Maker Pro
Maker Pro

There is a definition of "full spectrum"?

D

Don Klipstein

Jan 1, 1970
0
So far until tonight, I have thought that there was no official
definition of "full spectrum".

Now, I had a look at:

http://www.venturelighting.com/NaturalWhite/NaturalWhite_FAQs.html

My take is that this is a bit hyped, somewhat along the lines of hype
that could get worse if there was an "official definition" of "full
spectrum".

For one thing, that article does appear to me to overemphasize value of
higher s/p ratio, which I sense has been "fashionable" to do nowadays.
They say how higher s/p ratio makes a lamp better for outdoor
illumination (which I think is fairly true), but they talk about s/p ratio
until they say "Bring Daylight Indoors!"

I surely like most of my indoor lighting to be "warmer". I have a very
strong liking to about 3600-3800 K or so, 4100 K when I can practically
achieve a kilolux or two, and 3400-3500 K looks good to me at 300-500 lux
or so. At 100-200 lux, 3000-3200 K or so looks good to me!
At 50-100 lux, I like 2700-2865 K. At 30-50 lux, I like 2700 K. When
the illumination level needs to be less than 30 lux, I either like
"utility high-s/p-ratio-daylight" if I want "maximum illumination power
per watt" at expense of appearing "stark" and "dreary gray", or else I
like flamelike lower color temp. incandescent (or good high-CRI
equivalent) with color temp. close to 2500 K. And in a bedroom to enter
along with a spouse (or girlfriend/boyfriend) while carrying full wine
glasses and dressed skimpily, probably 1700-2000 K if the color is
going to approach blackbody.

One thing they did say:

"The IESNA Lighting Handbook (3-26) defines full spectrum lighting as
having spectral emission in all parts of the visible spectrum and in the
the near UV, with a correlated color temperature of 5000K or more, and a
CRI of 90 or more."

Does any edition of the IESNA Lighting Handbook actually define "full
spectrum" or "full spectrum lighting" as claimed above?

Why should 7500 K qualify as "full spectrum" while 4800 is "disqualified
from being full spectrum" when 5400 K is the CCT if "equal energy per unit
wavelength"? Should not mid-4,000's or maybe even low-4,000's qualify for
"full spectrum"?

Does 9300 K or 10,000 K shade of "lighter sky blue" qualify as "full
spectrum"? Does 20,000 K "deeper sky blue" qualify as "full spectrum"?
If so, should not a carbon arc at 3800-3900 K qualify as "full spectrum"?

Any and all comments please!

- Don Klipstein ([email protected])
 
A

Andrew Gabriel

Jan 1, 1970
0
As for "why 5000K?" Well, if we accept the definition that
the source "has spectral emission in all parts of the visual
spectrum", AND if we make a further assumption that the
amount of radiation at each wavelength should be of at least
the same general magnitude, that would favor high CCTs.

I would prefer a definition of full spectrum not to be tied
to any specific colour temperature, but to be that of a
black body at the same colour temperature (or within some
permitted deviation). i.e. it's a measure of a full
spectrum verses discrete spectra, and nothing to do with
actual colour temperature.
 
C

Chris Carlen

Jan 1, 1970
0
I.N. Galidakis said:
Hans-Christian Becker said:
I agree completely---to me that is the only definition that
makes any physical sense.

Here's a definition I came up with, yesterday:

A "full-spectrum" is a distribution S(nm) which has the properties:

1) Int(S(nm),nm=a..b) > 0, for any [a,b] in [400,700] [*]
2) dS(nm)/nm exists, for any nm in [400,700]

[*] Int(F(x),x=a..b) is the integral of F from a to b.

According to the above, a black body is a "full-spectrum", but not all
"full-spectrum" distributions are black bodies.

That is a desired consequence of a well designed definition.

For example, HPX lamps satisfy
the above as well, as does the Sulphur Lamp spectrum.
Right.

On the other hand, any spectrum which contains emission lines is not
"full-spectrum", because the distribution of such a spectrum is not
differentiable at the wavelengths of the emission lines.

What troubles me about this is that it's too physics oriented. What is
the context for this definition, light sources for spectrophotometers,
or light sources for visual illumination?

If the latter, then I see no reason why a spectrum consisting of
spectral lines shouldn't be considered as some measure of full spectrum,
if the lines are spaced closely enough to give high visual rendering
accuracy (CRI, if you will) for nearly all reflectance spectra of
practical objects.

Hence, the requirement #2 that you state above is too demanding, and
doesn't allow for any continuum between full spectrum or not. Can't we
devise a definition that gives a number which converges toward some
value representing 100% continuous spectrum when the test spectrum is in
fact continuous, and then approaches some other value such as zero when
a single line spectrum is analyzed?

To clarify, let's consider a densely populated line spectral
distribution closely matching the energy vs. wavelength distribution of
a blackbody. Since no practical visible object is likely to have a
reflectance spectral distribution that manages to completely miss say,
all the lines on the red end thereby rendering the color horribly
distorted, this dense line spectrum of the light source may very well
render the object as effectively as a black body source of equivalent
color temp. Thus, why shouldn't this line source be characterized
somewhere in the range of say 95% full spectrum?

If however, we were going to use it as a source in a spectrophotometer,
then we'd have big problems. So in that context, your original
definition is more suitable, albeit still somewhat problematic since
what spectral line is actually not differentiable? Every line has a
finite as opposed to infinitesimal linewidth and thus is differentiable, no?
The above definition is a bit "weak" with blended spectra (for example a
blended-light lamp spectrum is very close to a "full-spectrum", because its
distribution contains a black body distribution, but according to the above it
is not), but otherwise I think it's a good compromise.

What say ye?

Well, it's certainly some good thought as always.



--
Good day!

____________________________________
CRC
[email protected]
NOTE, delete texts: "REMOVETHIS" and
"BOGUS" from email address to reply.
 
D

Don Klipstein

Jan 1, 1970
0
I.N. Galidakis said:
Hans-Christian Becker said:
As for "why 5000K?" Well, if we accept the definition that
the source "has spectral emission in all parts of the visual
spectrum", AND if we make a further assumption that the
amount of radiation at each wavelength should be of at least
the same general magnitude, that would favor high CCTs.
I would prefer a definition of full spectrum not to be tied
to any specific colour temperature, but to be that of a
black body at the same colour temperature (or within some
permitted deviation). i.e. it's a measure of a full
spectrum verses discrete spectra, and nothing to do with
actual colour temperature.
I agree completely---to me that is the only definition that
makes any physical sense.

Here's a definition I came up with, yesterday:

A "full-spectrum" is a distribution S(nm) which has the properties:

1) Int(S(nm),nm=a..b) > 0, for any [a,b] in [400,700] [*]
2) dS(nm)/nm exists, for any nm in [400,700]

[*] Int(F(x),x=a..b) is the integral of F from a to b.

According to the above, a black body is a "full-spectrum", but not all
"full-spectrum" distributions are black bodies.

That is a desired consequence of a well designed definition.

For example, HPX lamps satisfy
the above as well, as does the Sulphur Lamp spectrum.
Right.

On the other hand, any spectrum which contains emission lines is not
"full-spectrum", because the distribution of such a spectrum is not
differentiable at the wavelengths of the emission lines.

What troubles me about this is that it's too physics oriented. What is
the context for this definition, light sources for spectrophotometers,
or light sources for visual illumination?

I would counter the disqualification-on-basis-of-differentiality
("differentiable" means that "1st derivative exists") on basis that no
emission "lines"/"bands" have "infinitely steep slopes". If an "emission
line" has its steepest edges going from 90% of that spectral feature's
peak to 10% of same in .01 nm or .001 nm or .0001 nm or .000001 nm,
the spectral power distribution remains a differentiable and continuous
function of wavelength.

- Don Klipstein ([email protected])
 
T

TKM

Jan 1, 1970
0
So far until tonight, I have thought that there was no official
definition of "full spectrum".

Now, I had a look at:

http://www.venturelighting.com/NaturalWhite/NaturalWhite_FAQs.html

My take is that this is a bit hyped, somewhat along the lines of hype
that could get worse if there was an "official definition" of "full
spectrum".

For one thing, that article does appear to me to overemphasize value of
higher s/p ratio, which I sense has been "fashionable" to do nowadays.
They say how higher s/p ratio makes a lamp better for outdoor
illumination (which I think is fairly true), but they talk about s/p ratio
until they say "Bring Daylight Indoors!"

I surely like most of my indoor lighting to be "warmer". I have a very
strong liking to about 3600-3800 K or so, 4100 K when I can practically
achieve a kilolux or two, and 3400-3500 K looks good to me at 300-500 lux
or so. At 100-200 lux, 3000-3200 K or so looks good to me!
At 50-100 lux, I like 2700-2865 K. At 30-50 lux, I like 2700 K. When
the illumination level needs to be less than 30 lux, I either like
"utility high-s/p-ratio-daylight" if I want "maximum illumination power
per watt" at expense of appearing "stark" and "dreary gray", or else I
like flamelike lower color temp. incandescent (or good high-CRI
equivalent) with color temp. close to 2500 K. And in a bedroom to enter
along with a spouse (or girlfriend/boyfriend) while carrying full wine
glasses and dressed skimpily, probably 1700-2000 K if the color is
going to approach blackbody.

One thing they did say:

"The IESNA Lighting Handbook (3-26) defines full spectrum lighting as
having spectral emission in all parts of the visible spectrum and in the
the near UV, with a correlated color temperature of 5000K or more, and a
CRI of 90 or more."

Does any edition of the IESNA Lighting Handbook actually define "full
spectrum" or "full spectrum lighting" as claimed above?

Why should 7500 K qualify as "full spectrum" while 4800 is "disqualified
from being full spectrum" when 5400 K is the CCT if "equal energy per unit
wavelength"? Should not mid-4,000's or maybe even low-4,000's qualify for
"full spectrum"?

Does 9300 K or 10,000 K shade of "lighter sky blue" qualify as "full
spectrum"? Does 20,000 K "deeper sky blue" qualify as "full spectrum"?
If so, should not a carbon arc at 3800-3900 K qualify as "full spectrum"?

Any and all comments please!

- Don Klipstein ([email protected])

I checked the 9th edition handbook and actually found the paragraph
where they take the wording from. From what I can tell, they are
lifting the statements out of context and using it for their marketing
purposes. The full paragraph actually reads as follows:

"While the spectral content of an illuminant might be expected to be
important for any task where the spectral content changes an important
aspect of the stimuli the task presents to the visual system, there is
little evidence that it is important for all tasks. This has not
stopped claims being made for whatt are called full-spectrum lamps.
These lamps, which have no widely accepted definition, are typically
fluorescent lamps with spectral emission in all parts of visible
spectrum and in the near UV, with a correlated color temperature of
5000K or more and a CIE General Color Rendering Index of 90 or more.
Claims have been made that use of such lamps benefit task performance,
human health, and happiness. These claims have little merit in most
cases"

They (the IES handbook) footnotes the paragraph with a study from the
National Research Council Canda. 1994. Full-Spectrum lighting
effects on performance, mood, and healh, edited by Jennifer A. Veitch,
Insitute for Research in Construction Report no. 659. Ottawa:
National Research Council Canada.

You can find the reports at the following link.

http://irc.nrc-cnrc.gc.ca/pubs/ir/ir659/index.php?lang=e

(One of the comment papers in the index was written by Terry McGowan -
perhaps he can chime in on your question too)


The NRC report brings back memories -- mainly of the passion surrounding the
claims made for so-called "full spectrum" lamps. As we've discussed
numerous times in this forum, however, full spectrum is and remains a
marketing term and I question whether it deserves anything more
formal.

Since the NRC report was published, we've learned that lamp
spectrum, by itself, is just one of many factors involving light that might
have non-visual effects on people. So, maybe we need some sort of
comprehensive "healthy lighting" index instead with the spectral component
factored in.

But, if a technical definition is needed, why not use use the definition
proposed in the LRC "Lighting Answers"? See:
http://www.lrc.rpi.edu/programs/nlpip/lightingAnswers/fullSpectrum/lightSources.asp
The relative cumulative SPD calculation which the LRC calls Full-Spectrum
Index (FSI) seems to do a good job of capturing the SPD variations and the
FSI easily answers what I think is the main question: How does lamp "X"
compare to outdoor daylight, incandescent light or any other sources that
claims to be full spectrum.

Terry McGowan
 
Top