Split filtering (film, not paper)

[DrPablo]

I'm posting this prematurely, because I don't have results yet, but I've been toying a bit with split-filtering some exposures.

My strangest was yesterday. I was shooting with Efke 820 IR film, and I did a double exposure split between a #87 (740nm) filter and a tricolor blue filter. I think about 50% of the total exposure was with each one.

This was a scene in the woods with sunlight filtering through. My idea was to get a woods IR effect on the sunlit foliage, but use the blue filter to add a bit of luminosity to the shadows (lest they be jet black with the IR filter). I might have been able to do this with a standard gray card pre-exposure on zone 3 as well.

Any other thoughts / experiences with this type of insanity?

 

[AgX]

Never thought about this way. But is doesn’t sound insane to me.
The idea with this sort of photography is to grossly overexpose the leaves while keeping the rest of scene, which without filtration would even count for more impact on the film, reduced. If this would yield too black shadows, one could use a filter with a cut-off a lower wavelengths. But at harshly lit scenes this would only reduce the contrast between the white leaves and the rest of the lit scene, not affecting the shadows.
So your idea seems sound as long as the blue filter has no transparency in the near infrared. Could there be a time-parallax issue?

So much for the theory. Tell about your results. (I assume you have been making comparative exposures on the same scene...)

 

[DrPablo]

I'll develop the roll tomorrow and let you know. I've done a comparison of IR alone to the split exposure. I also have a blue versus no filter shot on a role of Acros 100 that I was shooting at the same time. I was using a Hasselblad, so I was able to use different films in the respective backs.

The sunlit areas I'm sure had pretty high values for all wavelengths, so I'm sure the stuff that is brightly reflective of IR will remain bright. I'm hoping it's not grossly overexposed compared with a normal exposure. The idea was to cut the total exposure in half for each one. The blue filter has a filter factor of 4, and I'm not quite sure for the IR filter -- I was exposing it at EI 3 or 6 for routine IR shooting.

Even if I used a very low cutoff IR filter, like a red #25 or even a yellow #8, this would considerably darken the areas that were in shadow. So what I've done with the blue filter is given them a half exposure in which their values may lighten a bit, rather than a full exposure where they're completely filtered away.

Parallax issues might arise, but that's ok. There are some prominent trees in the image that will (I hope) be totally sharp.

An interesting experiment that I may try is to compare proportions of the exposure with different filters. For instance, I can shoot IR only, blue only, 50-50, 25-75, and 75-25 to see which gives the most pleasing effect.

My idea, again, is because I really like the luminosity that blue filtration gives shadows in the woods -- so the combination of that and the ghostly white vegetation would be very interesting if it works.

 

[jim appleyard]

Sounds like fun! I look forward to a grand result.

 

[Bob Carnie]

This dosen't sound crazy to me, hope it works so you can post your results.
In scanning with a difficult scene we will do scans for the highlight, midtone and shadow and then combine them back into one.
Same basic principle of split printing on paper.
good luck
I imagine your use of multiple filters is to maximize the colour pallette that is present in different areas of the scene.

 

[Helen B]

Depending on which blue filter you are using, you might find that the blue filter passes IR just as well as it passes blue.

Best,
Helen

 

[DrPablo]

Really?

Which ones? And what kind of cutoff do they have?

 

[gainer]

Get hold of the CRC Handbook of Chemistry and Physics. It has a section on Wratten filters and their transmission characteristics. It doesn't show anything above 700 nm for 47 or 47B. Usually if there is transmission outside the visual region, there is at least a note to that effect. Both 47 and 47B have some transmission below 400nm, but nothing indicated on the other end.

 

[DrPablo]

Mine is a 47. We have a CRC in the lab -- I didn't realize that filter info was in there as well (at least using the names I'd recognize).

 

[AgX]

Blue filters can be different. That was the reason I hinted at their IR-transmission. I have not got my catalogues at hand and so couldn’t state the figures for the separation filters. The common types of a blue filter are relatively broad with a transparency at 800nm of 40% and still 30% in the visible red, which would not spoil the intended effect, but a more pronounced blue filtration would be nice. As you intend to make use of the wood effect that IR-transmission wouldn’t be actually a problem as you are overexposing in this region anyway, it would rather be the red region.
There are fine broad bandpass mineral filter glasses like the Hoya B 380.

Anyway, this is theory; practise will tell.

 

[ben-s]

You can make an effective IR filter using a few congo blue and a couple of red gels.
So a congo blue alone might work well;

Transmission data for congo blue here: Lee filters

You might want to get a swatch book for initial experiments - you can get them either free or for a nominal price.

 

[AgX]

Ben,

The filters from the small sample book are bit small for the use in front of many lenses, though I have used them this way before.

Based on Dr Pablo’s idea I too thought about an IR-filter which in addition shows transparency in the blue. That Lee no.181 (I forgot about it) is something like that (extrapolating its curve into the IR) but most probably it has too much red transparency. The red sensitivity of those near-IR films is high, their sensitivity towards longer wavelengths is decreasing. A result could be generally overexposed highlights instead of just white leaves.
But as the Wood effect doesn't have a sharp wavelength limit and this is about esthetical effects it's worth a try anyway, with a reduced exposure.

 

[ben-s]

Ben,

The filters from the small sample book are bit small for the use in front of many lenses, though I have used them this way before.

Very true you can also get the "master edition" which is huge. I've emailed Lee to find out the price and how to get it

It may be possible to build a custom filter by stacking a few gels (after studying the SPD graphs) to pass blue and IR.
I think the graphs in the swatchbook extend further into the IR than the ones on the website do.

 

[Helen B]

Get hold of the CRC Handbook of Chemistry and Physics. It has a section on Wratten filters and their transmission characteristics. It doesn't show anything above 700 nm for 47 or 47B. Usually if there is transmission outside the visual region, there is at least a note to that effect. Both 47 and 47B have some transmission below 400nm, but nothing indicated on the other end.

The CRC Handbook appears to be an entirely inadequate source of information on the behaviour of most Wratten filters beyond 700 nm if the data stop at 700 nm. You cannot extrapolate from transmission data that stop at 700 nm. The standard Kodak transmission graphs (Kodak Photographic Filters Handbook, Publication B-3) cover the range from 200 nm to 900 nm for most filters, with an expanded range where applicable. The Kodak data show that nearly all Wratten filters are long-pass, starting somewhere between 690 nm and 780 nm.

Without referring to published data, users of digital cameras may know from practical experience that the commonly-used filters in the RGB Bayer matrix over the CCD or CMOS all pass infrared radiation. The IR transmittance of most coloured filters is quite a well known phenomenon, I think. This property if IR transparency is shared by the dyes in colour film, which is why you can make an IR filter (ie IR pass) from unexposed processed E-6 film, and why IR can be used to detect dust on colour film.

The blue Wratten 47 has a long pass 50% transmission at 780 nm. The maximum transmission in the visible part of the spectrum is also 50%. Between 780 nm and 900 nm it passes more radiation than at any wavelength between 200 nm and 780 nm.

I use a green Wratten 58 with IR film. The 58 long pass has 50% transmission at about 745 nm which is about the same as its peak transmission in the green at 520 nm, and above 750 nm it passes more than its peak in the green.

Best,
Helen

 

[gainer]

If you notice the transmission data for IR passing filters, you will see that data start at 700 and go to 950. The data in my copy of the CRC handbook (58th edition)were compiled for Eastman Kodak.

The 58 filter was intended for color separation. What happens that far beyond the visible spectrum would not be of interest unless the camera lens were capable of seeing it. I have to tell a funny story about the Space Task Group of NACA which became the core of NASA. With the Mercury vehicle outside of the atmosphere, here was a chance to see things we normally could not. We couldn't get out and walk around yet, but we could design a special camera with a lens that could pass UV and IR, and so it was done. The scientists did not take into account the fact that the Mercury's window had essentially the same transmission characteristics as the atmosphere.

Whatever set of filter characteristics you use, be sure that your camera's lens can see the spectrum for which you have data.

 

[DrPablo]

The blue Wratten 47 has a long pass 50% transmission at 780 nm. The maximum transmission in the visible part of the spectrum is also 50%. Between 780 nm and 900 nm it passes more radiation than at any wavelength between 200 nm and 780 nm.

I use a green Wratten 58 with IR film. The 58 long pass has 50% transmission at about 745 nm which is about the same as its peak transmission in the green at 520 nm, and above 750 nm it passes more than its peak in the green.

Irrespective of IR transmission, though, I'd imagine that filters that transmit both visible and IR will have different filter factors for these wavelengths. Metering visible light through my 47 and my 58 gives a 3-4 stop filter factor. This is considerably brighter than the 7 to 9 stop correction I use for pure IR. So while there might be a bit of IR transmission, it would probably be overwhelmed by visible light.

I also can't really imagine a film other than HIE that is sensitive enough at 750 for the IR transmission to make a big contribution (in the face of the visible transmission). Certainly not the Rollei IR film, and the Efke/Maco 820 film is really on the downslope of sensitivity beyond 750.

So it might still be advantageous to do a partial exposure with a visible filter and the remainder in pure IR.