Home of Professional Cinematography since 1996

Nanometer To Kelvin Conversion

Published : 26th December 2003


Does anyone know how to convert nanometers of light to kelvin?

John Lazear, SOC
http://www.geocities.com/no1camerasoc/



John Lazear writes:

>Anyone know how to convert nanometers of light to kelvin?

These terms refer to different things and are not convertible.

A specific example may make this a little clearer. A tungsten light -- 3200K -- is a full spectrum light, that is, it is emitting light over the entire visible spectrum. Red, orange, yellow, etc. Sunlight is also a full spectrum light, but at a much higher Kelvin rating, but still contains red, orange, yellow, etc. Many other light are also full spectrum xenon, carbon arc, but have different proportions of the various wavelengths.

Each of these full spectrum light sources contain light at a near infinite number of different wavelengths which what can be measured in nanometers. Visible light being the region between 400 and 750 nanometers.

A very specific color would have a very specific wavelength. For example, a ruby laser has a wavelength of 694nm.

Brian Heller
IA 600 DP



>A specific example may make this a little clearer. A tungsten light -- >3200K -- is a full spectrum light, that is...

I may be opening a can of worms here, but where do fluorescents lie? Since flos have to be measured in a convertible co related temp since they're technically not on the Kelvin scale, since Kelvin really only applies to incandescent sources.

Would you convert the flo to a correlated K temp and then figure the nm measurement of each spectrum? How accurate would that be?

This is definitely not practical information here, unless maybe you were shooting something for scientific analysis and such minute measurements mattered.

Marty Hamrick
Photojournalist/Cinematographer
WJXT TV,Jax.,Fl.



Brian is quite correct - nanometers refers to the wavelength of a single specific coloured light such as that emitted from a laser. Kelvin’s measure the colour temperature of a full spectrum light source.

However there IS a relationship between a colour temperature and the peak wavelength in its spectrum. It's called Wien's law.

Wavelength (nanometers) = 3,000,000 / Col temp (Kelvin).

So at 4,500K, the peak wavelength is 666nm (red) at 6,000K the peak wavelength is 500nm (bluish green) and at 7,500K the peak wavelength is 400nm (deep blue)

Outside these temperatures, the peak is outside the visible spectrum.

And in every case, that's just the peak wavelength - all other wavelengths are present as well, in slightly lesser intensities, adding up to a more-or-less white result.

Not sure if that's any use to anyone in a practical sense.

Dominic Case
Atlab Australia



> where do fluorescents lie?

Since they don't match any particular spectrum curve for any particular color temp, they don't lie anywhere!

The correlated color temp of a fluoro source is simply the color temp of an incandescent source that matches it closest in terms of the R, G, B tristimulus responses. It's not always possible to match at all, if the green component for example is too high or too low for the other two colours.

And once you've done that, there is no guarantee that the peak wavelength of that colour temperature appears in the fluoro's spectrum at all.

Dominic Case
Atlab Australia



Marty Hamrick asked :

>...where do fluorescents lie?...

Fluorescents just lie.

Wade K. Ramsey, DP
Dept. of Cinema & Video Production
Bob Jones University
Greenville, SC 29614



Dominic Case wrote :

>...Not sure if that's any use to anyone in a practical sense.

At least not to anyone on this list, probably, but it's nice to learn something new!

Wade K. Ramsey, DP
Dept. of Cinema & Video Production
Bob Jones University
Greenville, SC 29614



Another way to get a good idea of where nanometers lie in the spectrum and how peaks work is to check the Lee and Rosco swatch books where a handy graph is provided showing the relation between transmission and nanometers for each effect colour. This way you can also work backwards by finding the colour that matches your source and then reading the graph alongside it.

Just my two gels' worth.

Roger Simonsz
DP/Operator
Paris



>I may be opening a can of worms here, but where do fluorescents lie?

Everywhere and all the time

But seriously, folks... the light that is emitted from fluorescent tubes is made up of the specific wavelengths of light emitted by whatever witches' brew of phosphors are coating the inside of the tube (with no disrespect towards Wiccans) coupled with the emission lines that you get from the low pressure mercury arc which runs down the tube and excites the phosphors.

How this discontinuous spectrum of added up wavelengths is perceived depends on the eyeball, chip, film stock or other receptor's response curves.

A high CRI fluorescent (color rendering index) with a specified color temp (expressed in degrees Kelvin) might have a response curve that is not too different from an ideal black body radiator (of which a tungsten filament is a reasonable facsimile) at that temperature, but it might be a bit strong or weak in some wavelengths that might or might not be noticed by the film.

The 5th and 6th editions of the ASC manual had spectral distribution curves of some fluorescents in them so you could get a sense of this sort of thing.

I don't know if the dist. curves survived into the 7th and I did not notice them in the 8th, though I did not look that hard.

Mark Weingartner
LA based.