The QTR CMYK UV issue

[Colin Graham]

Trying to get my head around QTR, so please bear with me. I recently replaced a 3800 because of grain and banding, and while the banding has been fixed with the new one, grain has not- at least with QTR.

I'm getting perfect values in the shadows and midtones, but cannot get values above 30% to unblock without getting very pronounced grain. It would seem that with the 'dark' ink mix of CMYK controlling the highlights, the UV blocking properties of the 4 inks would need to be extremely similar, otherwise grain would be the result. Certainly the dither matrix is easy to see with yellow and cyan in with black ink. In other words if the UV blocking properties of each ink weren't the same, the dither pattern visible in the negative would be in the print. Am I missing something? It seems like that with 4 dark inks controlling the highlights, you would have to curve each ink separately, not just follow the K curve. Right?

Or am I being a tedious simpleton? I'm getting great gradients with green negatives and no apparent grain. I must be doing something wrong with QTR. It literally looks as if the inks aren't atomizing enough off the nozzles, and the dots of each dark CMYK ink each are plainly visible. Is this normal? Enclosed is a scan of the 'transition' to grain.

Maybe I'll try overlaying the neg on the print and try to see what inks aren't pulling their weight (so to speak) and turn those off, or at least scale them back. Or crank them up.

 

[keithwms]

Colin, your analysis makes some sense to me, though I profess no knowledge in this area.

I can offer to do some quick spectrometer scans on test patches if you wish. If you send little pictorico (or whatever) scraps with patches of each ink, I can give you absorption as a function of wavelength. Then, indeed, I suspect you will see that some inks absorb more at particular UV wavelengths than others, and you need to adjust the curve for each ink so they at least approximately match absorbance at the wavelength or wavelength range of your lamp. What lamp are you using? If it is a narrow-band lamp (mercury or such) then my guess is that these differences will be more pronounced.

If I had to guess which inks are or aren't "pulling their weight," I'd of course guess that the bluer inks are less UV transmissive than the redder inks. So then presumably you'd scale back the bluer inks until they match the redder. Might make no sense for your inkset though, if its CMYK.

 

[Ron-san]

It seems like that with 4 dark inks controlling the highlights, you would have to curve each ink separately, not just follow the K curve. Right?

I'm getting great gradients with green negatives and no apparent grain. I must be doing something wrong with QTR.

Colin--
I agree with you that in a perfect universe we might get smoother tones if all four of the dark inks (C, M, Y, and mK) had the same UV absorbance, the better to fill in the holes around each other's dots. But to accomplish that you would have to mix your own inks and devote a printer solely to making digital nagatives. The QTR profile I assume you are using is a compromise with the real world. It allows you to make very good negatives with the normal Epson inkset while retaining the ability to switch quickly to the Epson driver and print positive color or BW prints.

In the current version2 of my QTR manual, I suggest keeping all the dark inks at the same ink limit, and all the light inks at a different, single ink limit. This is done for convenience, because it makes writiing QTR profiles somewhat easier, and because I don't think it materially affects negative quality. However, if you want to play around with trying to get all the dark inks to have similar blocking power, here is the info you need. mK, Y, C and M inks (for the 3800) have relative UV blocking ability in the order 100, 60, 25, 12. In other words, for the Y ink to have the same blocking ability as the mK ink, the Y ink must be set to a limit 1.66 times the mK ink limit. C should be 4 times the mK limit and M should be about 8 times the mK limit. I tried that once and got a negative that was dripping wet with ink. When I lowered total ink limits, but kept them in the same ratio, I could not get enough overall UV blocking to print on palladium. Bottom line, you could equalize the mK and Y inks, and maybe the C ink, but forget aboout the M ink. I have not done the experiment, but negatives made with all dark inks at the same ink limit are sufficiently good that I doubt equalizing them a bit would make a discernable difference.

The fact that you see a transition on the negative, from smooth to grainy, at about 30%, suggests to me you are looking at the crossover point. This is the point where the dark inks kick in and start to take over from the light inks. And the fact that is it occurring at about 30% suggests that the Gray Curve function in your profile is turned off or not working. A Gray Curve that properly linearizes a negative for palladium printing shoves the crossover point down to a few percent -- at which place I think it will be invisible on either neg or print. So, does your Gray Curve function only have an address for a .acv curve following the = sign? If so, you may need to delete the existing address and put in the proper address for your computer. I can help if this is the problem.

You say you are getting a smooth gradient with a green neg but not with a QTR neg. Hmmmm. My guess is that green negatives are printed primarily by a combination of Cyan and Yellow inks with minimal contribution from other colors. In principle, printing with just two unbalanced inks ought not be smoother than printing with four unbalanced inks.

Final thought for now. I cannot tell from your post if you are judging smoothness from a pt/pd print, or from the negative itself. Since our eyeballs do not see in the UV, the print is really the only useful arbiter of smoothness.

If further trouble shooting is needed, it would be helpful to see the actual QTR profile you are using for these tests.

Cheers, Ron Reeder

 

[Colin Graham]

Keith- thanks very much for the offer. I guess at this point I'm primarily trying to get the printer to work like it should. The unrest about the grain I'm seeing got me to wondering about the UV transmission- though at this point I'm no position to pursue it much.

Ron, I hope it didn't sound like I was picking the process apart. I'm more than certain it's all due to something I'm doing wrong. The 3800 is brand new to me and I'm not used to the level of tweaking it offers, and coupled with the RIP, well, it's a little dizzying. At this point I'm still trying to make sure I'm setting up the basic procedures correctly, and that this dither I'm seeing is typical. The resulting grain is troubling me in the highlights, but I've never seen anything so smooth as the midtones and shadows the process is giving- even un-curved they are right on the money, in and out are aligning precisely.

I'm using the carbonprint profile from your website. The only thing I did was remark (=) out the gray curve line- mostly because I'm used to printing un-curved color stepwedges for initial tests for max black and pure white and wanted to get to appropriate ink densities before I applied a curve. If I understand what you're saying, is it better to always have a curve in place for the RIP to function properly, even if it's a 'flat line' curve with no adjustment?

Many thanks for the reply and info, and my apologies if my frustration sounded like it was due to anything other than my own inexperience.

BTW, the scan was from a direct scan of a printed stepwedge. But I'm seeing the same highlight grain in the prints, single transfer carbon on Arches HP. Maybe I should just use that as an indicator for ink density, and then curve the highlights from there? It's hard to get an accurate sample in PS due to the grain, even with a hefty does of Gaussian blur.

Another possibility I've considered is that the contrast range of my process isn't in the sweet spot of the 3800- that I should increase emulsion sensitivity and tone down the contrast and shoot for denser inks in QTR- the dither/grain pattern does seem to go down a bit with increased density. I hesitate to try it because my exposures are at 22 minutes already- but increasing the % of dichromate would most likely help with that too.

 

[Ron-san]

Ron, I hope it didn't sound like I was picking the process apart.

I'm using the carbonprint profile from your website. The only thing I did was remark (=) out the gray curve line- mostly because I'm used to printing un-curved color stepwedges for initial tests for max black and pure white and wanted to get to appropriate ink densities before I applied a curve.

Another possibility I've considered is that the contrast range of my process isn't in the sweet spot of the 3800- that I should increase emulsion sensitivity and tone down the contrast and shoot for denser inks in QTR- the dither/grain pattern does seem to go down a bit with increased density. I hesitate to try it because my exposures are at 22 minutes already- but increasing the % of dichromate would most likely help with that too.

Colin-- Sorry if I sounded defensive about QTR. I actually enjoy it when people pick the profiles apart rather than accept them as holy writ. As you can tell, there are many alternative ways of writing them and I am not at all certain we have hit on the best one yet.

Also, your approach (turning off Gray Curve until I have the overall negative contrast set) is exactly the way I prefer to go about making a profile as well. However, it is true that when Gray Curve is finally applied, it often shifts the cross over point dramatically -- which may (or may not) help diminish the graininess you are seeing.

The carbonprint QTR profile you are using was developed while trying to learn the carbon printing process from Sandy King at a work shop in Montana a month or so ago. It has not been tested very much as yet. My own carbon expertise is so vestigial that I am happy to get an image, much less start looking critically at the grain. Sandy has the same profile now and it will be of interest to see if he sees the same grain problems you are seeing.

Good luck, Ron Reeder

 

[Colin Graham]

Oh no, you didnt sound defensive. I just wanted to make doubly sure I wasn't coming off sounding like some ungrateful *head. I'm very grateful for the effort you've put into this and the generosity of sharing it with the community- even at the risk of dolts like me butchering it beyond recognition! Cheers and thanks again.

 

[sanking]

Colin,

You can reduce the appearance of grain significantly with carbon printing by using a sheet of thin mylar between the negative and the carbon tissue.The mylar kind of diffuses the grain without affecting sharpness, assuming you print with a vacuum frame.

I do this routinely because it is sometimes difficult to tell if the tissue has dried sufficiently, and if has not it will stick to the negative and ruin it. The problem is much more likely with carbon than with pt/pd.

Sandy King

Oh no, you didnt sound defensive. I just wanted to make doubly sure I wasn't coming off sounding like some ungrateful *head. I'm very grateful for the effort you've put into this and the generosity of sharing it with the community- even at the risk of dolts like me butchering it beyond recognition! Cheers and thanks again.

 

[Colin Graham]

Thanks Sandy, this is very timely advice- I just had a fresh step wedge stick to the tissue! Is there anything specific to look for when ordering mylar? I had been using one of those ultra-thin clear print envelopes as an interleaf, but it seemed like it might be blocking a bit of UV, so I just tried printing without it. Of course it pulled half the ink off when I separated the neg from the tissue!

I'm starting to think I picked the wrong process to try to learn Ron's QTR method with.. I probably should have started with kallitypes, VDB or another less time consuming process. An hour from test print to development is a bit trying with all these different things to try. Not to mention the tissue making, paper sizing, etc....Ever since I started with carbon though I want to do nothing else.

 

[sanking]

Hi Colin,

Well, the needs in terms of mylar kind of work against each other. You want a thin mylar to not lose detail, but you need a thicker one to diffuse the grain pattern. I think about 4 mil works well. I buy the mylar in boxes of 16X20 shees from Light Impressions.

You are right in that carbon is not the best process for learning about making good digital negatives with QTR. With either pt/pd, kallitype or VDB the texture of the paper prevents grain artifacts so there is a very smooth look.

Ironically, the silver profile that Ron posted prints with very little grain. I can not get my head around why the QTR negatives print with such fine grain on a smooth silver paper but not nearly as fine with carbon. And with carbon we really need fine grain if we transfer to smooth surface fixed out photo papers. I discussed the issue of ink blocking with Ron when we were at the Formulary earlier this summer but I still don't understand the issues well enough to adjust the inks to fully achieve my goal, which would be a grain free carbon print from a digital negative. However, what I am getting is not bad at all, at least as good as with the green negatives from the HP 9180 and about the same as the QTR negatives from the 2200. But banding has not been an issue at all for me with the 3800 so I see this printer as a big step up.

Sandy King

Thanks Sandy, this is very timely advice- I just had a fresh step wedge stick to the tissue! Is there anything specific to look for when ordering mylar? I had been using one of those ultra-thin clear print envelopes as an interleaf, but it seemed like it might be blocking a bit of UV, so I just tried printing without it. Of course it pulled half the ink off when I separated the neg from the tissue!

I'm starting to think I picked the wrong process to try to learn Ron's QTR method with.. I probably should have started with kallitypes, VDB or another less time consuming process. An hour from test print to development is a bit trying with all these different things to try. Not to mention the tissue making, paper sizing, etc....Ever since I started with carbon though I want to do nothing else.

 

[Colin Graham]

Wow, I feel pretty stupid. Apparently I was having a huge issue with that interleaf I had been using. It seems to have been blocking enough UV exposure to contribute considerably to the highlight grain.

Might be hard to see from the jpeg, but values over above 40 appear to cut out completely- it would seem even among the inks within the same step- and therefore give the appearance of heavy grain. The sample on the left was with the interleaf in place, the one on the right without (the spots are where the ink stuck to the tissue-oops.) Same negative for both exposures, same scan and B&W points adjustment.

It's odd, because it doesn't appear to have a linear impact on the wedge- the steps around 40-100% are very close because I'd adjusted exposure for the interleaf. But the 30-0% steps blow out almost immediately with the interleaf in place- there's must be just a little low-UV ink letting some exposure though. I should have thought to try this, but I wouldn't have thought the UV blocking could have been so disproportionate across the scale. But of course with the two different ink sets controlling each end of the step tablet, this is starting to make a lot of sense. I guess the grain get more obvious as the light inks crossover and give way to the dark inks.

Thanks for the info about the mylar Sandy- I'm off to order some more suitable stuff!