[argyllcms] Re: CCT of a Speedlight
- From: ternaryd <ternaryd@xxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Tue, 13 Jul 2021 06:42:47 +0200
On Mon, 12 Jul 2021 22:07:44 -0400
"SteveJ" <sj9000@xxxxxxxxxxx> wrote:
Wouldn’t a xenon flash have the spectrum of
xenon?
Thanks for your thoughts.
Well, it should. But I understand that the
spectrum at the discharge takes some time
greater that zero to develop. (Very) low
powered flash emissions may be too short for
that. Also the technique to stop the discharge
for that lower power may have an influence. But
I don't know the details. From the little I do
know, I think it's plausible that the spectrum
changes (somewhat) according to power, but not
by >1000°K, and I can't see an explanation why
it should change by distance. Maybe I made some
mistake measuring?
If there’s also mercury in the mix there’d be
a slight spectral shift, but probably not
enough to be of concern.
Sure, chemical purity must be an issue. But
high(er) quality flash devices should have
impurities below a significance in photography.
Photographers better than me would have
protested loudly, wouldn't they?. At that time,
the Nikon SB-910 was the flag-ship flash of
Nikon's, so I tend do assume a certain level of
quality.
Anyway a bare xenon tube (with no filters in
the way) produces a SPD very close to daylight
(D65).
I'm reading an older book, from the beginnings
of digital photography, and that says that high
quality studio flashes could reach a CCT of
6500 (as an extreme limit), which is still less
than the 6800 I measured.
But if you check the specifications of many
compact or studio flashes, if they do mention
the color temperature, all give a CCT of
something between 5400°K and 5600°K ±200°K. If
they do say anything at all; I couldn't find
where Nikon does so. There are too many
different brands of those who do, to ignore
them.
So if I use one of those 910 together with one
of those others, I would get a mixed light with
a difference of 1200°K (to the specifications)
or more. Isn't this a concern for any
photographer?
"Close" is a relative term. How many wrinkles
can a spectrum have and still be close? I don't
know of any commonly used measurement to
describe the difference between two spectra.
Such a value must be complicated, metamerisms
also come into play. The CCT, as well as Duv,
compare uv, not the spectra, and are to old to
qualify as representative for human perception.
In this sense, Ben Goren is doubtlessly right
to point out the fuzziness.
My point was to reinterpret the CCT just as a
key into the formulas of the CIE-D series
producing again some spectrum. And, as CCT
describes a shift between blue and yellow, the
second shift, the tint, is on the red and green
axis, as I imagine, pretty much like on the
horizontal plane of CIE-Lab.
Just out of curiosity, I applied some color
filter to see how the spectrum, the CCT and
Duv would change. In one case, I got a CCT of
4996°K (aiming for D50), but Argyll found that
the closest Daylight temperature CIE-D would be
4843°K rather than 5003°K. As Duv considers
both axes, the influence of "tint" must be
included. But I didn't try to reproduce the
math involved.
While it’s not as smooth as D65 and there
might be a bit more UV, most well behaved
sensors aren’t affected by these small
differences.
As it's said of my camera that it has filters
to cut UV and IR, I'm not really concerned with
that while considering the spectrum of a light
source by itself. I guess, when illuminating a
sheet of paper with FWA, there might be a point
to it. But I expect such influence to be much
less than that of a color shift of 1200°K.
Hence there still is my question: What's the
color of a flash, or then: does every flash
have such a different color temperature that
they can't be used together in any combination?
Other related posts:
