No, I have not yet given up! In fact, I’ve been making some progress on the color carbon transfer front. Last time, I made some cyan, magenta and yellow tissue with my newly acquired paints. Only five 4×5″ sheets per color, which is ample for some initial testing. This testing is underway, results so far are promising, and there’s even some progress on the digital negatives front!
First, those tissues. Of course, they’re difficult to compare, because I used a different brand and/or type of paint for each color. But my initial goal with this test was to look for blatant differences between them. Turns out there aren’t any…well, there’s bound to be small-ish differences in pigment load, but it’s not the landslide outcome I had expected. I made some glop with 2% (w/v) Windsor & Newton watercolor cyan (PB15), Van Gogh/Talens watercolor yellow (PY128) and Talens gouache magenta (PV19). Pigment load appears to be in the same ballpark across all three – but this is based on highly subjective eyeballing.
More importantly, none of the tissues appear to present any odd problems in this application. Basically, they just work. So that gives me something to choose from…I’ll settle on one brand and type of paint for further testing and I think I’m going to get some Windsor & Newton watercolors and the Talens gouache. The Van Gogh watercolor doesn’t seem to add much, and this way I’ll still have two horses in the race in case an unforeseen problem pops up.
And there’s some good news about the inkjet digital negatives. That is to say, I seem to have found a workaround for at least some of the issues, most notably the reticulation in the ink layer (scroll down on that link, it’s in there somewhere…). This problem made good contrast impossible, and the actual reticulation showed up on the prints if you examined them closely. Yuck. Ugh. Impossible. But, do not despair! Have a look at this:
Well, sometimes there’s an easy fix, or at least workaround: just overlay two identical negatives and Bob’s your uncle! With these small 4×5″ negatives, it’s easy to keep them in registration with some tape. Yeah, this project is going to be held together by tape, by the looks of it. I don’t really consider this as a good solution; if needs be, it can work, at least that’s what some initial testing leads me to believe. But I’d much rather have to print each negative only once. Still, as I’m waiting for some better (hopefully) transparency film to arrive, I can at least work on other fronts.
If you look at the image above, you might notice two things. First thing: I did a really stupid messup on the Y, M and C negatives. Having never made a color separation in my life, at least not that I can remember, I had a go with GIMP. You may remember from previous installments that I pretty much loathe GIMP, but it’s also kind of a love-hate relationship. We’re still together, you see. Anyway, turns out GIMP can actually make CMYK separations, at least in fairly recent versions of the software. It was buggy before, but it works now.
What I did not know, was that it actually creates separation negatives for all four layers. I did spot it for the black/key layer, but I overlooked the fact that C, M and Y are also negatives. So I inverted them (turning them back into positives) and printed away…of course I only realized my mistake just now as the carbon transfers I printed from them are drying. Live and learn!
The second thing you might notice is the image itself. It’s a new torture test I made for the whole color thing. I took inspiration from the nice patchwork quilts I spotted in Calvin Grier’s work; no doubt he does far smarter things with them than I do – for one thing, I don’t have any color profiling hardware except my own retinas. Not to worry; my aim is to make decent prints that sort of stand up to my own critical eye. They don’t have to measure 100% correctly. Maybe one day, but certainly not for the short term.
Anyway, I thought it would be good to have a pattern with some gradients across the primaries, but also a collection of various hues that are a little less obvious, but that may still prove important. Here’s what I came up with:
Working from left to right, we find…:
- There’s the obvious greyscale gradients, which I repeat on both ends, inverted, making it easier to spot bottom-top unevenness (e.g. a light source that does not quite cover the print area).
- There’s your C, M and Y and also R, G and B scales. I’m going to change those to range from white to full saturation instead of the way they are now, as they now confuse testing too much by involving the key/black layer too much.
- Then there’s a strip of skin tones pretty much spread across the entire spectrum of human diversity.
- Next up are the sky tones, including some sunset-inspired hues.
- Then there’s two columns of ‘earth’ hues which covers things like sands and stuff. Basically dirt, I guess.
- This is followed by a column of foliage greens and related hues.
- Then some more dirt, but in this case two columns of rock and stone hues; lots of low-saturation greys and browns and stuff.
- Then I included some palettes centered around one main color each: a column of reds, greens, blues, yellows, magentas and cyans.
- Finally I included a rainbow, because rainbows are cool.
I honestly don’t expect this to ever print decently. I mean, I couldn’t get this to print really very well on inkjet, let alone I’m going to pull it off with carbon. But at least it’ll allow me to spot major systematic problems, and hopefully solve the most glaring ones. And do opportunistic stuff, like decide which kind of image to print with which kind of tissue/pigment to avoid the weaknesses and exploit the strengths. All considered, it seemed like a useful thing to have around. The patches are 5x5mm when printed at 4×5″.
Of course, I had that silly little mishap with the inverted C, M and Y on the very first color separation I ever made. So now I’m stuck with this:
A very colorful kind of QR code…well, at least many things seem to work, but I’ll have to postpone the first ever color evaluation until tomorrow or so when I have the chance to reprint the whole thing. Then comes the whole business of figuring out the best layer order, see if the pigments are sufficiently transparent (this is probably going to be a bit of an issue) and many, many more issues, no doubt…