RA-4 Tricolor Printing...Sensitivity Overlap??

[gmikol]

Ok...So I was thinking a little bit about tri-color printing, and I was looking at the spectral sensitivity curves of Crystal Archive. It seems like there's a lot of overlap between the blue-sensitive (yellow-forming) and green-sensitive (magenta-forming) dyes. It makes me wonder how one would get a really pure yellow.

So I went a step further, and convolved the dye sensitivity curves with the filter responses of tri-color printing filters. I've excluded the red-sensitive (cyan-forming) layer from this discussion, as it doesn't have any overlap with the others. So I have graphs for 47B (blue exposure), and 3 different filters for the green exposure (58, 61, 99). The curves are relative sensitivity, shifted up from the base vaules, so I could see more of the curves above the zero axis.

One shortcoming of my approach...the spectral data I had for the wratten filters (Kodak Photographic Filters Handbook, Pub. B-3) didn't go below 0.1% transmittance (3.0 density or absorbance), so I don't know exactly how strong the filters truly are in their stop-band. If any one has any more detailed spectral data, that might help a bit. I approximated everything in the stop-band as being 3.0 density.

Even with the sharpest cut filter (99), there is still appreciable overlap. So is it just a compromise or shortcoming in terms of being able to achieve a really pure yellow, or am I totally missing something (entirely possible).

Someone please straighten me out. But keep in mind I've never actually done any tri-color printing. This is as much a thought experiment as anything else.

--Greg

 

[Photo Engineer]

I think that you are misunderstanding something.

Imagine the tricolor filters as monochromatic. In this case, they have no overlap whatsoever. Expanding them into the real curves by widening beyond monochromaticity will give broader and broader curves, but at normal exposures will give less overlap (in fact virtually none) compared to white light printing.

I have done this test many many times and can show considerable improvement in separation exposures vs white light with CC filters, and I can assure you that there is no problem (overlap).

PE

 

[gmikol]

Thanks for chiming in, PE...I understand the theory behind tri-color printing, and also that many people produce excellent results by doing so. But I'm trying to reconcile that with what (I think) the graphs are telling me.

I understand that the filters have virtually 0 overlap, but the sensitivity curves of the different dye layers in the paper do. Let's take the example of an exposure with the 47B filter. It has some transmission all the way up to about 500nm. In the region between 450nm and 500nm, the blue- and green-sensitive dye layers both have some sensitivity (according to the graph, differing by about 1.8 log E ) It stands to reason (in my mind) that both will be excited by the energy transmitted by the 47B filter, and that there will be some magenta dye formed by the blue-light exposure. Now, I realize that the magenta will be formed in a very small proportion to yellow, but it seems like it would happen. Now, is it just the case that it is too little to be meaningful, or is there some other mechanism at play that means absolutely no magenta dye is formed?

And on a related note, do I interpret the graphs correctly to mean that a 61 filter is probably not appropriate for tri-color printing, and a 99 would be preferable?

Thanks, PE, and anyone else who might chime in, for your help.

--Greg

 

[Photo Engineer]

Greg;

Ahh, this is also a misunderstanding. The blue layer is about 5 stops faster than the red layer and so when you expose it to blue light, you are only "skimming" the top of the sensitivity and the blue sensitivity of the red layer is totally missed.

The same is true of the green sensitive layer.

So, there is no worry. And, this speed differential is there to partly compensate for the mask colors and partly to make it overly tungsten sensitive. With white light printing you therefore have to use about 50R average to fix up any extra mismatch.

All of this holds for both narrow band and white light printing.

PE

 

[gmikol]

Thanks, PE...I was looking at the curves a bit more and comparing it to the response curves of C-41 films, and everything's starting to make sense, now. I suspected the mask color on the film had something to do with it.

It's been a long time since I printed color, and I never felt like I was any good at it back then. I'd love to be able to try again some time, I just don't know when.

--Greg

 

[Photo Engineer]

Greg;

I used to tech color printing at the Kodak Park Camera Club. I had lots of students say the same and leave the class making outstanding prints. It is not all that hard.

PE