David R Wyble Munsell Color Science Laboratory Rochester

David R. Wyble Munsell Color Science Laboratory Rochester Institute of Technology wyble@cis. rit. edu Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 Analysis of Spectrophotometer Specular Performance Using Goniophotometric Information

Introduction Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 • All integrating sphere spectrophotometers are not created equal • Standards allow a wide range of conforming devices • Sample gloss and specular port configuration can significantly affect measurements

Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 Reflectance factor A Pathological Case

Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 Reflectance factor A more expected example…

Purpose Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 • Determine a relationship between the size of the specular port and the effective performance of the spectrophotometer in SPEX mode • Create a method to correct measurements to allow comparison of results from instruments of differing configurations

Outline CIE 15. 2 Colorimetry Theory on Effective Specular Port Size Samples and Goniometric Measurements Results & Conclusions Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 • •

CIE 15. 2 Colorimetry c) Diffuse/normal (symbol d/0): The specimen is illuminated diffusely by an integrating sphere. The angle between the normal to the specimen and the axis of the viewing beam should not exceed 10°. The integrating sphere may be of any diameter provided the total area of the ports does not exceed 10% of the internal reflecting sphere area. The angle between the axis and any ray of the viewing beam should not exceed 5°. c) Diffuse/normal (symbol d/0): (Similar angular specifications) Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 1. 4 Illuminating and viewing conditions for reflecting specimens

Reflectance: d/8 SPEX Detector Incident light Baffle Sample 8° viewing, diffuse illumination, SPIN Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 Specular cap “black cap” excludes

Reflectance: 8/d SPEX Detector Baffle Sample Diffuse viewing, 8° illumination, SPEX Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 Incident light

CIE 15. 2 Colorimetry c) Diffuse/normal (symbol d/0): The specimen is illuminated diffusely by an integrating sphere. The angle between the normal to the specimen and the axis of the viewing beam should not exceed 10°. The integrating sphere may be of any diameter provided the total area of the ports does not exceed 10% of the internal reflecting sphere area. The angle between the axis and any ray of the viewing beam should not exceed 5°. Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 1. 4 Illuminating and viewing conditions for reflecting specimens

CIE 15. 2 Colorimetry Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002

CIE 15. 2 Colorimetry c) Diffuse/normal (symbol d/0): The specimen is illuminated diffusely by an integrating sphere. The angle between the normal to the specimen and the axis of the viewing beam should not exceed 10°. The integrating sphere may be of any diameter provided the total area of the ports does not exceed 10% of the internal reflecting sphere area. The angle between the axis and any ray of the viewing beam should not exceed 5°. Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 1. 4 Illuminating and viewing conditions for reflecting specimens

CIE 15. 2 Colorimetry Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002

CIE 15. 2 Colorimetry Note 1: For the conditions ‘diffuse/normal’ and ‘normal diffuse’ the regularly reflected component of specimens with mixed reflection may be excluded with a gloss trap. If a gloss trap is used, details of its size, shape, and position should be given, Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 1. 4 Illuminating and viewing conditions for reflecting specimens

CIE 15. 2 Colorimetry Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 By inference, the angle from the normal to the gloss trap will not exceed 10°.

CIE 15. 2 Colorimetry Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 No guidance on angular size of gloss trap.

CIE 15. 2 Colorimetry Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 • All we are told about the specular port is to report the configuration used • Still a range of configurations that meet the specification

Specular Port Size Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 As port size grows …

Specular Port Size Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 … more of the specular information …

Specular Port Size Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 … is lost …

Specular Port Size Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 … but how much?

Where do we go now? Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 We need to know the details of how our instruments handle this component of the reflected light.

Where do we go now? Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 To do this, we first need to accurately characterize a set of samples, by measuring their reflectance characteristics as a function of angle.

MCSL Goniometer Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 Sample holder Collimator IR filter Light source

MCSL Goniometer Technical description Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 • Light source • 100 cm sphere • 2 interior 19. 6 V GE bulbs • Lamp current monitored and manually maintained at 6. 00 A • IR filter • Collimation lens • Detector • Photoresearch PR 704 spectroradiometer • Aperture ~ 3° • Measurement units are integrated radiance

MCSL Goniometer Technical description Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 • Sample and detector angles independently adjustable within physical constraints • Vernier scales allow repeatable angle settings to 0. 5°

MCSL Goniometer Sample +q 0° -q Incident light Physical constraints limit measurement angles to -8° to +75°, always referenced to the specular angle. Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 Detector

Theory Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 • Measure spectral reflectance using SPIN and SPEX modes • Calculate the average difference between SPIN and SPEX data, in percent reflectance • Determine the effective size of the specular port that would account for the above difference

SPIN and SPEX Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002

Theory Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 • Measure spectral reflectance using SPIN and SPEX modes • Calculate the average difference between SPIN and SPEX data, in percent reflectance • Determine the effective size of the specular port that would account for the above difference

Average Spectral Difference Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002

Theory Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 • Measure spectral reflectance using SPIN and SPEX modes • Calculate the average difference between SPIN and SPEX data, in percent reflectance • Determine the effective size of the specular port that would account for the above difference

Effective Specular Port Calculation Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 • Measure radiance vs angle on goniometer • Calculate cumulative volume as a function of radius • Determine the radius that results in the percent cumulative volume that matches the SPIN-SPEX difference

Detection angle Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 Radiance

Radiance vs 2 D Detection Angle Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002

Effective Specular Port Diameter Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 • Measure radiance vs angle on goniometer • Calculate cumulative volume as a function of radius. Normalize this volume to the average reflectance data for the sample. • Determine the radius that results in the percent cumulative volume that matches the SPIN-SPEX difference

Cumulative Volume vs Radius Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002

Effective Specular Port Calculation Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 • Measure radiance vs angle on goniometer • Calculate cumulative volume as a function of radius • Determine the radius that results in the percent cumulative volume that matches the SPIN-SPEX difference

Effective Specular Port Calculation Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002

Effective Specular Port Calculation volume inside that radius current radius Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 reff

Effective Specular Port Calculation Search through radii until we match the average SPIN-SPEX spectral difference Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 effective radius equation:

Detection angle Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 radiance Searching…

“SPIN”-“SPEX” SPIN = total Detection angle Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 radiance Found

Experimental: Spectrophotometers – – Datacolor Spectraflash 600+ Macbeth Coloreye 7000 BYK-Gardner The Color Sphere Minolta 3600 -d • All are d/0 devices • All have reasonable calibration status Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 • Four commercial-grade benchtop devices:

Experimental: Spectrophotometers Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 • Specifically chosen for their range of specular port configurations, from <1” to 2” • One (Minolta) has a detector in place of the specular port. SPIN and SPEX are calculated using the signal gathered by that detector

Experimental: Samples Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 • Two different automotive plastic samples (gray, tan) • Three levels of gloss • Total of five samples, two highly glossy, and three at two levels of matte surface • Underlying color identical, since various gloss levels are stamped in the same piece of plastic

Experimental: Samples Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002

Experimental: Sample Set gray tan Description Glossy Smooth Matte Rough Matte Glossy Smooth Matte Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 Color

Experimental: Samples Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 Smooth matte Rough matte

For Reference: Measured Gloss Level 20° 60° 85° gray glossy smooth matte rough matte 49. 2 0. 7 0. 5 65. 6 4. 7 3. 5 94. 3 22. 3 6. 5 tan glossy smooth matte 56. 2 0. 6 69. 7 4. 5 94. 1 20. 1 Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 Color

Goniometric Results: measured radiance Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 Note alternate ordinate axis for glossy samples.

Experimental Results: Cumulative volume Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002

Experimental Results: Effective Specular Port Size (°) Macbeth CE 7000 BYKGardner TCS Minolta 3600 -d Gray Glossy 1. 9 2. 0 Gray Smooth Matte 7. 6 4. 2 3. 2 5. 1 Gray Rough Matte 8. 3 5. 0 3. 9 4. 7 Tan Glossy 3. 7 3. 2 3. 3 Tan Smooth Matte 6. 2 4. 8 5. 8 7. 5 “Actual” 4. 7 3. 0 3. 6 *** Minolta 3600 -d has specular sensor Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 Datacolor SF 600

Experimental Results: Effective Specular Port Size (°) Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 SF 6000+ CE 7000 TC S

Experimental Results: Effective Specular Port Size (°) Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 Minolta data at arbitrary port width

Assumptions and Limitations or “Opportunities” Specular port vs sphere wall Specular port uniformity More comprehensive sample set Viewing/illumination beam Relationship to Gonio collimation Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 • • •

Conclusions Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 • Overall trend among spectrophotometers agrees with the physical measure • Ability to compare traditional designs to those with electronic port detection • May aid in device selection to best accommodate the application • Goal of inter-instrument correction not realized

Acknowledgements Special thanks are due to: – Danny Rich – Mark Fairchild – Roy Berns – Mitch Rosen who all helped tremendously with many fruitful discussions and emails. Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 This work was supported by the Munsell Color Science Laboratory

Fourth Oxford Conference on Spectrometry, Davidson NC, June 9 -13, 2002 That’s all. Thanks for listening!