Filter Question for b/w film

[Steven Ruttenberg]

For certain shots where things are not moving very fast, I would like to use a red, green, blue filter separately when taking an image along with an image with no filter (luminence only). One goal, is to be able to combine them to create a color image, but also have three separate color channels plus the luminence channel (4 total) to be able to use the channel mixer in photo shop to create a final b/w image. Digitally, what I do, is take the color raw, and then follow a process that operates on each of the three color channels separately and combines them plus the luminence channel to create the final image. I would like to do that with straight up black and white film with a red, green, blue filter (the same ones used when they would capture 3 separate images at the same time where each passed thru one of the filters. I think they did this most for movies)

Looking to see where I can get the three filters of the proper red, green and blue, do not want to use gels or anything like that. I have found some places to get them, but want to be sure they are the right ones. I believe they are the Wrattens, R25, G58, B47B. I want to find a quality set at least with a 77mm thread. Yes, it is time consuming, but I am also wanting to do some experimentation to see if having all the color channels really provides the best b/w. The idea is that in a digital sensor with the bayer filter, you you 1/2 of the pixels are colored green, blue and red are 1/4 of the pixels each. So, no matter how you combine them, you are missing a lot of information that is interpolated by the software in processing the image. I want to compare that to having 100% of red, green and blue, plus the luminence channel (4th film frame no filter)

Can't use it for every shot, but can for patient people when doing portraits, certain landscape scenes, most definitely city scapes, architectural images and other landscapes if the wind is not blowing the daylights out of everything. Clouds won't necessarily be a problem unless the exposure are silly long.

[Jac@stafford.net]

The predominance of Bayer's Green is intended to mimic the color sensitivity of the human eye/brain. There have been filter wheels for each of the filters. The auction site might be the best source. A DIY would work, too.

For the correct filters, go to this site and search (control-f) for 'tricolor'.


This is a source showing Russian color photography before color film, using the
tricolor filter technique.

[Steven Ruttenberg]

The predominance of Bayer's Green is intended to mimic the color sensitivity of the human eye/brain. There have been filter wheels for each of the filters. The auction site might be the best source. A DIY would work, too.

For the correct filters, go to this site and search (control-f) for 'tricolor'.


This is a source showing Russian color photography before color film, using the
tricolor filter technique.

I found this http://www.tiffen.com/Black%20&%20White%20Filters.htm

Those shots are quite impressive of old Russia.

[Jac@stafford.net]

I found this http://www.tiffen.com/Black%20&%20White%20Filters.htm

That does not show the specific tricolor designations.

[Steven Ruttenberg]

You have to look at the wratten number. They are supposed to be blue 47b, red25 and green 58 wbich are allocated listed as the color separation filters. Thse are listed there.

[Corran]

There's a thread somewhere on the forum that has pages and pages of discussion and information pertaining to this. Perhaps someone can find it.

I really think you are making this way harder than you need...consider that you'll need to do extensive tests for filter factors and possibly development factors to equalize exposure and contrast. I don't really understand what you mean by "best b&w." Basically you can make a color image, and then use PS to selectively desaturate colors post exposure? Uh, just shoot color film...

Personally I have been experimenting with the technique on and off for different reasons:



And yes light/shadow will be extremely problematic even at shorter exposures. Look on the bottom right of this image and you can see color artifacts from the shadows moving over the course of changing filters.

[Steven Ruttenberg]

There's a thread somewhere on the forum that has pages and pages of discussion and information pertaining to this. Perhaps someone can find it.

I really think you are making this way harder than you need...consider that you'll need to do extensive tests for filter factors and possibly development factors to equalize exposure and contrast. I don't really understand what you mean by "best b&w." Basically you can make a color image, and then use PS to selectively desaturate colors post exposure? Uh, just shoot color film...

Personally I have been experimenting with the technique on and off for different reasons:



Very interesting picture. Yes, I see the issues with the shadows and such. It isn't for every scene you may come across. In those situations, I might choose a filter or shoot all three as b/w and a luminence channel and pic the best of the three for post. Or just shoot a color frame and if I see it as b/w, then convert it post, using the channel mixer method, then either digitally print or print a digital negative the size I want and contact print it.

And yes light/shadow will be extremely problematic even at shorter exposures. Look on the bottom right of this image and you can see color artifacts from the shadows moving over the course of changing filters.

Generally when someone just desaturates a color image, you lose all of the color information and then your digital filters have 0 effect. If you adjust the color image first then desaturate you generally get a less than desireable conversion, again because it just desaturates all the color channels and you end up with a single luminence channel.

If you use the channel mixer, you can adjust each color channel independent of the other and then combine them any way you want and do final tweaks to the image, even going into Nik Silverefx pro and the color filters will still work on your b/w image made with the channel mixer. I used to just desaturate or let LR or PS convert it for me, but since I started using this technique, I have gotten much better results. Which brings me to my point. The digital sensor does not record the full red/green/blue information, only 1/2, 1/4, 1/4 and then does a lot of interpolation as to what a full sensor of red/green/blue would look like. By taking either a color film image and scanning it in (but I am not sure if there is a bayer filter of sorts on the scanner that again reduces the amount of informatino you are getting to the digital file) and going thru the process. Assuming the scanner, like the Epson V850 I does not use a bayer filter, then you get 100% of the color in each channel with 0 interpolation. This would give the maximum capability for converting from color to black and white.

What I am interested in and have been for quite some time, is how to record an image that has the three color channels that have 100% of the color information for each channel. With digital, you have to have your bayer filter removed, (that is 2500 bucks, but then you loose the anti-alias filter (I suppose you could add it back on, other wise you actually have and see individual pixels in the image if you zoom in to high mag, this is not necessarily bad as the image will be sharper) then use the three filters red/green/blue and combine in photoshop. Also need to be sure to record a luminence only channel. This also has the effect of so called, pre-filtering and your choice of three images that looks best if you don't want to work on the individual color channels and create the b/w that way.

Now to film, for many reasons, I prefer film to digital, especially on large format. I can take a standard b/w, choose a filter, use any type of b/w film still in production for varying effect and then adjust that in LR or PS if I want to scan it in. If I have three color channels of b/w plus a luminence only channel, I can then create a b/w that goes beyond the traditional look of b/w so to speak (I have more choices in post than I do if I just shoot black and white.)

Plus, I think it would be fun to create images this, even color images by combining the individual color channels. Also, for my astrophotography I will be using an image sensor that is monochrome so to create the cool color pics, I need to shoot a red/green/blue set of images and combine. I would also like to do that with b/w film as well. Main reason they use monochrome sensors for astrophotography is the greater resolution for each image you get.

Yes, it is a lot of work, but I like to experiment and I learn a great deal in the process. I believe I heard once that Thomas Edison supposedly said, after like a 100 failures at trying to make the light bulb he finally succeeded and when asked about his failures, he replied that he didn't fail, he just found a thousand ways not to make a filament because he only needed one.

It would be cool to find the thread that has this discussion as well.

[Bill Burk]

You could use a Foveon sensor.

You could photograph on a color transparency film, and then use the tricolor filter set to make separated negatives from the slide.

[Corran]

For me, in digital, I edit in Lightroom. I use the color channel section to change the luminance of individual colors to adjust an image to taste when making it b&w. Firstly I think the issues with Bayer interpolation as you describe, and has been written about a lot online, is way overblown. I am aware of the uneven RGB distribution of the filter, but I've never found myself really thinking "huh, this photo would be waaaay better if I had a few more R and B pixels..."

Anyway, I'm big on KISS, so you'll excuse me if I think shooting 4 sheets of film to make a color image to make a b&w image is a bit over the top, but if you want to help fund Kodak/Fuji/Ilford for a bit longer be my guest . As for being fun, well we all are a bit crazy with what we call fun while lugging around lots of big cameras and film, so yeah, you do you. Oh, also registration for the images like above is a pain.

Scanners do not have Bayer filters though so there is no worry there. Using color film though you'll have to decide WHICH color film, as they all look different, you won't have control over development for highlight control or other techniques, and of course the cost of development is much higher. If your final output is darkroom prints like me you'll have to burn LVT negatives for EVERY image.

Here's one more of the tricolor images I've made:

\

[Steven Ruttenberg]

Nice. It seems the only issues are with moving objects like water. The argument over loss of detail with bayer filtering, may seem over-blown, but for example, compare the blue channel to the red or green and you will see a huge difference. Since blue is the least sensitive I believe then having only 25% of the image as "true" blue "can" rob you of some detail, same with the other colors. Would most people notice or care? Doubt it, but when it comes to astrophotography the loss of information is not tolerated well. And, this isn't for every image, and as the last two examples show, definitely not with moving water or shadows. So mainly architecture, ghost towns, portraits, things like that, and then not all the time. Once novelty wears off and I have accomplished it, I will probably grow board with it and not do it often, espcially if the work flow is annoying or too long for my tastes. Besides, someone has to keep the film makers happy and buy their stock to keep them making it for the rest of the world.