There is a formula in Friedman that can be used to calculate the size of individual dots needed for proper viewing.
PE
PE
Experiments with RGB-colored screens... a la Dufaycolor & Autochrome
-
A
- Thread starter holmburgers
- Start date
Recent Classifieds
-
For Sale Fujica Super Fujica-6 folder, 6x6
- Started by Dan Daniel
-
Want to Buy WTB: "As Maine Goes" by John McKee
- Started by TheTrailTog
-
For Sale Zenza Bronica S2 Outfit - S2 Body, 75mm lens, 135mm lens, 120 film back
- Started by indy_kid
-
Sold Hasselblad 500 C Outfit - Body, 3 Finders, A24 Film Back, More!
- Started by indy_kid
Forum statistics
Assuming you used an older screen of about 1280x1024 and photographed it with 4x5", that's a peak resolution requirement of 1280*3/5=768 lines/in = 30 lines/mm. That is easily within the resolution of your film+lens. If it's a little unsharp, try photographing the screen with an enlarger lens; you should be able to make out the shape of each pixel and the masking in-between.
Try taking a photo of your LCD with C-41 as well as E-6... then you can make coarsely pixelated yet analogue RA-4 prints from your "chromes"!
Try taking a photo of your LCD with C-41 as well as E-6... then you can make coarsely pixelated yet analogue RA-4 prints from your "chromes"!
OP
OP
I mentioned the starch size in my first post, but kind of "in passing". It's .005-.010mm, or 5-10 microns.
And PE is right, there is a formula in the Friedman book. It's something like 1/600th of the distance to your eye that colors appear to merge. And 1/1000th of the distance to your eye for a black and white line pair to become indistinguishable, but apparently colors merge sooner. Sooooo, I guess if we're talking about holding a 4x5" at arms length or approximately 2-ft. (~600mm), that comes out to 1mm, right? Hmmm.... maybe there's a different formula.... this doesn't seem right.
Polygot, I'll trust your math
But where did you get 3/5? But if it's well w/in limits then that's promising! And an enlarger lens eh? I guess this follows the same logic of a process lens; getting as flat a field as possible.
I agree that 1mm seems pretty big.
Is that the largest size that won't be resolved as dots, therefore smaller is better. Or is that an optimum size such that smaller results in smearing the image?
Well, on an added note, the particles must be larger than the largest grain of emulsion IIRC. Otherwise, you have overlap of color in a single grain causing aliasing of colors in a manner similar to digital. Friedman also discusses this.
PE
PE
OP
OP
I'll try to reread tonight and post the pertinent info. He does indeed mention the importance of having enough developable grains behind each element.... so it's a balance really.
1280 pixels, multiplied by 3 elements (R, G, B) in a row for each logical pixel, divided by 5 inches on the long side of your film gives 1280 * 3 / 5 = 768 dots/inch. I'm assuming here that the LCD has strict columns of R, G, B, which they (for computer use) mostly do. Trinitron CRTs are like that but others use a hexagonal tiling of circular dots. Some are more like a Bayer mask (GRGB with alternating R/B in each row). But the three-column approach is by far the more common as it simplifies the wiring to the cells.
If the pixels are horizontal (finer pitch vertically), you have 1024 * 3 / 4 = 768 dots/inch again: same resolution required because the aspect ratios are the same. In the other direction of course you have one third of the linear resolution because each colour element has a 3:1 aspect ratio so that you get square pixels in the final output. LCD monitors not designed for computer use will generally NOT have square pixels, so the resolution calculations are a bit funkier; they should of course still work for your process. A quick google image search ("LCD macro") came up with this, which is fairly illustrative: LCD image.
Obviously if you use a higher resolution LCD (2560x1600 30" screen! you can fit so much code on one it's not funny) then you need proportionally better resolution from your film. But even that's only 4MP so you should have no problem with recording it on 4x5" Provia and a decent lens (process/enlarger type designed for close focus not infinity); your biggest problem at that resolution is going to be aligning the screen with each B&W image after processing.
Further thought to confuse issues: photograph your LCD with C-41 and make a whole pile of prints of it onto transparent RA-4 materials. Then you can use those as masks to make up a huge number of colour prints, just don't ever change your enlarger height! Result: using multi-layer colour photography to make spatially quantised single-layer colour photography: the worst of all worlds. Bonus: it's going to look digital due to the coarse pixelisation!
If the pixels are horizontal (finer pitch vertically), you have 1024 * 3 / 4 = 768 dots/inch again: same resolution required because the aspect ratios are the same. In the other direction of course you have one third of the linear resolution because each colour element has a 3:1 aspect ratio so that you get square pixels in the final output. LCD monitors not designed for computer use will generally NOT have square pixels, so the resolution calculations are a bit funkier; they should of course still work for your process. A quick google image search ("LCD macro") came up with this, which is fairly illustrative: LCD image.
Obviously if you use a higher resolution LCD (2560x1600 30" screen! you can fit so much code on one it's not funny) then you need proportionally better resolution from your film. But even that's only 4MP so you should have no problem with recording it on 4x5" Provia and a decent lens (process/enlarger type designed for close focus not infinity); your biggest problem at that resolution is going to be aligning the screen with each B&W image after processing.
Further thought to confuse issues: photograph your LCD with C-41 and make a whole pile of prints of it onto transparent RA-4 materials. Then you can use those as masks to make up a huge number of colour prints, just don't ever change your enlarger height! Result: using multi-layer colour photography to make spatially quantised single-layer colour photography: the worst of all worlds. Bonus: it's going to look digital due to the coarse pixelisation!
OP
OP
Hey David, thanks for posting but for some reason it didnt' work for me; says its not available.
Polygot, having a C-41 master and printing on transparent RA-4 sounds like a great idea, at least, for someone set up to do RA-4. These transparent materials exist?
And the digital look will be very interesting indeed.... I'm hoping that it'll turn your picture into basically something that resembles a TV image; definitely a bonus!
You mention that alignment will be difficult, but I hope that you are wrong. Currently, it remains un-confirmed, but I "pin-registered" the reseau to my b&w film. Actually, I just used a 2 hole punch, but it seemed to do a great job. Only the results will tell us for sure if it's good enough though.
Lastly, update to the Friedman book. He doesn't give a formula per se, other than the one I mentioned yesterday (1/600th of the distance to your eye for color fusion) which was actually E.J. Wall's disclosure. But he does give the conclusion of Mees & Pledge who found that for a geometrical screen the individual elements should be between 1/300th and 1/600th of an inch, or .08mm - .04mm. Irregular screens should be between 1/900th and 1/1200th of an inch, or .03mm to .02mm.
Interesting to note that the autochrome elements were on average .015mm or 1/1666th" (according to Friedman), but were intended for projection as well. Due to the number of grains that needed to be behind each element the autochrome plate was slow, exactly twice as slow as Dufaycolor.
Polygot, having a C-41 master and printing on transparent RA-4 sounds like a great idea, at least, for someone set up to do RA-4. These transparent materials exist?
And the digital look will be very interesting indeed.... I'm hoping that it'll turn your picture into basically something that resembles a TV image; definitely a bonus!
You mention that alignment will be difficult, but I hope that you are wrong. Currently, it remains un-confirmed, but I "pin-registered" the reseau to my b&w film. Actually, I just used a 2 hole punch, but it seemed to do a great job. Only the results will tell us for sure if it's good enough though.
Lastly, update to the Friedman book. He doesn't give a formula per se, other than the one I mentioned yesterday (1/600th of the distance to your eye for color fusion) which was actually E.J. Wall's disclosure. But he does give the conclusion of Mees & Pledge who found that for a geometrical screen the individual elements should be between 1/300th and 1/600th of an inch, or .08mm - .04mm. Irregular screens should be between 1/900th and 1/1200th of an inch, or .03mm to .02mm.
Interesting to note that the autochrome elements were on average .015mm or 1/1666th" (according to Friedman), but were intended for projection as well. Due to the number of grains that needed to be behind each element the autochrome plate was slow, exactly twice as slow as Dufaycolor.
I can't download it either. Google says:
Sorry, the page (or document) you have requested is not available.
Please check the address and try again.
I am logged in to my Google documents account just in case, so it's not a generic login problem.
Apparently the transparent RA-4 materials exist and are used in illuminated signs and the like. Availability though...
One wrinkle that you'll have to sort out by experimentation is the spectral response of the dyes in whatever mask you generate (E-6, RA-4, whatever). All we know is that, when illuminated with white light, that they form most of the gamut we can perceive - after all, they are designed for display purposes. When you use them as a recording mask though, you might get weird tonality in your image because (for example) a certain shade of green may not actually pass the green dye in your mask, and will therefore be darker in the recorded image. Note that in colour processes there are two spectra of importance: the recording spectrum and the playback spectrum; they can differ quite dramatically yet still produce what looks (to human eyes) like accurate colour, so when you use the playback spectrum as a recording spectrum, things may get interesting.
One wrinkle that you'll have to sort out by experimentation is the spectral response of the dyes in whatever mask you generate (E-6, RA-4, whatever). All we know is that, when illuminated with white light, that they form most of the gamut we can perceive - after all, they are designed for display purposes. When you use them as a recording mask though, you might get weird tonality in your image because (for example) a certain shade of green may not actually pass the green dye in your mask, and will therefore be darker in the recorded image. Note that in colour processes there are two spectra of importance: the recording spectrum and the playback spectrum; they can differ quite dramatically yet still produce what looks (to human eyes) like accurate colour, so when you use the playback spectrum as a recording spectrum, things may get interesting.
Take care. Some of these materials are not clear transparency materials, but rather have a frosted support.
PE
PE
OP
OP
This is something I hadn't considered too much, but I'm sure you're right.
So let me see if I understand... you're saying that if we take a picture of a green apple, due to the dye's absorption characteristics and the apple's reflection characteristics, less or more green than we expect will make it thru, or in other words there might be spectral peaks.(?)
Many of these questions will (hopefully) be answered when results arrive. In the meantime all we can do is go ahead and see what we get.
By the way, I can see David's screen now. Thanks for posting that!
OP
OP
Microscope images of CPU screen-plate (fluorescent backlighting)
Hurah!
I got an image back from my friend with the microscope.
As you can see, the old S-K Symmar 180mm did the trick and resolved the elements fairly well. It's hard to know the quality of the picture-taking device in the first place, so I can't comment on whether it might actually be sharper (this was her first lesson with this microscope at school).
As for color balance, at least this tells me which direction to go. My source color was a cream-yellow, to the eye, and that might account for the very weak blue. So if I start with a white source color and add some magenta as per Kodak's suggestion for dealing w/ fluorescents, it might reign this balance in afterall.
If you have any other suggestions for filtration let me know. As of now I have exactly zero CC filters. A warming filter might be the closest thing I have, but I'll try to scrounge up something.
Have a great weekend everyone!
Hurah!
I got an image back from my friend with the microscope.
As you can see, the old S-K Symmar 180mm did the trick and resolved the elements fairly well. It's hard to know the quality of the picture-taking device in the first place, so I can't comment on whether it might actually be sharper (this was her first lesson with this microscope at school).
As for color balance, at least this tells me which direction to go. My source color was a cream-yellow, to the eye, and that might account for the very weak blue. So if I start with a white source color and add some magenta as per Kodak's suggestion for dealing w/ fluorescents, it might reign this balance in afterall.
If you have any other suggestions for filtration let me know. As of now I have exactly zero CC filters. A warming filter might be the closest thing I have, but I'll try to scrounge up something.
Have a great weekend everyone!
Attachments
OP
OP
The Proof is in the Pudding
At long last!
Proof of concept, results, tangible evidence!
It's far from perfect, but the verdict is in, the judge has ruled and a sentence dealt... Louis Ducos du Hauron was right, the Lumiere brothers were not total crack pots, and Dufay has rotated 360° in his grave!
I kid, I kid... but I am excited to finally have something to show for my efforts. And although there is room for improvement, a color image has been formed!
I've posted two pictures. The black & white positive and the composite screen-plate photograph (b&w slide + TV-screen-plate).
As you can tell, the positive is underexposed. There is definitely more detail to be seen in person, but my scanner isn't quite able to peer through it. Add that density to the screen plate and you've got a really dark image; hence the nasty looking scan that has been boosted beyond reason. In real life the colors are much richer.
Now, I've learned quite a few things. For one, the color balance of the screen is the least of my worries at the moment, as registration is more difficult than expected. The pin-registers only get you in the ballpark, and the real aligning has to be done visually. This in itself is difficult because the two films don't lie flush. I overcame this by taping the positive to my light table and then taping the screen plate to a piece of glass. Pressed emulsion to emulsion, I could then line them up visually and maintain good contact. With part of the screen lying off the edge of the glass, I was able to tape the two together with painter's tape once I achieved "good-enough" registration.
Moving the screen in relation to the positive causes all sorts of rainbow-colored moiré patterns to dash across the image. It's very strange and beautiful! But as it is, it's impossible for me to get the whole image to look color-correct in one position. However, I can get each individual part to look right at the expense of another; i.e. the banana will look yellow but the Budweiser logo will be purple instead of red, the white apple will be tinged green, and so on.
I think the problem lies in the fact that there was space between the screen & the film during the exposure. Getting the two taped-together films to slide under the grooves of a 4x5" film holder was no easy task and undoubtedly the crude registration (by means of a strip of masking tape) was not sufficient to hold them perfectly together. This needs improving, because without good contact, registration of the two films is a complete crap shoot.
Secondly, the color balance really doesn't look too bad, and the example here is utilizing the screen-plate (réseau) derived from a flat-screen CRT television; a.k.a. the one that I posted above and that appeared quite blue. I illuminated the subject with an ordinary tungsten light bulb, and I think the bluishness of the screen actually balanced it quite nicely to white.
Considering how finicky the registration is, I can't imagine working with anything higher resolution than SDTV at this point. I think I'm going to stick with TV (in lieu of LCD computer monitors) for now. Plus, I think it looks pretty cool!
So for next time, I need to produce a much brighter black & white positive, and process it myself (*cough *cough, dr5 took over a month) while figuring out some solution for maintaining better registration. Ultimately I'll mount the two together permanently with Canada balsam or the like and build some kind of opal glass viewing frame, but something temporary must be done for the exposure registration.
Cheers!
At long last!
Proof of concept, results, tangible evidence!
It's far from perfect, but the verdict is in, the judge has ruled and a sentence dealt... Louis Ducos du Hauron was right, the Lumiere brothers were not total crack pots, and Dufay has rotated 360° in his grave!
I kid, I kid... but I am excited to finally have something to show for my efforts. And although there is room for improvement, a color image has been formed!
I've posted two pictures. The black & white positive and the composite screen-plate photograph (b&w slide + TV-screen-plate).
As you can tell, the positive is underexposed. There is definitely more detail to be seen in person, but my scanner isn't quite able to peer through it. Add that density to the screen plate and you've got a really dark image; hence the nasty looking scan that has been boosted beyond reason. In real life the colors are much richer.
Now, I've learned quite a few things. For one, the color balance of the screen is the least of my worries at the moment, as registration is more difficult than expected. The pin-registers only get you in the ballpark, and the real aligning has to be done visually. This in itself is difficult because the two films don't lie flush. I overcame this by taping the positive to my light table and then taping the screen plate to a piece of glass. Pressed emulsion to emulsion, I could then line them up visually and maintain good contact. With part of the screen lying off the edge of the glass, I was able to tape the two together with painter's tape once I achieved "good-enough" registration.
Moving the screen in relation to the positive causes all sorts of rainbow-colored moiré patterns to dash across the image. It's very strange and beautiful! But as it is, it's impossible for me to get the whole image to look color-correct in one position. However, I can get each individual part to look right at the expense of another; i.e. the banana will look yellow but the Budweiser logo will be purple instead of red, the white apple will be tinged green, and so on.
I think the problem lies in the fact that there was space between the screen & the film during the exposure. Getting the two taped-together films to slide under the grooves of a 4x5" film holder was no easy task and undoubtedly the crude registration (by means of a strip of masking tape) was not sufficient to hold them perfectly together. This needs improving, because without good contact, registration of the two films is a complete crap shoot.
Secondly, the color balance really doesn't look too bad, and the example here is utilizing the screen-plate (réseau) derived from a flat-screen CRT television; a.k.a. the one that I posted above and that appeared quite blue. I illuminated the subject with an ordinary tungsten light bulb, and I think the bluishness of the screen actually balanced it quite nicely to white.
Considering how finicky the registration is, I can't imagine working with anything higher resolution than SDTV at this point. I think I'm going to stick with TV (in lieu of LCD computer monitors) for now. Plus, I think it looks pretty cool!
So for next time, I need to produce a much brighter black & white positive, and process it myself (*cough *cough, dr5 took over a month) while figuring out some solution for maintaining better registration. Ultimately I'll mount the two together permanently with Canada balsam or the like and build some kind of opal glass viewing frame, but something temporary must be done for the exposure registration.
Cheers!
Attachments
Congratulations!
You may be having a problem with the swell of the support. Estar support was used to remove this problem in Dye Transfer, otherwise you would get color fringing. On acetate support the change in dimension with processing can cause such a problem with screen materials. Back in the early days, the integral screens were used to avoid such problems.
PE
You may be having a problem with the swell of the support. Estar support was used to remove this problem in Dye Transfer, otherwise you would get color fringing. On acetate support the change in dimension with processing can cause such a problem with screen materials. Back in the early days, the integral screens were used to avoid such problems.
PE
OP
OP
How much of a dimensional change can occur on acetate supports and what can be done, if anything, to mitigate it? Where is the estar used in the DT process; on the matrix films?, or the separation negatives? Are there any films still coated on estar these days that I could use, either for the screen or b&w positive?
I'd like to think that in the present example this is not my problem, considering the resolution of the screen being fairly coarse and also the aforementioned problems with jamming the film into the holder. Plus, this is something I have no ability to remedy, so thinking about it is depressing!
But, only further testing and refinement will tell.
I'd like to think that in the present example this is not my problem, considering the resolution of the screen being fairly coarse and also the aforementioned problems with jamming the film into the holder. Plus, this is something I have no ability to remedy, so thinking about it is depressing!
But, only further testing and refinement will tell.
Right On Chris!
I'll keep watching to see how far you can take it. Looks like you should be able to get some rather unique images with a character all their own.
I'll keep watching to see how far you can take it. Looks like you should be able to get some rather unique images with a character all their own.
Estar or Melenex or this material under a number of names is used on all films in the Dye Transfer production sequence to minimize size changes. The change is enough to cause a "moire" or "rainbow" pattern on photos which experience such changes.
All sheet films from Kodak are on Estar IIRC.
PE
OP
OP
Oh, well that's reassuring. So then, no more Ilford Delta 100 perhaps...
R Shaffer; thanks! That's what I'm hoping for and I appreciate the encouragement, it makes it all the more worthwhile.
- Joined
- Mar 2, 2007
- Messages
- 1,464
- Format
- Medium Format
If you wanted to spend the money there's a new breed of LCD moniter out that uses white LEDs as a light source, and are brighter with a higher contrast ratio.
Either way you should be able to color balance the monitor screen to any color you want in spite of the light source.
OP
OP
Stephen, that result is stunning. What are you using for the screen? Is that 4x5"? Definitely of interest to me and others I'm sure! I'd love to hear your method.
As for fine-detail, I can't speak authoritatively as I've still only got one example under my belt. It does stand to reason that there will be a loss of fine detail, due to the fact that a pure green object (for instance) would only be using 1/3 or less of the screen in that area.
glbeas, that monitor sounds awesome, but I'm in no position to buy one at the moment. At some point I am going to shoot a roll of 36 with different filtration to see what comes out the most grey, and then use the best for my 4x5" screens.
As for fine-detail, I can't speak authoritatively as I've still only got one example under my belt. It does stand to reason that there will be a loss of fine detail, due to the fact that a pure green object (for instance) would only be using 1/3 or less of the screen in that area.
glbeas, that monitor sounds awesome, but I'm in no position to buy one at the moment. At some point I am going to shoot a roll of 36 with different filtration to see what comes out the most grey, and then use the best for my 4x5" screens.
I'm going to try dye some starch and mix it into a gelatin emulsion and paint it on top of some foma, I dont think it'll hold during processing, but I can re-apply a screen after scanning (or painstakingly on some RA-4 when I get an enlarger).
| Photrio.com contains affiliate links to products. We may receive a commission for purchases made through these links. To read our full affiliate disclosure statement please click Here. |
PHOTRIO PARTNERS EQUALLY FUNDING OUR COMMUNITY:  |