[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Green light

Peter Swinson wrote;

"It is worth remembering that today's CRTs put out a broad spectrum of light
that happens to peak in the green and this is why it looks biased towards
green when viewed directly."

Broad? If a light is perceived green its dominant wavelength IS green.
Although this tells us nothing about how the lights spectral distribution
is, green is what you want the least of.

"Conversely a Xenon lamp source biases towards the blue."

Xenon lamp is probably the best man made light source practically available
to provide a broadband spectral content. If compared to a green phosphor CRT
sure it looks blue....
In fact it its dominant wavelength is by far not as pronounced blue as the
green CRT is to green.

"What is important in film scanning is not the color of the light, or indeed
its intensity at any single point in the optical path."

Its intensity? Come on Peter!! Intensity IS number one issue how else could
you cope with density?
I would put my monthly salary on the fact that it IS the intensity and
spectral distribution of the Xenon light that attributes the Spirits quality
in extreme density detailing.
In the case of CRT then if you have slightly off spectral distribution maybe
then correcting filters could be used if you have enough intensity to trade
with. If you dont then what remains is gain in the pathe after the film has
been exposed. We all know what that means.

Ideally, to correctly reproduce film it should have had (tagged along with
the reel itself) a spectral ID of the light used to expose it. I.e same
light as exposed it ought to project it. The goal is to provide a linear
result to the viewer. Intermediary processes must obviously have to be taken
into account as these will modify the spectral distribution. As this is
somewaht difficult to do, then a scanning device that can produce lots of
wide spectral content at high intensities is a good second best.

"The only requirement that matters is that the complete path is accurately
matched, and that the detectors have the dynamic range and sensitivity to
capture accurately minute changes in light level,(corresponding to the
subtle density changes in negative and intermediate film images and the
extreme dynamic ranges of print stocks).

Well, if you begin designing a telecine the first thing I would look at is
what light sources are available to illuminate the film. Everything after
this is going to be a compromise, a fix to what could not be done correctly
in the first place. In terms of light source, what you want the least of is
green, second best is red and if you cant get an "ideal light source" the
the choice is blue bias. The reason for this is that RGB outputs from the
imagers is normalised to roughly equal levels at nominal white balance.
Since blue is the naturally smallest of the three components it has to be
subjected to the largest gain for normalisation. Hence if you can inject
more blue light then electrical compensation needs to be less.
"We are in danger of being brainwashed into the belief that 'blue-rich'
Xenon light is a pre-requisite for color negative."

Brainwashed? - these are simple scientific facts. But it is noted that the
only film loaded into a telecine is not color negative. But then Peter you
must know all of this said above so who is brainwashing who here?

I know it's several
years since I installed and serviced both contact and optical film
printers, but my recollection is that they all used Q.I. incandescent
"yellow biased" light. In film printing it is the print stock, that has the
"gain" in it to compensate for the color negative or inter-positive orange

I have never installed a single scanner or printer but this does not change
scientific facts.
Again, Peter, it is the light source that is what you should play with here.
Gain in film or electronics is just a post-fix.

"Incidentally have you ever seen the effect of exposing a frame of negative
with a concentrated slit of unfiltered incandescent light from a 150 watt
lamp, let alone Xenon light. It can take a worryingly short time for the
magenta layer to start to fade, permanently and irreversibly. "

Nope I have not, if this is true, then why would this be an issue in the
Spirit (to which you imply) where you can adjust the light intensity without
altering the spectral content? If the CRT you employ has such a wide
spectral content (albeit with faint green bias) then why would it not pose a
similar problem?

film does not take kindly to extreme light levels."

Correct, at some point the film will melt or ignite.

"Today's film scanning CRTs are optimised for the best illumination while
protecting the film's dye layers."

Hmmm. Most phosphors, and phosphor mixes , I have looked at have very spiky
spectral distribution indeed.

"The light source is just one component in a film scanning system, so beware
of judging it in isolation."

Yes, very true. It is not possible to design an ideal scanning system that
can reproduce each "single" wavelength of all colors AND at the same time
have acuity to detail.  Bit of a Heissenberg uncertainty theory over this.
The purpose of the scanning devices masking filters is to offer the best
compromise. It is here where the experience of a long term established
manufacturer weighs in favour to a newcomer who need a lot of empirical
experience to be fed back into its masking models. This takes time and since
we have talked about the Spirit, this is precisely what the early models had
a problem with. Corrected well enough since some time now.

However if the light source of the scanner is such that it restricts your
freedom to produce these optimal masking curves then there is a third
dimension of restrictions added that invariably will extend the compromise
to be made on either seeing all colours OR detail.

A final note. I have no relationship with the manufacturer if the Spirit nor
a particular grudge agains Cintel. What I do have a problem with is when the
laws of physics are conveniently "forgotten".
I may also have been noted to speak more often in favour for the Spirit than
the Cintel devices and I must say that I do admire the harmony between
theoretical and practical engineering as well as the visible results.  In
other words, nice work by Kodak and Phillips.

Mike Reichel

"I do not distinguish by the eye but by the mind which is the proper judge
of man"
"The future of television is the Internet"
Phone/fax/voicemail +46-(0)8-541 31500  Mobile +46(0)708-541792
Adress; Ytterbystrand 6, S-185 94 Vaxholm, SWEDEN
web; http://www.culinaire.se/JWA

Thanks to Casablanca Finish (Sao Paulo) for support in 1998..
No product marketing allowed on the main TIG.  Contact rob at alegria.com
952 subscribers in 36 countries on Sun Mar 15 07:07:04 PST 1998 
complete information on the TIG website http://www.alegria.com/tig3/