[argyllcms] Re: Fluorescence - Graeme's article

Roger Breton wrote: 
In what Graeme calls the normal spectral reflectance model, there is only
one incident light source, Phi sub I (lambda), which represent the incident
luminous flux upon the surface characteristic of the illuminant.

But in the FWA model, there is a second luminous flux called Phi sub I (mu),
which represent some normalized quantity?


No, it's just the illuminant at the wavelength mu, rather than wavelength 
lambda,
since Fluorescence involves transforming energy from one wavelength band to
another, there are separate variables to notate this.

But "(mu)" then finds its way into
the Transmittance of the colorant. And then into the "FWA normalized
excitation spectrum".


Right, the contribution of the Fluorescent radiation at wavelength lambda
is a function of the weighted sum of the incident radiation at each
wavelength mu.

To me, there is only ONE kind of light source incident upon the surface. But
I agree then that the UV portion of that light has something to do with how
much fluorescence gets excited into the paper.


But we need to analyse the light source at independent incident and
emitted wavelengths to be able to deal with Fluorescence.

I am not clear on the "Overall FWA excitation level" and the Quantum Yield
factor.


It's the weighted sum of excitation wavelengths that trigger fluorescent
behaviour. Quantum yield is the efficiency with which the excitation is
converted to emitted light.

The "Excitation Spectrum range -- typ. 300-600nm" is not defined either. 

The wavelengths to which the Fluorescent material responds.

If I was to operationalize all the variables in Equation 3, I would not know
what to plug into E, Q and f(lambda).


That's what most of the rest of the paper is about, finding ways to
put numbers to those variables.

Before going on with some questions with the rest of the article, I would
like to ask about the following idea which does not see too far fetched for
me. I understand now that I need to evaluate two quantities:

A) how fluorescent is the unprinted paper,
B) how different light sources contribute to this fluorescence.

Question B has some implications for future color matching under different
light sources but Question A poses the fundamental question. Suppose I have
a spectrofluorimeter and I can indeed separate the light reflected off the
paper itself from the light emitted as a result of UV excitation, then can I
assume that the UV excitation will be constant over any colorant level, or
will it vary according to the particular colorant?


I assumed that it varies according to the colorant level, hence the 
Tc(lambda)
factor in the light reaching the fluorescent part of the model.

If it coud be demonstrated that the UV excitation off the unprinted paper
remains the same for any colorant level (a hard to prove proposition I can
image as 400% of ink is likely to absorb so much of the incident light
source that hardly and UV excitation will ensue) then the "excitation
spectrum" could be subtracted or multiplied spectrally from the raw spectral
measurements of any colors?


This idea is not supported by measurements. As colorant is added to
the sample, the characteristic "peak" caused by fluorescent emission
decreases. One has to assume that the light reaching the FWA declines
with more colorant, just the way it declines as it reaches the paper
(since the FWA is in the paper).

Graeme Gill.


Other related posts: