0
# Quality Assessment
1

2
Comprehensive colour quality assessment functions for evaluating illumination source quality, colour fidelity, and spectral similarity. These functions implement industry-standard metrics including CRI, CQS, CFI, and advanced assessment methods for professional lighting evaluation and colour science applications.
3

4
## Capabilities
5

6
### Main Quality Assessment Functions
7

8
Primary entry points for colour quality calculations with automatic method selection and comprehensive result specifications.
9

10
```python { .api }
11
def colour_rendering_index(sd_test: SpectralDistribution, additional_data: bool = False) -> float | ColourRendering_Specification_CRI:
12
    """
13
    Calculate Colour Rendering Index (CRI) Ra of given test spectral distribution.
14
    
15
    Parameters:
16
    - sd_test: test illuminant spectral distribution
17
    - additional_data: whether to output detailed specification with individual sample data
18
    
19
    Returns:
20
    Float CRI Ra value (0-100) or detailed ColourRendering_Specification_CRI
21
    
22
    Notes:
23
    - Based on CIE 13.3-1995 standard using 8 test colour samples (TCS01-TCS08)
24
    - Uses CIE 1931 2° Standard Observer and CIE 1960 UCS colour space
25
    - Reference illuminant: blackbody radiator (CCT < 5000K) or CIE Illuminant D (CCT ≥ 5000K)
26
    - Individual sample indexes available with additional_data=True
27
    - Values: 100 = perfect colour rendering, lower values indicate poorer rendering
28
    """
29

30
def colour_quality_scale(sd_test: SpectralDistribution, additional_data: bool = False, method: str = "NIST CQS 9.0") -> float | ColourRendering_Specification_CQS:
31
    """
32
    Calculate Colour Quality Scale (CQS) of given test spectral distribution.
33
    
34
    Parameters:
35
    - sd_test: test illuminant spectral distribution
36
    - additional_data: whether to output detailed specification
37
    - method: computation method:
38
        * "NIST CQS 9.0": latest CQS method (default)
39
        * "NIST CQS 7.4": legacy CQS method with preference metrics
40
    
41
    Returns:
42
    Float CQS Qa value (0-100) or detailed ColourRendering_Specification_CQS
43
    
44
    Notes:
45
    - Uses 15 virtual test samples (VS) covering wider gamut than CRI
46
    - Employs CIE L*a*b* colour space with chromatic adaptation
47
    - Includes colour fidelity (Qf), preference (Qp), and gamut (Qg) metrics
48
    - More comprehensive than CRI, especially for LED and novel light sources
49
    - Values: 100 = excellent quality, 0 = poor quality
50
    """
51

52
def colour_fidelity_index(sd_test: SpectralDistribution, additional_data: bool = False, method: str = "CIE 2017") -> float | ColourRendering_Specification_CIE2017 | ColourQuality_Specification_ANSIIESTM3018:
53
    """
54
    Calculate Colour Fidelity Index (CFI) Rf of given test spectral distribution.
55
    
56
    Parameters:
57
    - sd_test: test illuminant spectral distribution
58
    - additional_data: whether to output detailed specification
59
    - method: computation method:
60
        * "CIE 2017": CIE 2017 Colour Fidelity Index (default)
61
        * "ANSI/IES TM-30-18": IES TM-30-18 method with gamut index
62
    
63
    Returns:
64
    Float CFI Rf value (0-100) or detailed specification object
65
    
66
    Notes:
67
    - Uses 99 test colour samples for comprehensive evaluation
68
    - Based on CAM02-UCS uniform colour space for improved accuracy
69
    - CIE 1964 10° Standard Observer for better correlation with visual assessment
70
    - TM-30-18 method includes gamut index Rg and local metrics
71
    - Superior to CRI for modern light sources and critical applications
72
    """
73

74
def spectral_similarity_index(sd_test: SpectralDistribution, sd_reference: SpectralDistribution, round_result: bool = True) -> NDArray:
75
    """
76
    Calculate Academy Spectral Similarity Index (SSI) between test and reference spectra.
77
    
78
    Parameters:
79
    - sd_test: test spectral distribution
80
    - sd_reference: reference spectral distribution  
81
    - round_result: whether to round result to integer (default True)
82
    
83
    Returns:
84
    SSI value (0-100), where 100 indicates perfect spectral match
85
    
86
    Notes:
87
    - Developed by Academy of Motion Picture Arts and Sciences
88
    - Evaluates spectral similarity rather than colour rendering
89
    - Uses 375-675nm range with 1nm resolution and specialized weighting
90
    - Ideal for comparing LED sources to reference illuminants
91
    - Does not require colour space transformations
92
    """
93

94
# Available methods for each quality metric
95
COLOUR_FIDELITY_INDEX_METHODS = {
96
    "CIE 2017": colour_fidelity_index_CIE2017,
97
    "ANSI/IES TM-30-18": colour_fidelity_index_ANSIIESTM3018
98
}
99

100
COLOUR_QUALITY_SCALE_METHODS = ("NIST CQS 7.4", "NIST CQS 9.0")
101
```
102

103
### CIE Colour Rendering Index (CRI) Functions
104

105
Traditional colour rendering evaluation based on CIE 13.3-1995 standard.
106

107
```python { .api }
108
def colour_rendering_index(sd_test: SpectralDistribution, additional_data: bool = False) -> float | ColourRendering_Specification_CRI:
109
    """
110
    Calculate CRI Ra using 8 standard test colour samples.
111
    
112
    Algorithm:
113
    1. Determine CCT of test source using Robertson's method
114
    2. Select reference illuminant (blackbody < 5000K, CIE D-series ≥ 5000K)
115
    3. Calculate tristimulus values for test samples under both illuminants
116
    4. Apply chromatic adaptation using CIE 1960 UCS
117
    5. Compute individual rendering indexes Ri for each sample
118
    6. Calculate general rendering index Ra as average of R1-R8
119
    
120
    Test Colour Samples:
121
    - TCS01: Light greyish red
122
    - TCS02: Dark greyish yellow  
123
    - TCS03: Strong yellow green
124
    - TCS04: Moderate yellowish green
125
    - TCS05: Light bluish green
126
    - TCS06: Light blue
127
    - TCS07: Light violet
128
    - TCS08: Light reddish purple
129
    - TCS09-TCS14: Special samples (not included in Ra)
130
    """
131

132
def tcs_colorimetry_data(sd_t: SpectralDistribution, sd_r: SpectralDistribution, sds_tcs: dict, cmfs: MultiSpectralDistributions, chromatic_adaptation: bool = False) -> tuple:
133
    """
134
    Calculate colorimetry data for test colour samples under test and reference illuminants.
135
    
136
    Parameters:
137
    - sd_t: test spectral distribution
138
    - sd_r: reference spectral distribution
139
    - sds_tcs: test colour sample spectral reflectance distributions
140
    - cmfs: colour matching functions
141
    - chromatic_adaptation: apply von Kries chromatic adaptation
142
    
143
    Returns:
144
    Tuple of DataColorimetry_TCS objects containing XYZ, uv, UVW data
145
    """
146

147
def colour_rendering_indexes(test_data: tuple, reference_data: tuple) -> dict:
148
    """
149
    Calculate individual colour rendering indexes for each test sample.
150
    
151
    Parameters:
152
    - test_data: colorimetry data under test illuminant
153
    - reference_data: colorimetry data under reference illuminant
154
    
155
    Returns:
156
    Dictionary mapping sample number to DataColourQualityScale_TCS
157
    
158
    Formula: Ri = 100 - 4.6 * ΔE*uv
159
    """
160
```
161

162
### Colour Quality Scale (CQS) Functions
163

164
Advanced quality assessment using virtual test samples and improved colour space.
165

166
```python { .api }
167
def colour_quality_scale(sd_test: SpectralDistribution, additional_data: bool = False, method: str = "NIST CQS 9.0") -> float | ColourRendering_Specification_CQS:
168
    """
169
    Calculate CQS using 15 virtual samples with comprehensive quality metrics.
170
    
171
    Quality Metrics:
172
    - Qa: Overall colour quality scale
173
    - Qf: Colour fidelity scale (similar to CRI)
174
    - Qp: Colour preference scale (rewards increased chroma)
175
    - Qg: Gamut area scale (relative gamut size)
176
    - Qd: Relative gamut area scale (deprecated in CQS 9.0)
177
    
178
    Virtual Samples (VS):
179
    - 15 synthetic reflectance spectra covering extended colour gamut
180
    - More saturated colours than CRI test samples
181
    - Better representation of real-world object colours
182
    """
183

184
def gamut_area(Lab: ArrayLike) -> float:
185
    """
186
    Calculate gamut area covered by given L*a*b* coordinates using triangulation.
187
    
188
    Parameters:
189
    - Lab: L*a*b* colourspace matrices for samples
190
    
191
    Returns:
192
    Gamut area G using Heron's formula for triangular areas
193
    """
194

195
def vs_colorimetry_data(sd_test: SpectralDistribution, sd_reference: SpectralDistribution, sds_vs: dict, cmfs: MultiSpectralDistributions, chromatic_adaptation: bool = False) -> tuple:
196
    """
197
    Calculate colorimetry data for virtual samples under test and reference illuminants.
198
    
199
    Uses von Kries chromatic adaptation with CMCCAT2000 transform
200
    when chromatic_adaptation=True for improved accuracy.
201
    """
202

203
def CCT_factor(reference_data: tuple, XYZ_r: ArrayLike) -> float:
204
    """
205
    Calculate correlated colour temperature factor penalizing extremely low CCT lamps.
206
    
207
    Returns factor ≤ 1.0 that reduces quality scores for sources with
208
    very low CCT that compress the reference gamut.
209
    """
210

211
def scale_conversion(D_E_ab: float, CCT_f: float, scaling_f: float) -> float:
212
    """
213
    Convert ΔE*ab value to CQS scale using logarithmic transformation.
214
    
215
    Formula: Qa = 10 * log(1 + exp((100 - scaling_f * ΔE*ab) / 10)) * CCT_f
216
    """
217

218
def delta_E_RMS(CQS_data: dict, attribute: str) -> float:
219
    """
220
    Calculate root-mean-square average for CQS data attributes.
221
    
222
    Used to compute RMS values for ΔE*ab, ΔE'*ab, and ΔC*ab.
223
    """
224

225
def colour_quality_scales(test_data: tuple, reference_data: tuple, scaling_f: float, CCT_f: float) -> dict:
226
    """
227
    Calculate individual quality scales for each virtual sample.
228
    
229
    Computes chroma shift ΔC*ab, colour difference ΔE*ab,
230
    and colour preference difference ΔE'*ab for each sample.
231
    """
232

233
# CQS reference gamut area constant
234
GAMUT_AREA_D65 = 8210  # Gamut area for CIE Illuminant D65
235
```
236

237
### CIE 2017 Colour Fidelity Index Functions
238

239
State-of-the-art colour fidelity assessment using 99 test samples and CAM02-UCS colour space.
240

241
```python { .api }
242
def colour_fidelity_index_CIE2017(sd_test: SpectralDistribution, additional_data: bool = False) -> float | ColourRendering_Specification_CIE2017:
243
    """
244
    Calculate CIE 2017 Colour Fidelity Index using 99 test colour samples.
245
    
246
    Technical Requirements:
247
    - Spectral range: 380-780nm (recommended)
248
    - Maximum interval: 5nm
249
    - Observer: CIE 1964 10° Standard Observer
250
    - Colour space: CAM02-UCS for uniform colour differences
251
    - Missing values filled with zeros per standard
252
    
253
    Test Colour Samples:
254
    - 99 samples covering comprehensive colour space
255
    - Real reflectance spectra from natural and artificial materials
256
    - Superior coverage compared to CRI's 8 or 14 samples
257
    """
258

259
def load_TCS_CIE2017(shape: SpectralShape) -> MultiSpectralDistributions:
260
    """
261
    Load CIE 2017 test colour samples dataset for given spectral shape.
262
    
263
    Parameters:
264
    - shape: spectral shape of tested illuminant
265
    
266
    Returns:
267
    MultiSpectralDistributions containing 99 test samples
268
    
269
    Notes:
270
    - Supports 1nm and 5nm intervals
271
    - Cached for performance on subsequent calls
272
    - Labels: TCS0-TCS98 (CIE 2017)
273
    """
274

275
def CCT_reference_illuminant(sd: SpectralDistribution) -> NDArray:
276
    """
277
    Determine reference illuminant CCT and Δuv using Ohno 2013 method.
278
    
279
    Uses optimized temperature range 1000K-25000K for performance.
280
    Returns CCT and distance from Planckian locus.
281
    """
282

283
def sd_reference_illuminant(CCT: float, shape: SpectralShape) -> SpectralDistribution:
284
    """
285
    Generate reference illuminant for given CCT following CIE 2017 standard.
286
    
287
    Reference Selection:
288
    - CCT < 4000K: Planckian radiator only
289
    - 4000K ≤ CCT ≤ 5000K: Weighted mixture of Planckian and CIE D-series
290
    - CCT > 5000K: CIE D-series illuminant only
291
    
292
    Mixture ensures smooth transition between reference types.
293
    """
294

295
def tcs_colorimetry_data(sd_irradiance: SpectralDistribution | list, sds_tcs: MultiSpectralDistributions, cmfs: MultiSpectralDistributions) -> tuple:
296
    """
297
    Calculate test colour sample colorimetry under CIE 2017 viewing conditions.
298
    
299
    Viewing Conditions:
300
    - Adapting luminance LA: 100 cd/m²
301
    - Background luminance Yb: 20% of white
302
    - Surround: Average (CIECAM02)
303
    - Discount illuminant: True
304
    """
305

306
def delta_E_to_R_f(delta_E: ArrayLike) -> NDArray:
307
    """
308
    Convert CAM02-UCS colour difference to CIE 2017 fidelity index.
309
    
310
    Formula: Rf = 10 * log(1 + exp((100 - 6.73 * ΔE) / 10))
311
    
312
    Where ΔE is Euclidean distance in CAM02-UCS colour space.
313
    """
314

315
# CIE 2017 constants
316
SPECTRAL_SHAPE_CIE2017 = SpectralShape(380, 780, 1)  # Standard spectral shape
317
ROOT_RESOURCES_CIE2017 = "datasets/"  # TCS data directory
318
```
319

320
### ANSI/IES TM-30-18 Functions
321

322
Extended colour fidelity assessment with gamut index and local colour fidelity metrics.
323

324
```python { .api }
325
def colour_fidelity_index_ANSIIESTM3018(sd_test: SpectralDistribution, additional_data: bool = False) -> float | ColourQuality_Specification_ANSIIESTM3018:
326
    """
327
    Calculate ANSI/IES TM-30-18 Colour Fidelity Index with comprehensive metrics.
328
    
329
    Based on CIE 2017 CFI with additional TM-30-18 specific metrics:
330
    - Rf: Colour fidelity index (same as CIE 2017)
331
    - Rg: Gamut index (relative gamut area)
332
    - Rfs: 16 local colour fidelity indexes per hue bin
333
    - Rcs: 16 local chromaticity shifts (%)  
334
    - Rhs: 16 local hue shifts
335
    
336
    Hue Binning:
337
    - 16 equal hue bins of 22.5° each in CAM02-UCS a'b' plane
338
    - Bin assignment based on reference sample hue angles
339
    - Local metrics provide detailed spectral performance analysis
340
    """
341

342
def averages_area(averages: ArrayLike) -> float:
343
    """
344
    Calculate area of polygon formed by hue bin averages using shoelace formula.
345
    
346
    Used for computing gamut index Rg = 100 * (test_area / reference_area).
347
    """
348

349
# TM-30-18 hue bin configuration
350
# 16 bins × 22.5° = 360° coverage
351
# Bin centers: 11.25°, 33.75°, ..., 348.75°
352
```
353

354
### Academy Spectral Similarity Index Functions
355

356
Direct spectral comparison without colour space transformations.
357

358
```python { .api }
359
def spectral_similarity_index(sd_test: SpectralDistribution, sd_reference: SpectralDistribution, round_result: bool = True) -> NDArray:
360
    """
361
    Calculate Academy Spectral Similarity Index for direct spectral comparison.
362
    
363
    Algorithm:
364
    1. Normalize spectra to 375-675nm at 1nm resolution
365
    2. Integrate using weighted trapezoidal rule over 10nm bands
366
    3. Calculate relative differences with reference offset
367
    4. Apply wavelength-dependent weighting (higher weight 500-600nm)
368
    5. Convolve with smoothing kernel [0.22, 0.56, 0.22]
369
    6. Compute SSI = 100 - 32 * √(Σ(weighted_differences²))
370
    
371
    Spectral Requirements:
372
    - Range: 375-675nm (extrapolated with zeros outside)
373
    - Resolution: 1nm (interpolated from input)
374
    - Normalization: area under curve = 1.0
375
    
376
    Weighting Function:
377
    - 375-425nm: 4/15 to 22/45 (low weight, blue)
378
    - 425-575nm: 32/45 to 1.0 (high weight, visible)
379
    - 575-625nm: 1.0 (maximum weight, yellow-green)
380
    - 625-675nm: 11/15 to 3/15 (decreasing weight, red)
381
    
382
    Applications:
383
    - LED source matching to reference illuminants
384
    - Spectral metameric evaluation
385
    - Light source development and quality control
386
    """
387

388
# SSI constants
389
SPECTRAL_SHAPE_SSI = SpectralShape(375, 675, 1)  # Required spectral shape
390
```
391

392
## Specification Classes
393

394
Comprehensive data classes for structured quality assessment results.
395

396
```python { .api }
397
class ColourRendering_Specification_CRI:
398
    """
399
    CRI computation results with detailed sample-by-sample data.
400
    
401
    Attributes:
402
    - name: test illuminant name
403
    - Q_a: general colour rendering index Ra (0-100)
404
    - Q_as: individual sample rendering indexes R1-R14
405
    - colorimetry_data: test and reference colorimetry computations
406
    
407
    Individual Sample Data (DataColourQualityScale_TCS):
408
    - name: sample name (e.g., "TCS01")
409
    - Q_a: individual rendering index Ri
410
    """
411

412
class ColourRendering_Specification_CQS:
413
    """
414
    CQS computation results with comprehensive quality metrics.
415
    
416
    Attributes:
417
    - name: test illuminant name
418
    - Q_a: overall colour quality scale (0-100)
419
    - Q_f: colour fidelity scale (0-100)
420
    - Q_p: colour preference scale (0-100, None in CQS 9.0)
421
    - Q_g: gamut area scale (relative to D65)
422
    - Q_d: relative gamut area scale (None in CQS 9.0)
423
    - Q_as: individual sample quality data
424
    - colorimetry_data: detailed colorimetric computations
425
    
426
    Sample Data (DataColourQualityScale_VS):
427
    - name: virtual sample name
428
    - Q_a: sample quality scale
429
    - D_C_ab: chroma difference ΔC*ab
430
    - D_E_ab: colour difference ΔE*ab
431
    - D_Ep_ab: preference-weighted colour difference
432
    """
433

434
class ColourRendering_Specification_CIE2017:
435
    """
436
    CIE 2017 CFI results with advanced colorimetric data.
437
    
438
    Attributes:
439
    - name: test illuminant name
440
    - sd_reference: reference illuminant spectrum
441
    - R_f: colour fidelity index (0-100)
442
    - R_s: individual sample fidelity indexes (99 values)
443
    - CCT: correlated colour temperature (K)
444
    - D_uv: distance from Planckian locus
445
    - colorimetry_data: CAM02-UCS based computations
446
    - delta_E_s: colour shifts in CAM02-UCS space
447
    
448
    Colorimetry Data (DataColorimetry_TCS_CIE2017):
449
    - name: sample names (99 TCS labels)
450
    - XYZ: CIE XYZ tristimulus values
451
    - CAM: CIECAM02 appearance specification
452
    - JMh: lightness-chroma-hue coordinates
453
    - Jpapbp: CAM02-UCS uniform colour coordinates
454
    """
455

456
class ColourQuality_Specification_ANSIIESTM3018:
457
    """
458
    TM-30-18 extended colour quality specification with local metrics.
459
    
460
    Extends CIE 2017 specification with:
461
    - R_g: gamut index (relative gamut area)
462
    - bins: hue bin assignments for 99 samples
463
    - averages_test: test hue bin average coordinates
464
    - averages_reference: reference hue bin averages  
465
    - average_norms: reference coordinate magnitudes
466
    - R_fs: local fidelity indexes (16 hue bins)
467
    - R_cs: local chromaticity shifts (16 values, %)
468
    - R_hs: local hue shifts (16 values)
469
    
470
    Local Metrics Applications:
471
    - Identify specific hue regions with poor performance
472
    - Evaluate gamut compression/expansion patterns
473
    - Optimize LED spectra for targeted applications
474
    """
475

476
# Supporting data classes
477
class DataColorimetry_TCS:
478
    """CRI test colour sample colorimetry data."""
479
    # name: sample name, XYZ: tristimulus, uv: chromaticity, UVW: uniform coordinates
480

481
class DataColourQualityScale_TCS:
482
    """CRI sample quality scale data."""  
483
    # name: sample name, Q_a: rendering index
484

485
class DataColorimetry_VS:
486
    """CQS virtual sample colorimetry data."""
487
    # name: sample name, XYZ: tristimulus, Lab: L*a*b*, C: chroma
488

489
class DataColourQualityScale_VS:
490
    """CQS sample quality scale data."""
491
    # name: sample name, Q_a: quality, D_C_ab: chroma shift, D_E_ab: colour shift, D_Ep_ab: preference shift
492

493
class DataColorimetry_TCS_CIE2017:
494
    """CIE 2017 test sample colorimetry data."""
495
    # name: sample names, XYZ: tristimulus, CAM: appearance, JMh: appearance coordinates, Jpapbp: uniform coordinates
496
```
497

498
## Constants and Datasets
499

500
Essential constants and datasets for quality assessment computations.
501

502
```python { .api }
503
# Test colour sample datasets
504
SDS_TCS = {
505
    # CRI test colour samples (TCS01-TCS14)
506
    "TCS01": SpectralDistribution,  # Light greyish red
507
    "TCS02": SpectralDistribution,  # Dark greyish yellow
508
    "TCS03": SpectralDistribution,  # Strong yellow green
509
    "TCS04": SpectralDistribution,  # Moderate yellowish green
510
    "TCS05": SpectralDistribution,  # Light bluish green
511
    "TCS06": SpectralDistribution,  # Light blue
512
    "TCS07": SpectralDistribution,  # Light violet
513
    "TCS08": SpectralDistribution,  # Light reddish purple
514
    "TCS09": SpectralDistribution,  # Strong red
515
    "TCS10": SpectralDistribution,  # Strong yellow
516
    "TCS11": SpectralDistribution,  # Strong green
517
    "TCS12": SpectralDistribution,  # Strong blue
518
    "TCS13": SpectralDistribution,  # Light yellowish pink (Caucasian skin)
519
    "TCS14": SpectralDistribution   # Moderate olive green (leaf)
520
}
521

522
SDS_VS = {
523
    "NIST CQS 7.4": {
524
        # 15 virtual samples for CQS 7.4
525
    },
526
    "NIST CQS 9.0": {
527
        # 15 virtual samples for CQS 9.0  
528
    }
529
}
530

531
# Sample index mappings
532
INDEXES_TO_NAMES_TCS = {
533
    1: "TCS01", 2: "TCS02", 3: "TCS03", 4: "TCS04",
534
    5: "TCS05", 6: "TCS06", 7: "TCS07", 8: "TCS08", 
535
    9: "TCS09", 10: "TCS10", 11: "TCS11", 12: "TCS12",
536
    13: "TCS13", 14: "TCS14"
537
}
538

539
INDEXES_TO_NAMES_VS = {
540
    # Virtual sample index to name mapping
541
}
542

543
# Reference gamut area
544
GAMUT_AREA_D65 = 8210  # CIE Illuminant D65 gamut area for CQS normalization
545

546
# Spectral shapes
547
SPECTRAL_SHAPE_CIE2017 = SpectralShape(380, 780, 1)  # CIE 2017 standard shape
548
SPECTRAL_SHAPE_SSI = SpectralShape(375, 675, 1)      # SSI computation shape
549
```
550

551
## Usage Examples
552

553
Practical examples demonstrating quality assessment workflows for different applications.
554

555
### Basic Quality Assessment
556

557
```python
558
from colour import SDS_ILLUMINANTS, SpectralDistribution
559
from colour.quality import (
560
    colour_rendering_index, colour_quality_scale, 
561
    colour_fidelity_index, spectral_similarity_index
562
)
563

564
# Load test illuminant
565
sd_test = SDS_ILLUMINANTS["FL2"]  # Fluorescent F2
566

567
# Basic quality metrics
568
cri = colour_rendering_index(sd_test)
569
print(f"CRI Ra: {cri:.1f}")  # CRI Ra: 64.2
570

571
cqs = colour_quality_scale(sd_test)  
572
print(f"CQS Qa: {cqs:.1f}")  # CQS Qa: 64.1
573

574
cfi = colour_fidelity_index(sd_test)
575
print(f"CFI Rf: {cfi:.1f}")  # CFI Rf: 70.1
576

577
# Spectral similarity to daylight
578
sd_d65 = SDS_ILLUMINANTS["D65"]
579
ssi = spectral_similarity_index(sd_test, sd_d65)
580
print(f"SSI: {ssi:.0f}")  # SSI: 75
581
```
582

583
### Detailed Assessment with Specifications
584

585
```python
586
# Get detailed CRI specification with individual sample data
587
cri_spec = colour_rendering_index(sd_test, additional_data=True)
588
print(f"General CRI: {cri_spec.Q_a:.1f}")
589
for i, (sample_num, sample_data) in enumerate(cri_spec.Q_as.items(), 1):
590
    if i <= 8:  # Show first 8 samples used in Ra
591
        print(f"  R{sample_num} ({sample_data.name}): {sample_data.Q_a:.1f}")
592

593
# Detailed CQS with all quality metrics
594
cqs_spec = colour_quality_scale(sd_test, additional_data=True, method="NIST CQS 9.0")
595
print(f"CQS Quality Metrics:")
596
print(f"  Qa (Overall): {cqs_spec.Q_a:.1f}")  
597
print(f"  Qf (Fidelity): {cqs_spec.Q_f:.1f}")
598
print(f"  Qg (Gamut): {cqs_spec.Q_g:.1f}")
599
# Qp and Qd are None in CQS 9.0
600

601
# TM-30-18 with local hue metrics
602
tm30_spec = colour_fidelity_index(sd_test, additional_data=True, method="ANSI/IES TM-30-18")
603
print(f"TM-30-18 Metrics:")
604
print(f"  Rf (Fidelity): {tm30_spec.R_f:.1f}")
605
print(f"  Rg (Gamut): {tm30_spec.R_g:.1f}")
606
print(f"Local Fidelity by Hue Bin:")
607
for i, rf_local in enumerate(tm30_spec.R_fs):
608
    hue_center = 22.5 * i + 11.25  
609
    print(f"  Bin {i+1} ({hue_center:.1f}°): Rf={rf_local:.1f}")
610
```
611

612
### LED Source Development
613

614
```python
615
# Evaluate custom LED spectrum
616
led_data = {
617
    380: 0.001, 385: 0.002, 390: 0.003, 395: 0.005,
618
    # ... LED spectral data ...
619
    450: 0.850, 455: 0.900, 460: 0.950,  # Blue peak
620
    # ...
621
    550: 0.120, 555: 0.150, 560: 0.180,  # Green component
622
    # ...
623
    630: 0.280, 635: 0.300, 640: 0.320   # Red peak
624
}
625
sd_led = SpectralDistribution(led_data, name="Custom LED")
626

627
# Comprehensive quality assessment
628
print("LED Quality Assessment:")
629
print(f"CRI Ra: {colour_rendering_index(sd_led):.1f}")
630
print(f"CQS Qa: {colour_quality_scale(sd_led):.1f}")  
631
print(f"CFI Rf: {colour_fidelity_index(sd_led):.1f}")
632

633
# Compare spectral match to target illuminant
634
target = SDS_ILLUMINANTS["D65"]
635
ssi_match = spectral_similarity_index(sd_led, target)
636
print(f"SSI to D65: {ssi_match:.0f}")
637

638
# TM-30-18 for detailed analysis
639
tm30 = colour_fidelity_index(sd_led, additional_data=True, method="ANSI/IES TM-30-18")
640
print(f"Gamut Index Rg: {tm30.R_g:.1f}")
641

642
# Identify weak hue regions
643
weak_hues = [(i, rf) for i, rf in enumerate(tm30.R_fs) if rf < 80]
644
if weak_hues:
645
    print("Weak Performance Hue Regions:")
646
    for bin_idx, rf_val in weak_hues:
647
        hue_deg = 22.5 * bin_idx + 11.25
648
        print(f"  {hue_deg:.0f}° (Rf={rf_val:.1f})")
649
```
650

651
### Method Comparison Study
652

653
```python
654
# Compare all methods for multiple sources
655
sources = ["A", "D65", "FL2", "FL11"]  # Incandescent, daylight, fluorescent
656
methods_cfi = ["CIE 2017", "ANSI/IES TM-30-18"] 
657
methods_cqs = ["NIST CQS 7.4", "NIST CQS 9.0"]
658

659
print("Quality Assessment Comparison:")
660
print("Source    | CRI Ra | CQS 7.4| CQS 9.0| CFI 17 | TM30-18| SSI->D65")
661
print("-" * 70)
662

663
for source in sources:
664
    sd = SDS_ILLUMINANTS[source]
665
    
666
    # Calculate all metrics
667
    cri = colour_rendering_index(sd)
668
    cqs_74 = colour_quality_scale(sd, method="NIST CQS 7.4") 
669
    cqs_90 = colour_quality_scale(sd, method="NIST CQS 9.0")
670
    cfi_17 = colour_fidelity_index(sd, method="CIE 2017")
671
    cfi_tm30 = colour_fidelity_index(sd, method="ANSI/IES TM-30-18")
672
    ssi = spectral_similarity_index(sd, SDS_ILLUMINANTS["D65"])
673
    
674
    print(f"{source:9} | {cri:6.1f} | {cqs_74:6.1f} | {cqs_90:6.1f} | "
675
          f"{cfi_17:6.1f} | {cfi_tm30:7.1f} | {ssi:7.0f}")
676
```
677

678
## Technical Notes
679

680
### Computational Considerations
681

682
- **CRI**: Fastest computation, suitable for real-time applications
683
- **CQS**: Moderate complexity, good for general LED evaluation  
684
- **CFI**: Most computationally intensive due to 99 samples and CAM02-UCS
685
- **SSI**: Fast spectral-only computation, no colour space transforms
686

687
### Method Selection Guidelines
688

689
- **CRI**: Legacy applications, regulatory compliance, basic evaluation
690
- **CQS**: LED development, broader gamut assessment than CRI
691
- **CIE 2017 CFI**: Research applications, most accurate colour fidelity
692
- **TM-30-18**: Comprehensive lighting design, local performance analysis  
693
- **SSI**: Spectral matching, metameric source development
694

695
### Spectral Requirements
696

697
- **Range**: 380-780nm recommended (CRI/CQS/CFI), 375-675nm (SSI)
698
- **Resolution**: ≤5nm intervals for CFI, 1nm preferred for SSI
699
- **Extrapolation**: Zero-padding for missing spectral regions
700
- **Normalization**: Methods handle relative spectral data appropriately
701

702
### Accuracy and Limitations
703

704
- All methods assume 2° or 10° standard observer viewing conditions
705
- CRI limited by 8 test samples and CIE 1960 UCS colour space
706
- CQS virtual samples better represent real object colours than CRI
707
- CFI most correlates with visual assessments due to CAM02-UCS uniformity
708
- SSI evaluates spectral similarity, not colour rendering quality
709

710
The quality assessment module provides comprehensive tools for professional lighting evaluation, supporting both traditional CRI assessment and advanced methods for modern light sources including LEDs, OLEDs, and laser-based systems.