Sammy said:
Does anyone know the color temp and the CRI of the famous "Ott Light"?
I did not see it published in their web site.
Color Rendering Index is not a helpful or explanatory measure of a
light's spectrum. It is of course popular and often available.
The original publication of the CIE's color rendering document in 1965
represents an unfortunate turning point. Some of the great authority
figures of color and vision in the 20th century were active, including
Dorothy Nickerson, Dean Judd, David MacAdam, Gunter Wyszecki. These were
important people because they made solid progress in their lifetimes and
had positions of authority from which they could, in effect, force
people to take notice of new ideas.
Unfortunately, the Color Rendering Index document contains hidden
assumptions. I know that many will shrug and say "Hidden assumptions,
who cares?" The problem is that hidden assumptions cause the discussion
to grind to a halt. Within the CRI document there is an explicit (not
hidden) assumption that natural daylight has a smooth spectrum and that
it can be modeled by the 1964 formula of Judd, MacAdam, and Wyszecki.
That is very well, it makes sense, it has been tested by other authors
and is a clearly stated useful idea. But what about object reflectances?
We know that they are smooth also, and the CRI authors knew that but
they did not say it. They imply it because of the 8 Munsell chips that
play a key role, but are there really 8 degrees of freedom?
The smoothness of object colors was analyzed and put forth by Jozef
Cohen in 1964, the same year that Judd, MacAdam and Wyszecki put forth
their daylight model. Cohen used the same method, Principal Components
Analysis, and reached a similar conclusion, that 3 vectors could go far
to simulate object colors.
Are the 8 chips a good proxy for other objects? Are 8 chips really
needed, or would 3 suffice? If 3 chips would give a similar result, this
is important, because it means that color rendering effects are
systematic, not random or obscure. In fact, the major effects ARE
systematic. They do not have to do with some infinitely subtle
interaction between tiny wiggles in the object spectral reflectance and
corresponding tiny wiggles in the lamp spectrum. Historically the color
rendering issue arose because many commercial lights are deficient in
red and green.
What does that mean, "deficient in red and green?" See this page on my
web site:
http://www.jimworthey.com/basicfacts.html
Scroll down to to item 5 and look at the graph. Three types of
broad-band light are compared to Cool White fluorescent, with all lights
matched for color and intensity. Anybody could prepare a graph like
this, but it takes a certain diligence.
So, the real problem of color rendering involves a systematic loss of
color information, but the CRI document does not say this. Perhaps the
authors had some understanding about it, but they did not say so. Later,
Judd proposed a further "Color Preference Index" that would have said
something about the systematic loss of red. Believe it or not, this is
still an active idea. (Not going anywhere, but an idea that gets kicked
around.)
I say, why not look at the basic issues in the clearest language
possible? That is what I do in the articles on my web site, especially
the introductory article, "Color Rendering: Asking the Question." It
will be published soon.
What is the practical "bottom line?" For one thing, some lights enhance
reds and greens, rather than dulling them. In particular, 3-phosphor
lights do this. They get downgraded by CRI because they are not exactly
daylight; but they do address the key issue. This is the kind of simple
fact that is forever obscured by mention of the CRI.
There have been several posts here concerning full spectrum lighting and I
was wondering if anyone researched this company.
They have done quite a job in the commercial market. I was at the Office Max
the other day and noticed one of their lamps.
The public has a legitimate concern that much lighting is inappropriate
for seeing. Some products are offered to address the problems, but
there is little clear discussion of what is needed.
The lamp and fixture I saw looked like it is possibly made by Philips and
then "Off Branded" to Ott Light maybe using their phosphor blend to get the
CRI in the upper 90's?
I don't know!
Any thoughts?
Those are my thoughts. I should add that color is not the only issue
with fluorescent lights. They are also too low in luminance and
correspondingly too large in area. This turns highlights and
shading---good sources of information---into veiling reflections, which
take away information.
Jim Worthey
