Re: Studying Direct Sunlight and Lamps Rendering

Roger Breton <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]
> I am trying to compare the lighting provided by artificial lighting such
as
> filtered tungsten-halogn (i.e.Solux4700K 50W) and 5000K fluorescent
> (Phillips Colortone D50 40W) to that of direct sunlight at 10:00 am,
eastern
> part of Canada.
>
> My experimental setup is as follows:
>
> I use two ColorChecker charts. Both of them are placed on the floor. One
is
> being lit directly by sunlight entering my studio whereas the second one
is
> being lit alternatively by the Solux lamp and the Phillips lamp.
>
> I know there must be 100s of fatal flaws in my experimental setup but I
find
> that none of the two lamps I used are capable of approaching the sheer
> appearance of direct sunlight in terms of quality.
>
> I know Jim Worthy recent discussions is touching on CRI so I thought that
my
> experiment might triggered or continue feeding into his discussion.
>

<SNIP>

> Roger Breton
> Laval Qc

Well, for starters, are you using sunlight illuminance levels on both
charts?  That makes a significant difference from the perceptual standpoint.

Unfortunately, the evaluation/measurement metrics that we have available to
us in lighting -- primarily chromaticity and color rendering -- aren't much
help with a task like you are talking about.   They give so little
information about what the eye is seeing as to be almost useless.  But, they
were never intended to be used for such things anyway.

I've had some luck with normalizing the spectra (plotted as SPD curves) of
light sources and then overlaying the spectra one on the other.  That shows
clearly where the sources do and don't match so proper filters can then be
applied.  However, what I've learned by the resulting experiments over some
years is that:

- Continuous spectra (like incandescent) are much easier to match to
sunlight than discharge spectra with all of the inherant lines and gaps.
- Enough red output, especially far red, is impossible to get except with
incandescent and high pressure xenon.
- Incandescent spectra take a lot of filtering to match any kind of
"high-noon" daylight so matching illuminances over other than small test
areas takes lots of lamps and watts.
- Illuminance level really counts.
- UV-A probably matters.  If you are looking at object colors with no UV
sensitivity, then O.K.  But many natural object colors utilize UV-A for
color and depth.  They fluoresce.  UV-A also causes the interior of the eye
to fluoresce slightly and that can affect perception.
- Sunlight has a high degree of variability.   The main variations are due
to latitude (atmospheric scattering) and altitude (water vapor) assuming
that you're after clear-sky conditions.
- If you must be limited to one source, high pressure xenon is a good choice
overall although it is a dangerous and expensive lamp to handle and use.  It
does have some lines, but the spectrum is continuous and only minimal
filtering is needed for spectral balance and UV.

Terry McGowan