Light Measurement Instrumentation
By David Brainard and Denis Pelli
- General considerations
Other considerations
Making an ICC profile
Vendors
References
-
-
-
We get asked about light measurement instrumentation (in the context of vision research) frequently enough that it seemed worth writing out a relatively full answer. Our experience with such instruments is mainly as users, not as designers or builders.
Thanks to Bob Cole for correcting some broken links and providing additional ones, to Alex Wyer for updating us on the sale of Graseby Optronics, and to Erich Hohl for updating the information for LMT.
Please send comments and suggestions to brainard@psych.upenn.edu.
General considerations
It is helpful to understand the different types of information you can get from different light measurement devices. Here is a quick taxonomy.
- Photodiodes: If you don't need calibrated data but just need to have a signal that varies with the amount of light incident on a detector, you can get an uncalibrated photodiode. These are very cheap and respond monotonically (and maybe even linearly) to the light energy incident upon them. Typically they respond fast. I use these (hooked to an oscilloscope) for purposes such as verifying that my software is really flickering the screen at the rate I think it is or measuring phosphor decay times. My understanding is that these devices are quite stable over time, so they can also be incorporated into feedback circuits for stabilizing light sources, although I have not personally tried this.
- Radiometers: These devices are calibrated and provide a calibrated measurement the amount of light power incident on the detector. They are univariate devices and do not distinguish between light of different wavelengths. The ones I have used are based on some type of silicon diode technology and come with electronics that allows them to be used over several decades of light intensity. These devices are very useful purposes such as calibrating Maxwellian view optical systems, where the light has usually been made monochromatic before it reaches the eye.
- Photometers: These are similar to radiometers but rather than providing a reading of incident energy, the contribution of light at each wavelength is weighted by the photopic luminous efficiency function.
- Colorimeters: These devices have three separate univariate channels, rather than the single channel found in photometers. The three channels are designed to have the spectral sensitivities of a standard human observer, so that measurements made with them can be used to predict human color matches. The standard observer used is almost always the CIE 1931 standard observer.
- Spectral radiometers: These devices provide measurements of the light energy at a series of sample wavelengths throughout the visible spectrum. For color research, which is what I am most familiar with, you want either a colorimeter or a spectral radiometer. There are a number of tradeoffs that are worth bearing in mind.
Other considerations
- Spectral radiometers generally cost more than colorimeters, although mass market applications for color light measurement (primarily color printing) seem to be bringing some inexpensive spectral radiometers to market.
- The accuracy of colorimeters is determined by how well the spectral sensitivities of the three channels match the standard observer. Often, these devices are designed so that the match is very good on average with canceling deviations across the spectrum. This is fine if you are measuring spectrally broadband sources. But if you are measuring narrowband sources or sources with sharp spectral peaks, you want to pay careful attention to how well the curves match at the wavelengths you care about. Note that the red phosphor on most monitors contains sharp spectral peaks. Much of the cost of colorimeters is getting the filters just right, and this is one way that expensive ones can be expected to be superior to cheap ones. Some colorimeters may have curves optimized to work well with monitors.
- Colorimeters are typically made to match the CIE 1931 standard observer. In color research, one often wants to use other standard observers (e.g. the Judd-Vos modified version of the CIE standard observer, the CIE 1964 large-field observer, or observers derived from Stiles' color matching functions.) Colorimeters based on the 1931 observer do not provide enough information to compute tristimulus coordinates with respect to these other observers unless some additional facts about the stimulus are known.
- As colorimeters age, their filter properties can change, sometimes drastically. Although I don't think you really have to send the devices back to the manufacturer for recalibration as often as the manufacturers recommend, you do want to keep an eye on your device. One good strategy is to pick some standard lighting configuration in your lab and measure it when you first buy your instrument. You can then remeasure this configuration periodically. Drastic changes are a cause for concern. For spectral radiometers, measuring a fluorescent lamp is useful because you can make sure the lamp's spectral lines show up where they are supposed to.
- Although spectral data generally tells you more about a stimulus than colorimetric data, there is a performance cost for using spectral radiometers, namely that the signal-to-noise ratio is not as good. A radiometer takes the incident photons and measures those in each waveband
separately, so that the noise for each measurement is greater. My radiometer has difficulty at the low end of my monitor's gamma curves.
- Spectral radiometers vary considerably in how finely they sample the spectrum. It is possible to pay a lot of money and get sampling every nanometer or pay a lot less money and get sampling every 10 nm. In addition to how finely they sample the spectrum, these instruments differ in the bandwidth measured at each sample point. One thing that you want to be true is that light at every wavelength is accounted for somewhere. That is, it would be bad to have a device that sampled every 10 nm with a 1 nm bandwidth, because then your instrument would be completely blind to light between the sample points. I assume that most devices you can buy are setup correctly in this regard, but I have never actually checked. If anyone out there knows more, let me know.
- You really want to have your device interfaced to your computer. This will allow you to make automated measurements. Many devices have a simple RS-232 interface, but others require a special card or simply give you an analog signal. Make sure that your computer can tell the device to take a measurement as well as get the data back.
- One important way that devices differ is in what geometric configuration they use to measure light. Some devices actually allow you to image stimuli from a distance, while others simply contain a detector that you must get the light onto. Of the latter kind, some are set up with a little rubber hood that is designed to allow you to paste it onto the front of a monitor for monitor calibration.
Making an ICC profile of your monitor
You can buy a characterization package consisting of a program (e.g. ColorBlind ProveIt! or Monaco EZColor) and a simple colorimeter to easily create an ICC profile for your monitor. Using the Psychtoolbox IccProfile.mex, it should be easy to use that calibration data to do your color corrections, but we haven't actually tried that yet. Here are a few calibrators that seems to be good, from what we've read. We have not evaluated these ourselves. We've included links to the manufacturer (if available) and to Jon Cone's inkjetmall web site because he knows a lot about digital printing and has good advice about choosing among these products.
ColorBlind ProveIt! with Sequel Chroma 4 CRT & LCD monitor colorimeter $279.95
Available from Jon Cone's Inkjetmall.
Monaco EZColor software $299
MonacoSENSOR colorimeter $249
(or both together for $499)
The ColorBlind+Sequel package, for 279.95, is cheaper than the $499 Monaco package, and the ColorBlind software seems to be at least as highly regarded as Monaco's by Jon Cone and others using it to calibrate their monitors for critical digital color printing. Monaco or ColorBlind also sell software that may be used with other colorimeters or radiometers.
ColorVision Monitor Spyder (calibration puck) with PhotoCal software $199
ColorVision Monitor Spyder (calibration puck) with OptiCal software $399
ColorVision Profiler Pro software, uses a spectrophotometer.
The ColorVision products are getting rave reviews in the discussion groups. The Spyder has a cool look, but I don't know if its specs are at all remarkable.
Here's a 2001 comparative review of the Monaco and Colorvision products that says they're both great.
Vendors
Probably what you are really looking at this page for is vendor information. Here are some vendors and what I know about them. I'd be happy to include others if they are brought to my attention. The list is arranged in alphabetical order.
- Analytical Spectral Devices
- 303-444-6522 (voice)
- 303-444-6825 (FAX)
- http://www.asdi.com/
They design and manufacture portable spectroradiometers.
Ancal, Inc.
702-896-5787 (voice)
602-532-7018 (FAX)
http://www.ancal.com/
Distributors of miniature fiber optic spectrometers, accessories, and software.
Cambridge Research Systems Ltd
+44 (0)1634 720707 (voice)
http://www.crsltd.com/
OptiCAL photometer ($1,500) and ColorCAL colorimeter ($1,875) designed for monitor calibration.
Control Development Incorporated
(219) 288-7338 (voice)
(219) 288-7339 (FAX)
http://www.controldevelopment.com/
Dalotech
(603) 606-2881 (voice)
http://www.dalotech.com
Data loggers connected to various light measurement instruments.
Datacolor
(800) 924-2189 (Voice)
(609) 895-7472 (FAX)
http://www.datacolor.com
They make spectrophotometers and densitometers and provide software and systems for color formulation, color matching, color quality and color measurement.
- ELDIM
- (33) 231 947 600 (voice)
(33) 231 940 950 (FAX)
- http://www.eldim.fr
-
- They make photometers and colorimeters based on Fourier transform optics.
Gamma Scientific
(800) 637-2758 (voice)
(619) 576-9286 (FAX)
http://www.gamma-sci.com/
A source for photometers, spectral radiometers, colorimeters. They used to be EGG Gamma.
Gigahertz-Optik Inc
(978) 462-1818 (voice)
(978) 462-3677(FAX)
http://www.gigahertz-optik.com/
A source for UV-Vis-NIR radiometers, photometers, color meters, integrating spheres, laser/optical power meters, R/T instruments and optically diffuse materials.
Hagner Photometrics Instruments
+44 1243 575723 (voice)
+44 1243 575907 (FAX)
http://www.hagnerlightmeters.com/
They sell illuminance meters and photometers.
- IMPHORA Inc.
- (510) 689-2039 (voice)
- (510) 689-2788 (FAX)
- kmuray@msn.com
-
- They make high precision photometers and colorimeters as well as custom
- light measurement devices.
-
- Instrument Systems
- (613) 729-0614 (voice, North America)
- http://www.instrumentsystems.com/
-
- They make spectroradiometers (diode-array, CCD-array, and scanning) for a variety of applications such as color measurement.
They make a variety of light measurement instruments and detectors. I have never used their products. If you call them you can probably get them to send you a little booklet they produce, called the Light Measurement Handbook.
LI-COR
(800) 447-3576 (voice)
402-467-2819 (FAX)
http://env.licor.com/
They make a variety of light measurement instruments and detectors.
LMT
01149-30-393-4028 (voice, Germany)
01149-30-391-8001 (FAX, Germany)
http://www.lmt-berlin.de/
They make high-end photometers and colorimeters. The web page is not currently (5/21/01) very informative. You might try sending email to: LMT@LMT-Berlin.DE. (Thanks to Jason Barrett for pointing this out.)
- Lumetronics
91-22-8202170 (voice, India)
- 91-22-8360730 (FAX, India)
- http://www.lumetron.com/
They make photometers as well as a wide range of test equipment for all kinds of stuff.
- Microdaq
877-275-9606 (voice)
603-456-3529 (fax)
http://www.microdaq.com/
-
Small inexpensive data loggers, including ones that can measure light intensity and light on/off states.
Minolta
(201) 825-4000
http://www.minolta.com/
- Photometers, colorimeters, spectrophotometers, and spectroradiometers. I used the CS-100 when I was at Rochester (1990) and was quite happy with it. It was easy to interface to a computer over an RS-232 line, although I think you need the add-on control box to do that.
-
- In the US, they like you to buy through a vendor, not directly. The number above is a main Minolta number and you will have to work from there to a vendor. In the UK you can buy direct.
- Minolta (UK) Limited
- Tel: +44(0)1908 200400
Fax: +44(0)1908 618662
http://www.Minolta.co.uk/
-
- Molectron Detector, Inc.
- (800) 366-4340 (voice)
(503) 620.896 (FAX)
- http://www.molectron.com/
They manufacture laser power and energy meters incorporating pyroelectric and thermopile sensors.
- Ocean Optics
- http://www.oceanoptics.com
-
- They make a variety of radiometers and measurement accessories.
-
- Onset Computer
- (800) 564-4377 (Voice)
- (800) 759-9100 (FAX)
- http://www.onsetcomp.com
-
- They make inexpensive stand alone, battery-operated data loggers for measuring and recording changes in light intensity (e.g. HOBO H8 RH/Temp/Light External).
Photo Research
(818) 341-5151 (voice)
(818) 341-7070 (FAX)
http://www.photoresearch.com
They make spectral radiometers, colorimeters, and photometers. I own their PR-650 spectral radiometer, which I am quite happy with. It cost me about $13,000 in 1991. I've had to send it back to be fixed twice in six years (shutter problems). It talks over a serial line (RS-232) and is easily interfaced to a Mac. (You are welcome to my code if you'd like it, it is included in the Psychophysics Toolbox.) It samples between 380 and 780 at 4 nm increments, which is good enough for most vision research. If you do use a 650, it would be very wise to check its calibration regularly, either by sending it back to the factory at the recommended 6 month interval or by measuring a stable standard in your lab.
- Pico Technology
- 01480 - 396395 (UK) +44 1480 396395 (INT)
- 01480 - 396296 (UK) +44 1480 396296 (INT)
- http://www.picotech.com/
-
-
- Low cost data logger that has built in sensors for light, temperature, sound level, etc.
-
- Radiant Imaging, Inc.
- (773) 832-1647 (voice)
- (773) 388-2578 (FAX)
- http://www.radimg.com/products_prometric.htm
-
- CCD photometry and radiometry measurement systems, light source characterization solutions, optical and illumination engineering consulting services, and illumination design and analysis software products.
Semicoa Optoelectronics
(714) 979-1900 (voice)
(714) 557-4541 (FAX)
- A source of cheap photodiodes with build-in amplifiers. I'm sure lots of sources make these. I found a couple lying around my department and have never had to buy one.
Formerly a division of CVI Laser. They make a variety of radiometers and measurement accessories
B.W. Tansley & Associates, Ltd.
(613) 692-2089 (voice)
(818) 692-0872 (FAX)
http://www.iqcam.com
They make an imaging photometer (the IQCam), which would be useful for purposes such as measuring spatial variation in stimulus luminance.
They make radiometers, photometers, and colorimeters.
TRICOR Systems Inc.
(847) 742-5542 x40 (voice)
(847) 742-5547 (FAX)
http://www.tricor-systems.com/
They make a variety of photometers and imaging photometers.
- UDT Instruments
(800) 891-2709 (voice)
(407) 673-7005 (FAX)
- http://www.udtinstruments.com/
They sell colorimeters and radiometers. They used to be called United Detector Technologies and then Graseby Optronics. I have one of their radiometers and my impression is that they are widely used in the vision community.
They make various inexpensive radiometers intended for use in color reproduction applications.
References
If you want to know more about light measurement, see the relevant sections of my psychophysics bibliography.
-
visitors since 20 March 2000.
-
- David Brainard, Denis Pelli & Allen Ingling.
psychtoolbox@yahoogroups.com
27 March 2005
