Light Meters
Versus
The Human Eye

The health department came to my work for a routine inspection.  We were assigned a new inspector.  It has been my experience over the past that each inspector has a different area they are more concerned with.  This inspector happened to be very interested in getting the lighting in the work areas up to the 50 foot-candle rating the health codes called for. 

Using the meter and talking with the inspector it soon became apparent that some areas appeared better lit than others, but registered lower on the light meter (foot-candles).  This was confusing since a candle power meter should measure the amount of light and it should be that simple.  The amount of light is the amount of light right?

Not understanding why some areas appeared brighter than others but were lower on the light meter’s reading bothered me.  That is why this is being written. 

Kelvin

The Kelvin rating was something no one seemed to be able to explain, other than quoting what they heard somewhere.  For example Kelvin refers to the color of the light, or Kelvin is a temperature rating.  In a way both of those statements are true.

Kelvin is a temperature scale similar to the Celsius scale.  The only difference between the two is what temperature is called zero.  Zero on the Kelvin scale is set at Absolute Zero (where there is no molecular movement).  Zero on the Celsius scale is 273.15 on the Kelvin scale and the increments per degree are the same.  Room temperature of 75°F would be 23.9°C or 297.04°K.

Kelvin to Celsius                     K = °C + 273.15
Fahrenheit to Celsius              F = 1.8(°C) + 32

It is not uncommon for a fluorescent light to have a Kelvin rating of 3500K.  If you convert 3500K to Fahrenheit you would get 5840°F.  I know light bulbs are normally hotter than room temperature when operating but I have never heard of one reaching over 5,000°F!  Obviously a lights Kelvin rating is not referring to its operating temperature.

To understand what the Kelvin rating of a light means we need to learn what a “Black Body” is.  Bear with me while I go over a couple things that will help understand what a Black Body is.

Light hitting an object can be absorbed or reflected by an object depending on what the object is made of.  Light can also pass right through it in the case of most gases.  The material an object is made of determines how much light is absorbed and how much is reflected.  Reflected light can change the color an object appears to be.  For example if you shine a blue light on a piece of paper the paper will take on a bluish color.

Light can be emitted by an object as well.  When an object is heated it begins to emit energy, the hotter it gets the shorter the wave length gets (the frequency increases).  When enough heat is present the frequency is high enough that the energy becomes visible to the human eye (light).  This is why iron starts to glow when you heat it.  The material an object is made of determines how much energy is emitted at a given temperature and the frequency distribution of that energy.

Different materials will glow different colors at the same temperature.  The different materials will have different wavelengths of peak energy radiation and will appear to be different colors.  The color they appear to be can also be affected by the surrounding light since they will reflect some of the light energy they receive.

A black body is a theoretical object that is a perfect absorber so no energy is reflected back and the object appears black, that’s where the name black body came from. It is also a perfect emitter meaning it will emit the maximum amount of energy at any given temperature, which is referred to as the black body radiation.

When a black body or any other material is at absolute zero there is no molecular movement and therefore no energy being radiated.  A black body at any temperature above absolute zero emits energy, at ALL wavelengths.  Although ALL wavelengths are being emitted there is a peak of energy at one particular wavelength and the wavelengths around it. A black body radiation curve shows how much energy will be emitted at each wavelength by a black body at a given temperature.  Each temperature has its own black body radiation curve that will show how much energy is radiated at each wave length and what the peak wavelength is.  As the temperature of the black body increases the wavelength of its peak output decreases (frequency increases).  When a black body is hot enough the peak out put is in a frequency range that is visible to the human eye (light).

Since the human eye sees different wavelengths of light as different colors, different temperatures of a black body would appear as different colors.  So a color can be directly related to a temperature.  A black body radiation curve shows that a black body does not emit a peak of energy at just one wavelength.  There is a range of wavelengths it will emit with the peak in the center of that range.  So the black body may not appear to be a particular shade of red or blue or green etc.  Instead it can look like white light but with a blue or red or green tinge to it.  Candle light is around 1,800K, normal day light is around 5,600K and the Kelvin scale goes up to 28,000K.

The temperature scale used when talking about the temperature of a black body is the Kelvin scale.  The Kelvin rating for a light refers to the overall color the light will appear to have.  A light with a 2400K rating means the light would appear to have the same color as a black body would at 2400°K.  The Kelvin rating of a light is also referred to as its color temperature.  Higher color temperatures are bluer in color while lower color temperatures are redder in color.

Here is the Black Radiation Curve for 5500K. These pictures were taken from a web page for Southwestern College’s physics department. http://cat.sckans.edu/physics/black_body.htm?textBox=2500
Radiation Curve for 5500K

Color chart for estimate of hue vs kelvin

The Kelvin rating of a light bulb can be misleading depending on the type of bulb. A black body will radiate at ALL wavelengths.  The amounts at each wave length may be high or so low it is immeasurable but all wave lengths are emitted from a black body.  Incandescent lights and the sun both represent a black body well in that ALL wavelengths are emitted.  With these types of light one can measure the energy out put at two points in the light spectrum (normally near red and blue) and calculate the Kelvin rating.

Fluorescent lights as well as all gas discharge style lamps do not fit in this category.  These have an interrupted spectrum.  Light is not emitted at all wavelengths.  There can be sections of the spectrum missing or there can be large spikes at various wavelengths.  Since the Kelvin rating refers to a black body it does not accurately apply to fluorescent or gas discharge lamps.  In the past, lighting companies used to use the term Apparent Color Temperature meaning that the color appeared to be at a particular Kelvin but was in fact a little off, due to the interrupted spectrum.  This term was replaced with Correlated Color Temperature (CCT) and soon lighting companies just used the term Kelvin to refer to a lights apparent Kelvin rating.

This misrepresentation of Kelvin for fluorescent and discharge style lamps is normally only a problem where accurate color reproduction is needed like TV and photography.  Objects will not appear to have the same color under fluorescent lights or gas discharge lamps as they do under incandescent lights or sun light at the same Kelvin rating.

As the selection of lighting styles grow there are more lights that have an interrupted spectrum and the term Kelvin is somewhat miss leading.  The term is still used since it helps relate the apparent color in familiar terms used for incandescent bulbs and day light.

The interrupted spectrum of a fluorescent light can make the color of objects appear different.  The lighting industry has a rating called the Color Rendering Index (CRI).  This index is an indicator of how well the color reproduction of a light is.  A CRI of 100 means it will render colors the same as a light source that does not have an interrupted spectrum that is rated at the same Kelvin rating.  A 4000K, 100CRI fluorescent bulb would have objects appear the same color as an incandescent bulb rated at 4000K.  A CRI value of less than 100 would indicate that the color objects appear would be different than an incandescent with the same Kelvin rating.  The CRI is calculated by an average of 8 different colors.  This means a bulb with a high CRI say 85 would render most colors fairly accurately.   With a high CRI rating a bulb could still be missing a portion of the spectrum that will make certain colors appear very different from what they really are.

The picture below was taken from the website for the International Cinematographers Guild http://www.cameraguild.com/technology/kelvin.htm and shows the typical spectrum of a fluorescent.
Typical spectrum of a fluorescent bulb

Light Output

Light output can be referred to in lumens, candle power, candela, foot-candles and others.  I started looking into the ratings of each in an attempt to find conversions or at least and understanding of the terms.  The definitions of each of these terms are fairly simple to understand if looked at individually.  Trying to relate one to the other or convert from one to another became a totally different story.  Some fall into the category of Radiance while others fall into the category of Illuminance.  And converting from one to another is not always possible even though we think it should be based on what the name is.

One worth pointing out is that candle power and foot-candles are not the same thing.  Candle Power is pretty much what the name implies.  It means that the light source is equal to that many candles.  Different candles will be brighter than others so a scientific method for standardizing this has been established.  Candle power is something that measures the amount of light that a source emits and is measured at the source.

A foot-candle is measured at the object being lighted rather than at the source.  It is a measure of how much light is actually making it to the object.  One foot candle is the amount of light present one foot away from one candle.  Increasing the distance obviously decreases the amount of light and this happens as the inverse square function.

Human Eye Response

When I was young I was taught about the two portions of the eye that respond to light, rods and cones.  The cones allowed us to see color and worked well in brighter light while rods worked well in low light levels and were pretty much a black and white vision. I was taught the rods were less concentrated around the center of the eye and that’s why you couldn’t see something well looking straight at it in very low light conditions.  The cones were not able to function in the low light levels, so you could see better if you looked off to the side in low light to see more with your rods.

There are almost no rods in the center of vision and there are almost no cones in the peripheral vision.  There is a zone where rods and cones both exist and is called Mesopic vision.

The rods do not see red.  That is why a red light used at night does not interrupt your night vision as much as other color light.  If you use a red flashlight at night then turn it off your eyes will still be adjusted to the darkness.  The dilation of the pupils is related to the Scotopic curve.  This is why high pressure sodium and fluorescent lighting can be bad on a person’s eyes.  There may be an absence of light at the wavelengths around the Scotopic curve.  When this is the case they pupils do not dilate and too much light is getting in hurting the eyes.  For this reason it is recommended to have an incandescent light as a supplement in work areas where only fluorescent lighting is used.

Picture taken from http://www.guidetohometheater.com/features/204eye/index1.html
Scotopic vesus Photopic curve.

Turns out it is not just the amount of available light it also has to do with the frequency of light.  The cones in the eyes respond better to a different frequency range of light than the rods. The curve showing the cones response is called the Photopic curve and the peak is around 555nm (around the yellowish orange portion of the color spectrum).  The curve for the rods is called the Scotopic curve and its peak is around 507nm (around the greenish blue portion of the color spectrum).

Light Meters

Most light meters are designed to measure the light mostly around the peak of the Photopic curve.  It was believed that the Photopic curve was primarily what they human eye responded to in normal lighting conditions.

Lights with interrupted spectrums like fluorescents or gas discharge lamps can give inaccurate readings on some light meters.  For example, there could be a spike of light around the wavelength the meter is most responsive to so it will give a higher reading on the meter.  On the other hand, there could be an absence of light around the wavelength the meter is most responsive to and give a lower reading.  This is why some light meters have a setting for various types of lighting like incandescent or fluorescent.

Recent research has shown that the rods are also active during normal lighting conditions and do contribute to how bright the light appears and how well we see.  Most light meters are designed to read mostly in the Photopic curve rather than taking both the Photopic and Scotopic curves into account.  For this reason one light may appear brighter than the other but the meter will tell you it is not.

This is where we return to the Kelvin rating.  A higher Kelvin rating means a shorter wavelength and therefore a higher frequency.  Since the peak of the Scotopic curve is at a higher frequency, a higher Kelvin rating may appear brighter.   The pupils react to light with wavelengths near the Scotopic curve, lights with a higher Kelvin rating may cause the pupils to dilate more and therefore appear brighter.

Things are not always so simple though.  A fluorescent light with a higher Kelvin rating may appear brighter because it produces more lumens in the Scotopic peak range.  But at the same time it may show higher or lower on a light meter.  The fluorescent light may have a spike or an absence of light around the peak of the Photopic curve (where the meter is looking).  This would have the light meter show you a higher or lower reading depending on this spike or absence.  But your eye will also be responding to the peak in the Scotopic curve which the light meter doesn’t see.

Lumens measured by the light manufacturers are weighted toward the Photopic curve as well.  Just looking at the lumen rating of a light can give an incorrect representation of how bright the light will appear.  Currently lighting is rated based on the Photopic curve so the specs can be misleading.

There are companies that make Photopic and Scotopic light sensors.  Lighting design professionals use light meters that measure light in the Photopic and Scotopic ranges.  These meters also measure the ratio between the Photopic and Scotopic curves (S/P Ratio).

When the pupils are dilated further open they allow more light in than is usually needed allowing aberrant rays to reach the retina.  Aberrant rays are rays let in by imperfections in the eye’s lens that almost all people have.  These aberrant rays cause blurred vision.  To dilate the pupils smaller a light that is brighter in the Photopic region can be used.  A more efficient way of dilating the pupils is with lights that have more light in the Scotopic region where the pupils react more.  This can mean energy savings to a large company.  You get more perceived brightness and better working conditions for less power.  Using lights, of the same power level but with more light in the Scotopic region, will increase the apparent brightness and give better light for working conditions.

Natural day light has an S/P Ratio of about 2.5. 

Lighting Specs

Looking at the S/P Ratio of a light can help determine how well it will light an area as far as perception goes.  Scotopic divided by Photopic (S/P ratio).

Let’s look at two lights.
Light A
32W, 3500K, 3000 Lumens and an S/P ratio of 1.45  

The efficacy (efficiency) of this bulb is 93.75 Lumens per watt.  (3000/32=93.75)
Since lighting manufactures use the Photopic curve this is the Photopic efficacy.  

Using the S/P ration we get something different.
S/P ratio of 1.45 means we have 4350 Lumens in the Scotopic curve.  (3000*1.45=4350)
The Scotopic efficacy of this bulb is 135.9 Lumens per Watt. 4350/32=135.9)    

Light B
32W, 6500K, 2800 Lumens and an S/P ratio of 2.10  

The efficacy (efficiency) of this bulb is 87.5 Lumens per watt.  (2800/32=87.5)
Since lighting manufactures use the Photopic curve this is the Photopic efficacy.  

Using the S/P ration we get something different.
S/P ratio of 2.10 means we have 5880 Lumens in the Scotopic curve.  (2800*2.1=5880)
The Scotopic efficacy of this bulb is 183.75 Lumens per Watt. 5880/32=183.75)

The lighting manufactures specs give ratings within the Photopic curve, so light A is said to be more efficient.  Light A has 93.75 Lumens per watt while Light B has only 87.5 Lumens per watt.  Since light meters also measure light in the Photopic range, light A will be the brighter light.

The human Eye uses the Scotopic range as well as the Photopic range, so to the Human Eye Light B would be more efficient.  Light B has 183.75 Lumens per watt while Light A has only 135.9 Lumens per watt.  Although light A would measure brighter on a light meter the human eye would disagree.  The human eye reacts to the Scotopic curve as well and would see the difference.

While Light B would appear dimmer on a light meter it would give better lighting for working conditions and strain the eyes of people in the area less.  This would mean better working conditions for the same number of watts used.

I hope this spreads a little light (pun intended) on why the meter appears to lie.

 

A helpful list of electrical terms can be found here https://electricalschool.org/  Thanks to Adrian for letting me know about this list.

Sources:

  1. http://www.aquabotanic.com/lightcompare.htm
  2. http://www.windows.ucar.edu/earth/Atmosphere/temperature/temp_scales.html
  3. http://www.cameraguild.com/technology/kelvin.htm
  4. http://galileo.phys.virginia.edu/classes/252/black_body_radiation.html
  5. http://www.egglescliffe.org.uk/physics/astronomy/blackbody/bbody.html
  6. http://cat.sckans.edu/physics/black_body.htm
  7. http://www.lightsearch.com/resources/lightguides/formulas.html 
  8. http://www.theledlight.com/lumens.html
  9. http://en.wikipedia.org/wiki/Photometry_%28optics%29
  10. http://www.energy2002.ee.doe.gov/Presentations/Tech/S3-Richman2.pdf#search='scotopic%20lumens%20photopic'
  11.  https://www.sylvania.com/en-us/innovation/education/light-and-color/Pages/color-characteristics-of-light.aspx (updated link)
  12. http://hyperphysics.phy-astr.gsu.edu/hbase/vision/rodcone.html#c5
  13. http://www.instrumentationguide.com/article/photopic_sctopic.htm
  14. http://ecom.mysylvania.com/miniapps/LightingCenter/PDFs/faq0016-0297.pdf#search='photopic%20scotopic%20curve'
  15. http://www.guidetohometheater.com/features/204eye/index1.html
  16. http://www.solar.com/spmeter.htm