Experiments with RGB-colored screens... a la Dufaycolor & Autochrome

[holmburgers]

Hi y'all,

So I've posted a thorough explanation of what my proposed method is on the "autochrome recipe" thread and another older thread "restarting my autochrome experiments....". But I felt bad about hijacking these more historical discussions and thought I should just create a new thread to share my experiment/plan.

The goal is to recreate the look of autochrome & dufaycolor photographs. Instead of attempting to reproduce these methods exactly, which were developed 103 years ago, I figured there had to be some modern methods that would be much easier.

I came up with the idea of using slide film to record the reseau, RGB-colored screen, by photographing a white TV screen (RGB elements of equal amplitude) TV screen.

Then, thru pin registration, this screen will be attached to ordinary panchromatic b&w film, exposed, separated, reversal processed and re-registered.

Voila!

Naturally, the irony is how retrogressive this technique is... using color film to make a reseau only to permit making another color picture on b&w film. .....doesn't bother me a bit though....

So, after going thru all the theoretical steps, I've moved to the practical steps, namely, calculating magnifications. However, I've come up against some problems.

To make a .05mm TV pixel (approx.) equal to a .005mm autochrome starch we need a magnification of 1/10th. Take a 12"x16" TV screen and 4x5" film and my calculations say that the TV can't fill the whole frame at that magnification. SOO either we have to sacrifice resolution or find a higher-definition screen.

I haven't done the math yet, but perhaps a 1080 hi-def screen would do the trick. As I don't have one ($), a computer monitor is probably the next best thing.

However, this got me thinking about lens/film resolution. I haven't done a ton of looking into it yet, but it seems like 60 lp/mm is about the limit of resolution.(??) That equals .017mm per line pair.

Would it be fair to say that it's impossible to reach the resolution of the autochrome starch grains with a typical lens/film combination? I thought that making a reseau thru optical means would be significantly easier, and indeed it is, but at a cost I guess.

I would appreciate any input on my calculations & in determining the maximum resolution available by the method I've outlined. Either way, I'm still going to attempt this.

Thanks!

 

[holmburgers]

Took a couple pictures of a TV last night using EPP 4x5". I found it surprisingly difficult (at least on this set) to get a clean, blank image on the screen w/o the channel number or "no signal", etc. I had to hook up a DVD player just to get it to stop flashing at me.

I ended up using a relatively dark screen and exposures at f/11 were 60 and 90 seconds. That's giving a 2/3 stop increase for reciprocity failure and 1 stop for bellows factor. I'm concerned with reciprocity failure and how it might affect the color balance.

Therefore, it has occured to me that I need to use a bright white screen, which will mean shorter exposure times. Also, while looking thru a loupe and adjusting the color balance on the TV set, the phosphors didn't change color at all. So, it appears that I need a signal before I can actually manipulate the balance. The blue looked signficantly weaker than red & green did to my eye.

Where can one find a pure white television signal? How about creating one, any ideas? You wouldn't realize how rare such a thing is until you need it!

 

[Mustafa Umut Sarac]

may be you need to visit your sony repair service or a video editing service. They have bunch of signal generators , ossiloscopes etc. May be your problems answers lie at electronic forums , after learning you want , you come back here.

 

[holmburgers]

good call on signal generators, maybe i can find some cheap somewhere. but hey, this stuff is still analog

 

[Mustafa Umut Sarac]

If you are talking about video signal generators , I think they are expensive. May be you can try circuit bending like musician Gazzala and blow yourself or touch 25000 volts electron tube and say goodbye to forum Or you can give a add to your weather station tv at Kansas for 5 seconds and cost you may be 1000 dollars.
Than you photograph the tv at bulb and you get a white background and a us cloud map

 

[Photo Engineer]

It is so difficult to do that using a color dot matrix printer is much easier.

PE

 

[Mustafa Umut Sarac]

Agree. if you have forced to continue at hybrid photo , I will be there and help you. May be you can connect your pc to tv via tv card and load tv with photoshop if you want to do this way. And you need flat screen tv or buy a flat screen computer monitor , may be lcd help you but I am not keen on them. But you are after specific white RGB values and for each monitor technology they are different.
And white color is generaly 345 , ymc. I dont know the equivalent at rgb but same cmyk valus give brown , if you want grey , they must be aligned in harmonic order and different. You will be needed to real professional equipment to calibrate your monitor. I squeezed my brain and this is the post that I tried to help
And if you want to print rgb with inkjet , you will need specific inks , you can do with getty museum formulas and they will be original otherwise you will get a autochrome with may be strange modern colors. Dont forget to add perlin noise to the screen

 

[Ray Rogers]

What about using a digital camera to shoot a white background,
then hooking that up to the tv monitor?

I cannot even begin to do the math, but large screen (outdoor) displays are another option... you might get them to give you a white screen for a few seconds before their broadcast begins

Or just make your own...
get ahold of some unneeded white... (or black!) studio Background Paper and
spend the summer painting teeny tiny RGB dots on it.
When your finished, stand back and shoot it.

Are these good ideas, or do I need some more sleep?

 

[Mustafa Umut Sarac]

summer time , when the life is easy !

 

[htmlguru4242]

I'll put my two cents in here ... To create a white screen, you can always find a camcorder, small security camera, etc. that has TV output, and trick it into overexposing so far that the screen goes pure white ... maybe shine a flashlight into the lens at close distance?

Why not shoot a computer screen (CRT or LCD) instead?

 

[Q.G.]

However, this got me thinking about lens/film resolution. I haven't done a ton of looking into it yet, but it seems like 60 lp/mm is about the limit of resolution.(??) That equals .017mm per line pair.

Would it be fair to say that it's impossible to reach the resolution of the autochrome starch grains with a typical lens/film combination?

You're not hoping to reproduce the aliassing problems the Bayer-pattern filters in digi-thingies create, are you?

 

[Mustafa Umut Sarac]

If you want to force upon your idea , you will have to buy a monitor calibrating spectrophotometer. We were using 15 years ago on mac 100 Mhz , I dont know they are still at the market with a flexible software.
I guess , you will have many empty areas with monitor photographing , If dots are even still visually visible

 

[holmburgers]

I like htmlguru's idea... I have an old VHS camcorder, I could see what can be done with that. But actually, you raise a great point about using a computer monitor. That would completely simplify this signal generating portion of it. Just open up Windows paint... voila! That's brilliant.

However, can anyone comment on my concern about resolving power? At what point will I simply not be able to resolve the pixel elements?

And no, I'm not hoping to reproduce any aliassing problems, but we'll see what I get !

I realize a dot matrix printer would work, but I don't have one, and it wasn't my idea, so what do I have to gain?

I like the idea of painting a wall.... NOT! I'm not that zen! But I did have the idea of planting my backyard with three flowers that produce RGB, and then taking a photo of them for my screen at the peak of bloom. Groovy, eh?

Thanks for all the input, the slides are on their way to the processor, so I'll report back once I see them.

 

[Photo Engineer]

I'm surprised that QG has not suggested that you stop working on this tricolor approach!

PE

 

[holmburgers]

Howdy everyone!

Alright, well I got back my 4x5 slides and I have to say that I'm optimistic, at least for my 1st attempt.

There is a very pronounced blue cast in the reseau, which is odd since I thought the blue TV "pixels" looked pretty weak on the ground glass. Can anyone comment on typical color shifts that result from reciprocity failure? Lemme guess.... blue?

To the unaided eye it appears fairly continuous, though I'm sure the aperture-plate pattern from the TV will be apparent in the resultant photograph. Either way, I see red, I see green & I see blue; which in my book spells "additive color synthesis".

The attachments show the whole slide and then a close up. The scans look pretty miserable; it looks better in person.

Next step is pin registration (*cough cough*, office hole punch) and reversal processing. I think I might send a couple tests off to dr5 just because my darkroom is in tatters at the moment and I've never done reversal processing. Plus, photo formulary is still down.

Anyways, just wanted to share an update.

 

[Q.G.]

I'm surprised that QG has not suggested that you stop working on this tricolor approach!

PE

Unlike you, PE, i know when it is appropriate, and when not.

 

[Q.G.]

It may be a good idea (maybe not?) to see what you can do creating a digital file in PS or similar software, and then have that turned into a slide by a lab.

 

[Photo Engineer]

There is a very pronounced blue cast in the reseau, which is odd since I thought the blue TV "pixels" looked pretty weak on the ground glass. Can anyone comment on typical color shifts that result from reciprocity failure? Lemme guess.... blue?

To the unaided eye it appears fairly continuous, though I'm sure the aperture-plate pattern from the TV will be apparent in the resultant photograph. Either way, I see red, I see green & I see blue; which in my book spells "additive color synthesis".

Holmburgers;

The result is quite impressive.

You can expect to see the pattern in the final "color" photograph. There is a formula for calculating the required dot size for any size of reseau - film format match to minimize this and maximize color. It is in "The History of Color Photography" by Friedman.

At a distance, the slide should look gray if properly exposed. By manipulation in PS, I was able to get a more distinct gray result, which suggests to me underexposure and low contrast. This might help.

I don't think that the problem is reciprocity.

PE

 

[Mustafa Umut Sarac]

QG is right , some labs print computer file on to slide film with laser.
I cant wait the result holmburgers.

 

[JOSarff]

I shot tape to film transfers for a number of years so perhaps I can help.

You transparencies look about right. The image of a television is balanced at 6500 degrees K, not 5500K as color film, so it will come out blue. You need to either attach a color balancing filter set to the 4X5 camera, or rebalance the monitor. We balanced the monitor and bought 1200' rolls of 16mm ECN in multiple case lots of the same emulsion. Even with freezing the balance would change slightly as the film aged.

Also, with the video camera make sure it is in B&W, so you get an even gray scale, and no color shift from the camera electronics

 

[Photo Engineer]

JO;

Very good suggestion!

I had forgotten the color balance of the TV. I had in the back of my mind that color TV was balanced more blue than film daylight but I was not sure of myself. I have forgotten the reasons, but I should have remembered that. I did little work with TV applications and so this type of information is largely "filed away".

PE

 

[holmburgers]

Hey, thanks guys. JOSarff, that's great to know to color balance of a TV. What type of filter would be recommended? 85, 81, yellow?

When looking thru the slides it's not as much of a rich blue as the scan would suggest; more like looking thru an 80c filter or something. I was thinking that maybe I've made a tungsten balanced reseau!

As for the PS/digital recommendations, shame on you! This is APUG!

I'll try to get it more gray the old fashioned way.

PE, thanks for the book suggestion; I'm do for a library run and I'll definitely look for it when I'm there.

Again, thanks for the support everyone, you'll be the first to know what happens next.

 

[Q.G.]

As for the PS/digital recommendations, shame on you! This is APUG!

I'll try to get it more gray the old fashioned way.

Uhm...
It's exactly what you are doing: get a raster-image on film.
Only easier. And better.

 

[holmburgers]

(truth be told, I don't even have photoshop)

 

[JOSarff]

As I said, we rebalanced the monitor, but a weak to medium yellow should help. Ideally, a color temperature meter would give you the filter you would need. without one, try a cc10y, cc15y, cc20y etc. until you get a neutral film image.

I've attached wat is called a chip chart that you could try to use. It would help you plot the curve of the film image, and give you an idea of where you are in color balance.