering
Excellent color rendering
(CRI=100) color rendering
??
No color rendering (CRI= -47)
Daylight (400-700 nm)
There are some ways to enhance both color quality and energy efficiency
400 500 600 700
13
Luminous Efficacy of a Source [lm/W]
=
Luminous flux [lm] Electrical power [W]
Radiant Efficiency
X
Optical power [W] Electrical power [W]
(External Q.E.)
Luminous Efficacy of Radiation [lm/W]
Luminous flux [lm] Optical power [W]
Example: 200 lm/W 50 % 400 lm/W
Luminous Efficacy of a Source
Theoretical maximum lm/W of a given source.
Determined only by the spectrum of the source.
400 500 600 700
14
Luminous Efficacy of Radiation - LER (lm/W)
Theoretical lm/W for a given spectrum
<Examples of LER>
Tri-p FL (3300 K) ~350 lm/W
CW FL (4300 K) ~340 lm/W
MH (4300 K) ~300 lm/W
HPS (2100 K) ~380 lm/W
B-Y phosphor LED ~310 lm/W
683 lm/W
V(λ) S(λ)
15
What we can learn from fluorescent lamps
Cool White Tri-phosphor Cool White improved for color
Narrow-band may be the solution for both high efficacy and high color rendering.
16
Examples of real LED lamps (sold in USA)
Phosphor-type white LEDs are meeting current needs, but ….
303 lm/W 291 lm/W 305 lm/W 305 lm/W Ra=80~83 R9=10~30
318 lm/W 301 lm/W 320 lm/W 318 lm/W
17
Narrowband can be more efficient (available LEDs)
Narrow-band: + 20% higher theoretical lm/W, plus another + 20 % gain from no Stokes loss
305 lm/W 303 lm/W 357 lm/W 363 lm/W Ra=81 Ra=30 Ra=90 Ra=90
18
Simulation of RGB/RGBA (not limited to available LEDs)
378 lm/W 382 lm/W 364 lm/W 352 lm/W Ra=85 Ra=91 Ra=96 Qa=93 R9=90
21
NIST Spectrally Tunable Lighting Facility
• Real-size room (2.5 m x 2.5 m cubicle)
• 2 cubicles side by side. • 1800 high power LEDs in 22
channels of LED spectra. • > 500 lx illumination with many
simulated white light spectra
29
Current broadband phosphor white LEDs meet current urgent needs for energy saving, but color quality and theoretical efficacy of most of them are not great.
Narrowbands can be theoretically more energy efficient.
Narrowbands can create as good color rendering as smooth broadband sources
Narrowbands can make color-enhanced sources (preferred in many applications.)
Summary
Thank you for your attention. Contact: [email protected] [email protected]