Am currently doing high quality digitizing of hundreds of photos, collected from dozens of sources, for a forthcoming book. These are everything from monochrome prints from glass plate negatives (1910 era) to digital images taken last week. Among these photos are some taken in 1950 on either Ektachrome or Anscochrome film--the mount is plain paper, and it is more than my life is worth to dismount them! Not surprisingly, they have developed a very strong red cast (really a lot more than a cast!). Red snow, red pavement,light pinkish sky, etc.. Unfortunately, or maybe fortunately, the subject matter is a basically one of those colours between orange and red, or at least I think that is what it is supposed to be--I do not have a proper reference for it. BTW masking of the sky is pretty much out of the question -- streetcar trolley wires are a critical component of those photos. Another problem, no idea if this is from the camera or if that is the way Anscochrome and/or Ektachrome deteriorate, is that they are heavily vignetted. The main subject matter is fine, but progressively into the four corners they get darker by as much as two or more f-stops I would think. Some heavy duty POM with an oval gradient fill sort of looked after that problem; however the colour breaks up very badly if the full amount required of lightness curve is applied, so the best correction so far still leaves the equivalent of about half an f-stop vignette. I have managed to get the red out of the snow, and the overall colour cast pretty much out of it. The darker portions -- near blacks and either pavement or gravel, still come out reddish brown, not brown or wet asphalt colour. And the very light neutral tones (overcast skies, etc.) have a barely perceptible green tinge now (passable). Most of the correction was done using all the curves in L*a*b. So, finally the question: has anyone developed a general procedure for dealing with late 1940s through early 1950s Ektachrome and/or Anscochrome? Would better results be attained through channel mixer perhaps? So far I have only encountered three slides, but as there are several hundred items I have not even looked at yet, and apparently hundreds more that I have not even received yet, I have no idea how many surprises await me!
EDIT: slight possibility those slides could be copy-slides on who knows what film; if so, copied prior to 1975, and the problems could either be the copy process, the source, or the copy film (or all!)
(BTW the book probably will not contain even a tenth of the material submitted; the publisher likes lots to choose from. But at least 50% of the material will be before 1950; coffee table history books are such fun!)
Hunter said:hmm I wouldn't think they were Ektachrome as that film stock tended to lean more towards the greens and blues... Kodachrome was always warmer IMO.
Hi Hunter. Since posting I have had some reasonable success with the Channel Mixer lens,. But painting (grey scale) on the lens to reduce the impact on the neutral colours is an adventure -- you actually see some weird collection of pixels that are best described as looking like a text pattern on TV as you paint. I have found that they are copies of March 18, 1950 slides, the copies were made about 1970 or before. So it's hard to say whether the originals or the copies have had the problem of the huge colour shift.
I have attached an almost finished, but low res JPEG of the "after" version along with the "before," the "after" having been done with the channel mixer lens, along with a contrast enhancement/gamma lens. A lot of finicky work left to get the channel mixer lens trimmed around the post and a couple of other places--unfortunately the publisher has said that YES! -- he wants that one in finished form, for the cover, no less. Not sure if it is good enough to enlarge to the cover (the working copy is well over 300 ppi, closer to 400 ppi at the size he would use for the cover, so pixels not a problem, but grain/sharpness of original may preclude use at 8 inch or more width, In any case I have to find a solution to the problems of trying to "paint" a pseudo mask on the channel mixer, to say nothing of ultimately getting it into CMYK. (The author just told me he saw large display prints from those in the early 1970s and they were awful! So the problems date back at least that far.)
(I wish the images in this forum would display using NNTP access -- I know you hate "forum" access instead of NNTP even moe than I do!)
I'm a little late to the game, and not to start a big "which is better" brouhaha, but I took you red version into PhotoShop and used;
Image > Adjustments > Auto Color.
This was the result:
Sometimes being forced to work in more than one platform can be good.
This took about 30 seconds in PP Image Adjustment Lab:
Clicked Auto adjust, then applied these settings:
OldBob... beautiful... looks great and exactly how I would wanna see it...
Roy... just ok... too "pink" for my liking... I would expect PP could get closer to the PS version though without a whole lot more work.
I just haven't tried it yet.
the 'dd'
OK... so I just tried for about 5 min using PP X4... still a bit "pinky" for my liking on a little LCD SAMSUNG 150s... and it certainly took more than just an auto adjust on the "red" original...
and I just noticed on my CRT screens it looks too greenish blue... so I'd have to calibrate my monitors on this one.
Most of my work along these lines isn't that critical so if I got close on an LCD screen, I'd be happy.