[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Algae and Incandescent Lighting

"Tom Wood" <tomwood2 at flash_net> wrote:

> So now I'm confuzzled. I thought CRI was an attempt to quantify an esthetic
> parameter of the light source. If the CRI is at or near 100, but you still
> think the colors are not esthetic (David), or are unnatural (Ivo), then why
> bother with CRI at all? I guess what I was saying was that the lower Kelvin
> lamps that the plants prefer are not compatible with the higher Kelvin that
> our eyes prefer. Which is what I thought CRI is supposed to
> measure...grumble.

CRI is a measure of how well a given light source renders colors of opaque
surfaces. The standard for comparison is sunlight. That is, sunligth has
by definition a CRI of 100, and lower values mean that the colors 
perceived under the ligth source look shifted from their "true" values.

However, the operational definition of CRI, that is, the definition from 
which one can build an equipment to measure CRI in practice, is such that 
it depends on *relative* color shifts in between a limited set of standard 
color patches (7 if I recall correctly). Because of this narrower color 
space, one can have certain types of light that *operationally* measure at 
a high CRI, but whose result nevertheless does not look good to the eye. In 
particular, any black body light source within a relatively large range of 
temperatures, that is, any source that emits a spectrum in the form of a 
Planck curve within a relatively large range of *color* temperatures, will 
have a CRI near or at 100. The reason is that the solar spectrum can be 
very closely approximated by a Planck curve, and thus the *ratios* of 
illuminances gotten from the standard patches are the same, or approximately 
the same, when the patches are illuminated by either sunligth or by the 
given source with Planck spectrum.

Incandescent bulbs emit a perfect Planck spectrum, however with a much lower
color temperature than the sun. So they *operationaly* pass the criterion
for a very high CRI. But since their color temperature is much lower than
the sun, that is, they create a much reddish light, the resultant colors
do not look as natural as they would under sunlight. In practice, one
should not care for the CRI of a ligth bulb if the color temperature of
this light bulb is way too different than the solar value (about 5,500 K).
To say that, for example, a Triton bulb has a CRI of such and such, or
a HPS sodium bulb as a CRI of such and such, does not make much sense.
The color temperature being too far from the solar value, the color 
rendering index looses meaning. One should only pay attention to CRI values
when dealing with light bulbs that have color temperatures in a narrow
range. I would venture and say in between 4,000 and 7,000 K. For anything
out of this range, CRI is meaningless since the light is already "un-natural"
due to its large excess of redness of blueness relative to sunlight.

Sounds a bit complicated, but I hope I made myself clear enough...

- Ivo Busko
  Baltimore, MD