'hybrid' digi negs

[Colin Graham]

Believe me, I agree! it really doesn't make much sense to me either. Maybe I'm setting up QTR differently? But the results are hard to argue with. I'll grab a screen shot of my latest profile, maybe it might help to clear something up. All I can say is one thing plus another thing sometimes makes a completely different thing- anyway that's my understanding of the colorized negative processes.

Here's the best combo so far. 70 limit alone for the yellow ink results in a 25%-30% gray in the 0 step without the cyan bump...With the 15 cyan limit addition I had to bump both the exposure and the sensitizer strength and still can't print through the 0 step. I don't have a densitometer so can't be anymore specific. Maybe there's something obvious I'm doing wrong.

But there are odder things at play today- anyone else watching the US Open semi between Federer and Djokovic!?

 

[sanking]

Ron,

Thanks for the thoughts.

But any thoughts on converting a RGB value of G=255, R=110 to a QTR profile. The thing is that after over a year I finally figured out a color negative that prints very smooth with carbon and I don't want to start over again and wind up testing another two tons (your words!!) of Pictorico.

BTW, that RGB color gives me very high UV blocking value, a little bit over log 3.0. And my understanding is that black is not used in the PDN colors. Could be wrong but someone explained it to me that way.

Sandy

 

[Ron-san]

Believe me, I agree! it really doesn't make much sense to me either. Maybe I'm setting up QTR differently? But the results are hard to argue with. I'll grab a screen shot of my latest profile, maybe it might help to clear something up.

Colin-- If I read your profile correctly you have set up a 2 part gray system with yellow as the 1st gray part and cyan set to follow the Y curve. I assume LK is the 2nd gray part and the other light inks are set to follow LK. But I am just assuming that since I cannot tell from the screen grab if indeed it is a two part gray system or some other set up. Anyway, I still cannot understand how adding a bit of weak cyan to the yellow has such a profound effect on the negative density. It might be fun if you were to drop by Mercer Island some day and we work on this together. Cheers, Ron Reeder

 

[Colin Graham]

Here's the txt file if that's any help.

That would be great to get your help with this. I will certainly try to contact you the next time I'm in the area. Many thanks for your input here.

 

[Ron-san]

Thanks for the textfile, Colin. I will play with it a bit and get back to you. Cheers, Ron Reeder

 

[Ben Altman]

Ron's results sound a bit different from what I get from the same experiment. I suspect that the blocking of inks in UV processes vary a bit with light sources and maybe the glass in the vacuum frame. Also I think ink settling in the cartridge can cause changes, so a new cartridge may look like a different formulation just because it has been shaken up recently. So I doubt that it's possible to develop magic bullet ink profiles - they have to be built to match the set-up you are using and checked from time to time.

Also things can change from one type of OHP to another - including which inks can be used at high volume.

Ben

 

[Colin Graham]

I'll probably embarrass myself more here but I'm used to that.

About the minor additions to 'prime' QTR inks- It's my impression when contact printing an HSL array that slight differences in hue angle can result in vastly different qualities of blocking color. Would it not then stand to reason that the same would be true when printing slight color differences from within QTR? This is sort of what this whole thread hinges on, and would like to clear that up if I have it wrong.

Additive color is just that, or do I have the theory backwards? In my tests, the blocking didn't seem to hang on what green it was, just that it was green. In otherwords, the yellower green seem roughly as effective at blocking UV as the more cyan ones, where as the plain yellow- which is supposed to be the prime blocker- didn't work well at all.

 

[Colin Graham]

Ron's results sound a bit different from what I get from the same experiment. I suspect that the blocking of inks in UV processes vary a bit with light sources and maybe the glass in the vacuum frame. Also I think ink settling in the cartridge can cause changes, so a new cartridge may look like a different formulation just because it has been shaken up recently. So I doubt that it's possible to develop magic bullet ink profiles - they have to be built to match the set-up you are using and checked from time to time.

Also things can change from one type of OHP to another - including which inks can be used at high volume.

Ben

My experience bears this out as well. I once had to RMA a 3800 and pretty much had to start over in a lot of ways when the new printer showed up.

 

[Colin Graham]

Here's the green next to the plain yellow. None of the profiles use dark black ink and all use the same light ink mix for the shadows. None are curved. Please forgive the spots & blotches- the first printing of the strip was too soon and some of the ink stuck to the mylar interleaf, so all the subsequent test exposures look pretty rough. Notice how much more effective C20 is from C15, both using same yellow, but they are much closer together than the yellows....even the 75 limit yellow is actually probably closer to a 50% gray in the 0 step. The white dots are a key so I don't mix the prints up from the negatives.

1. cyan 20 yellow50
2. C20 Y60
3. C15 Y60
4. Y60
5. Y75

 

[Ron-san]

But any thoughts on converting a RGB value of G=255, R=110 to a QTR profile?
Sandy

Sandy-- I have thought about that a bit and at present have not a clue. Maybe some smart person reading this will tell us how to do it.

I take it that you finally have, by whatever method, a blocking color that gives satisfactory results on carbon? If so, that is very good news whatever the method since carbon is perhaps one of the most demanding of the alt processes. I would like to be able to quote you on that when promoting digital negatives to the masses.

Cheers, Ron Reeder

 

[Colin Graham]

Here a comparison of green and red, same general densities, but the magenta addditons are much more than homeopathic, without adding nearly as much blocking. The magenta ink also take much much longer to dry, risking pizza wheels. Again, no K inks, same light ink set for the shadows and no curve applied.

1. cyan20 y60
2. c15 y70
3. >magenta 30 y60
4. M70 Y40
5. >back to cyan15 Y65

 

[sanking]

Ron,

OK, please let me know if you come up with any good ideas on the RGB to QTR profile.

And yes, the RGB color combo I mentioned works great with carbon, and you can quote me on that.

Sandy

 

[Ron-san]

Friends-- I get the feeling from reading this thread that some (many?) of you feel that somehow adding cyan to yellow to get green results in a magical color that is more than the sum of its parts. To try and test this I did the following experiment:

I wrote a QTR profile that caused a step tablet to be printed with yellow ink only, from 0 to 100%. (I set up a one part system with Y ink as the first and only part and set the Y ink limit arbitrarily at 70%). This resulted in a normal looking step tablet, printed on Pictorico, except that it was entirely in yellow.

Next, I wrote another QTR profile and printed the same step tablet but this time in only cyan ink, setting the ink limit to 15%.

Then I wrote a third profile in which the Y ink limit was set to 70% and the Cyan ink was set to copy the Y ink distribution but was set to a limit of 15%.

In other words, I printed out one tablet only in yellow, a second only in cyan, and a third in a combination of exactly the same two inks (it came out green, of course).

Then, I used my Xrite 316T to read the UV optical density of selected steps in all three tablets. What I expected was that the density a step in the green tablet (Y plus C) would equal the combined densities of the same steps in the Y and C tablets.

What I found was that they did, more or less. The 10% and 30% steps in the green tablet were exactly the sum of the same steps in the Y and C tablets. In other steps the green tablet was denser by 5-10% than the sum of the Y and C steps. I cannot explain why there was an apparent 5-10% increase in some steps, but I think this anomalie does not invalidate the conclusion.

If you mix cyan with yellow you get pretty much the sum of the densities of the two individual components. There is no magical amplification of density when yellow and cyan are combined to form green.

Thus, there is still order and reason in the universe. I still do not know how to explain Colin's results.

Cheers, Ron Reeder

 

[Colin Graham]

Well, I wouldn't call it a magical color either, just a starting point. I did mention I just started experimenting with this. When printing a HSL array and contact printing it with my processes, the green hues were generally the smoothest rendered and cleanest to final print, so that's where I'm starting. I don't mean to affix any mythic properties to the color, but so far it seems like exploring further.

Why would QTR be excluded from producing the same general hues of effective blocking colors? Surely not because it's a B&W rip, hehe. The HSL array and the QTR profiles were printed from the same printer. The fact is that green is a good blocking color, it's not a myth, and it's plain from looking at both the HSL output and the file in photoshop that there is minimal black (if any at all) in the sweet spot of the hue angle. I don't understand how else to get green except add cyan and yellow? Surely my results support my reasoning here, or I've really gone round the bend.

 

[Ben Altman]

Then, I used my Xrite 316T to read the UV optical density of selected steps in all three tablets. What I expected was that the density a step in the green tablet (Y plus C) would equal the combined densities of the same steps in the Y and C tablets.
Cheers, Ron Reeder

Hi Ron,

As the densities we read are negative logs of the actual density, adding them together is not totally straightforward, I believe - adding logs is equivalent to multiplying the underlying numbers. So I think the sum of the densities of your two inks should be -log(inverse log dY + inverse log dC). But it's a long time since I was in college, so if anyone has a better grasp of the theory, I'd be delighted to hear it...
Might be an interesting comparison to lay your Y and C step tablets on top of each other and measure that density, subtracting out the density of the second layer of substrate, again by the above method.
Best, Ben

 

[Colin Graham]

I'll spare everyone from further updates, but if anyone does have a question please don't hesitate. I'm pretty excited about the results I'm getting and hope it's not a matter of having a magic cyan ink cartridge- because it's starting to run a little low, hehe.

If I do discover an anomaly with the printer, QTR setup, or the carts I'll be sure to update the thread. And if anyone else sees anything peculiar in my profile or reasoning, I encourage them to sing out as well.

 

[Ron-san]

Hi Ron,

As the densities we read are negative logs of the actual density, adding them together is not totally straightforward, I believe
Best, Ben

Ben-- Actually, I think it is straightforward. You just add the densities together. And when I put the cyan step tablet on top of the yellow step tablet, the densitometer gave the same density reading as when I printed both the cyan and yellow inks on the same tablet. No need to make things more complicated than they already are. Cheers, Ron Reeder

 

[Colin Graham]

Sorry to nag you Ron but I do still have this big question: Green is a great UV blocker as determined with RNP and as Sandy attests with PND as well. The best shades are by most accounts free of black ink. How do you think that PS/Epson driver method might be mixing green that QTR cannot?

I just don't see how green can be 'magical' when achieved with QTR on one hand, and on the other hand it's news worth telling all your friends interested in digital negatives about with PDN. There seems to be a fundamental conflict here, but maybe I've misunderstood something along the way. My main question of this thread, really the whole point of the thread, is getting two separate answers here:

Sandy-- I have thought about that a bit and at present have not a clue. Maybe some smart person reading this will tell us how to do it.

I take it that you finally have, by whatever method, a blocking color that gives satisfactory results on carbon? If so, that is very good news whatever the method since carbon is perhaps one of the most demanding of the alt processes. I would like to be able to quote you on that when promoting digital negatives to the masses.

Cheers, Ron Reeder

Friends-- I get the feeling from reading this thread that some (many?) of you feel that somehow adding cyan to yellow to get green results in a magical color that is more than the sum of its parts.

Cheers, Ron Reeder

My apologies if I've misunderstood. Any clarification would be most appreciated.

 

[Ben Altman]

Ron,


But any thoughts on converting a RGB value of G=255, R=110 to a QTR profile.
Sandy

Hi Sandy - I thought about this a little. It might be do-able, but not sure. As you can't see inside the Epson driver, you can only guess and experiment.

Using a simple-minded approach, you could separate the G and R and work on them separately.
Step 1: Print a patch of G 255 and another of R 110 through the Epson driver.
Step 2: Write a QTR profile that prints variations on 50% Y + 50% C and another that does a range of variations around say 23% M + 23% Y.
Step 3: Measure all the above with a spectrophotometer and find the closest matches for the Epson G and R.

One problem is of course that even if you hit the right combo at 100%, there's no way to know how the Epson driver blends in LC and LM, so the less dense parts of a QTR profile might not match the PDN profile you are using. That would probably be manageable with your curve, though, I'd think.

Best, Ben

 

[Colin Graham]

Ben, you think most of the inks are getting added to the mix somewhere along the way with PDN? Interesting that a cmyk sample of the HSB color wouldn't include shades other than Y and C in it's makeup- but I'm likely making way too much of that. I had always though it that as more or less the instructions that get sent to the driver. Clearly there's more to it that that- but adding more ink colors just seems like you'd run into gamut problems fairly quick with some of these garish negative colors. From my experience with gum printing if you want a screeching color like this 110R 255G (76% C 100%Y) you don't muddy the waters with magenta and grays. And as the extra ink cart inclusions would not improve UV blocking according to the densitometric (sic?) evidence presented, I'm not sure what the benefit would be unless one of those carts was MK or PK. I would be really surprised if Sandy's color has black in it.

But now I'm curious about PDN- is it printed from the epson driver directly, or using PS to manage the colors? When I used RNP I tested many combination of rendering intent and driver combinations and seem to remember that with an HP printer PS handles the array best and gave the smoothest density and blocking colors. Printing out tests strips with PS management against the epson driver that now seems reversed, especially since you are able to tweak the ink density in the paper setup of the epson interface. So I'm even more curious now about the driver's interpretations of colors against their CMYK 'dropper' representations in PS. I swear though- louping a 'pure' green it's difficult to see anything other than cyan and yellow in these daylgo greens printed through PS/ Epson's driver

 

[sanking]

I'm curious about PDN- is it printed from the epson driver directly, or using PS to manage the colors? When I used RNP I tested many combination of rendering intent and driver combinations and seem to remember that with an HP printer PS handles the array best and gave the smoothest density and blocking colors. Printing out tests strips with PS management against the epson driver that now seems reversed, especially since you are able to tweak the ink density in the paper setup of the epson interface. So I'm even more curious now about the driver's interpretations of colors against their CMYK 'dropper' representations in PS. I swear though- louping a 'pure' green it's difficult to see anything other than cyan and yellow in these daylgo greens printed through PS/ Epson's driver.

Colin,

When you print with PDN you are using the Epson driver directly. There is nothing in PDN that controls the printer in any way.

My understanding of PDN when printing with the RGB values is that you are *not* mixing in any black when printing. I don't know this for sure but some people who understand the Epson driver very well assured me that this is the case. Course, that discussion was made in reference to one of the older Epson printers so it may not apply to all Epson sprinter drivers.

Another important issue is whether the color you choose can print all 255 tones. It is my understanding that all of the PDN colors will do this but that when you mix in black this may not be the case. I experimented some with an RNP array and found that some colors did not print 255 tones. Whether or not this makes any difference in practice I don't know.

So my thinking is that in order to create a QTR profile from my PDN color of G=255, R=110 it would not be necessary to use any black. But if you needed more blocking density the only way to get it with PDN is to either 1) add some black, or 2) use a +Ink density setting. The latter approach is the easiest way and does not present a problem for either Pictorico or Inkpress OHP so long as you don't exceed +30. If you need more Pictorico Ultra will handle up to about +50.

Sandy

 

[Colin Graham]

Thanks Sandy, I appreciate the clarification. That is in line with my experiences with RNP- the smoothest colors had no black in the mix.

I wonder where a 100% limit in QTR falls in with the maximum ink deposits of the epson driver? I was getting some fairly heavy pooling with the high magenta limits yesterday, but the yellow either dries much quicker or is clipped somewhere in QTR because it came out of the printer quite dry.

 

[Ben Altman]

From my experience with gum printing if you want a screeching color like this 110R 255G (76% C 100%Y) you don't muddy the waters with magenta

Yes, Colin, thanks, I wasn't thinking clearly - some of the Cyan and Yellow balances out the Magenta so could probably just work with one "diagnostic" QTR profile that does variations (doing some quick mental arithmetic) around 39% C + 61% Y.

With respect to the 100% limit in QTR, I've always assumed that it's about the same as a normal max ink output for the Epson driver, though that would be easy to test by printing test squares from each. It can be kind of irrelevant though, as when combining inks the total ink load the substrate can hold is what determines the amount of ink one can use - with several inks mixed it doesn't take much of each to reach the puddle point.
Measurement of the QTR ink separation print shows that with the QTR dither at least, the UV blocking densities shoulder off at around 60%, so I try not to use more than about 50% of any one ink.
I'm not convinced that black ink is a problem, as long as it is used in proportionate amounts. I think it does absorb into OHP less readily, so using black could lower the max you can lay down without puddling. For that reason I taper off some of the PK and replace with C and Y for the densest couple of steps on my negs. Thus my 100% step probably looks not unlike Sandy's but with a bit of PK and LK added.

Best, Ben

 

[R Shaffer]

I would agree with Ben, that I don't think the green is necessarily just cyan+yellow. The printer has a look up table ( LUT )

www.factor.bg/epson/SP9900/Maximum_colour_fidelity.pdf

that determines what inks to combine to create a particular color.

I would try printing a block of the green you are after on plain old epson glossy paper. Then print a couple of blocks of guestimates of QTR combinations and see how close you get. A spectrometer would probably be a big help, but eyeballs should work.

But I like your approach and your early trials look like they have potential.

 

[Colin Graham]

Thanks Rob, I sort of started with the eyeball method. I'm a little embarrassed to admit just how much so it was. I printed the best blocking color I could ever get on my 3800 in with different printer management and intents and just stared blending the dropper cmyk amounts to get a rough visual match in QTR. Pure guesswork, but they can be confirmed or rejected easily enough by a quick test exposure. Clearly not a easy as just contact printing a HSL array, but I think it's a reasonable secondary step after the array and got me in the right neighborhood fairly quickly.

Thank you Ben, that does sound like a good practical limit. The samples I posted with Yellow above 60 are likely on the verge of the pizza wheel problem, even with minimal cyan in the mix.

I just want to stress again that I don't really think black inks are necessarily a problem, or should be avoided at all costs. Except with a very picky process on very picky surfaces, I've have nice results with a full compliment of inks. I can't say that enough.

I was just recently frustrated with a certain highkey QTR negative, remembered how well colorized negatives used to handle highlights, then also remembered that most if all the best colors I could come up with didn't use black, so thought I would try that to see if it helped in this situation.

I'd use colorized method exclusively but I personally prefer the way QTR handles shadows and midtones, and I also vastly prefer the QTR interface to the PS/epson driver route (I'm on a PC- macs may be less likely to reset to defaults with every single print command.) So I thought it would be cool to blend what I like about each in one. Again these are just personal preferences.