I have been reading a 1950s paper that recommends splitting the visible spectrum into two parts, one blue green and the other orange-red. It suggests using Wratten 40a and 28 filters, but as far as I can see these are no longer available.

Are there any direct modern equivalents? Failing that, are there any reasonably priced filters that will do the job?

I know where I can get shortpass and longpass dichroic filters and beamsplitters, but they tend to be both small and expensive. For now I want a tinkerer's solution. Grateful for any and all suggestions.

Haven't got my catalogues at hand, but for the blue-green you just could stack CC-cyan filters to get high absorbtion.

Haven't got my catalogues at hand, but for the blue-green you just could stack CC-cyan filters to get high absorbtion.

Thanks AgX. My first instinct is to take a pair of colour temperature correction filters of opposite mired value. An 80A and an 85B for example. My worry is that I won't let enough green through in the middle - which I think will also be a problem with two of a three-colour seperation set.

I probably need to see what the Bichromie gang are using.

Anyway, further suggestions welcomed.

>> It suggests using Wratten 40a and 28 filters... <<

Struan, don't know exactly what those would be, however I looked thru a Kodak Filter Handbook (1990) and summarized at the bottom of this post. Without knowing your exact intention, I'd suggest to try either a #24, 25 or 29 red (maybe you already have one of these) and a Wratten #44 (#44A is similar, but passes a small amount of UV). There will be virtually no crosstalk with any of these.

All of the filters around your missing numbers are sharp cutting, so I imagine the author of your paper also intended this. If you don't mind a lot of crosstalk, the stack of cc cyan might work,

I don't see #44 listed in general catalogs, but Tiffen still lists it on website. Tiffen took over the Kodak Wratten filter line some years back, as I recall.

Here's the more detailed data. I did NOT double check, so be forewarned. you should find the same Wratten data, in tabular form, in the CRC Handbook. (I think CRC only lists 400-700 nm)

- red filters: include Wratten 24, 25, 26 and 29. There seems to be a pattern. These are sharp cutting filters, blocking virtually everything from a cut point down to about 200 nanometers, where the graph ends. Here are the cut points where only about 1% of the light gets through (roughly, looking at graph). Wratten #24 -> 575 nm, #25->585 nm, #26 -> 590 nm and #29 -> 605 nm. Presumably, the missing #28 might have cut somewhere between 590 and 605 nm.

- blue-green filters: there is not much choice in this range. The number sequence runs: #39 (blue), #44 (light blue-green), #44A then #47 (blue tricolor). The jump between the #39 and #44 is so large that you can't really nail down where the missing #40 might have been. If you want to let some green through, only #44 and #44A seem to fit, so the choice gets real narrow. These two filters are pretty similar, except that the #44 mostly blocks UV whereas the #44A lets a small amount of UV through (<~ 1%). The #44 transmits (at 1% cut point) from about 425 to 565 nm. Beyond 700 nm it begins to transmit again.

Mr Bill, thanks very much.

The paper I've been reading is this:

Robert D. Bensley: "Natural Color Photography in Colloidal Silver"
Science, Vol. 112, pp 552-557 (Nov. 10 1950)

I can't find any followup work, citations or discussion of Bensley's work, so the whole thing may be a dead end, but he claims to have found a reliable way to tune the natural colour of fine-grained silver particles to reproduce the original colour of the light in the scene.

In short, you make two negatives with filters that pass opposite halves of the visible spectrum. You then make a multi-stage print with full-spectrum light (tungsten will do) by printing the two negatives in register, but between the first and second exposure you develop, fix, and rehalogenate with silver chloride.

Again: I have no idea if this works. The deafening silence in the published record may be a clue :-)

It's related to some of the research I do in my day job (I'm a physicist working with, among other things, the optical properties of nanostructures), but mostly it's just plain cool. I'd like to try the idea, and even if the result is an inexact Art rather than a reproducible Science, it's fun to play.

I suspect my first attempts will be with digital negs, but one motivation is the desire to do colour in ULF without being a millionaire, and two-exposures seems 50% better than tri-colour work.

Anyway, armed with the probable characteristics of the filters, I can have a play. Thank you, and AgX again too. If I find time to do this before the next ice age, I'll report back here.