Multiple Linearizations


#1

Hello fellow DN printers,

l have a question about multiple Linearizations for the group. I started over with my PdPt calibration scheme by going back creating limiter curves as my starting point for each of the coating/paper combinations that I print. That process went well and then I created the Linearization curves and prints and it’s a great improvement over the first versions where I started with the IJM Pd80Pt20 and PdNa2 curves. My question is, if the results are not quite as linear as I want can I just take the latest .quad that I Linearized and then run it through the QTR droplet again with the new .txt file? How many times can I do that? What’s the groups experance?.


#2

You can relinearize to your heart’s content, but you are unlikely to improve upon the first one in my experience. I’m a little obsessive about trying to attain perfect linearization, but over the last 2 years of working with PiezoDN I have come to realize that with all the sources of variation in the printing system (I mean in the darkroom, not making the negatives) such as slight variations in coating or fluctuation in exposure light output, that doing so is an exercise in futility.

I think the best practice would actually be, rather than rely on a single print of the target negative, to print it several times and average the results, then use that to make the linearization, and do the same to verify the linearized quad. Most of us, me included, don’t want to do that because it is so time consuming, but in the end it may actually save time. This is just a recent thought - I haven’t put it into practice - yet.

Here’s an example where I made 3 iterative linearizations:

[attachment file=29939]

In the end, I have chosen to use the first linearization because I like the way it looks in actual prints. The others look too dark, especially in the highlights. I suspect that there was a fluctuation in exposure light output that resulted in a target print that appears to be too light, but that if I reprinted it a few times and averaged the results I would find it was an outlier.


#3

Hi Keith,

Thanks for the response. I was surprised and starting to feel neglected that nobody had responded to my fascinating question!

I totally agree with your comments above and as time permits will try your idea of averaging the readings from a few prints over a short period of time. It is amazing that no matter how hard I try to be consistent and standardize everything, the variability on the “wet-side” of the process is always a bigger factor than probably most realize. I am always thinking about factors like; the humidity in the air, the age of my Ferric Ox, the Ph of KOx, etc… I guess that’s part of the allure of the process that no two prints are exactly alike.

Screen Shot 2018-04-12 at 1.10.19 PM.png

Here’s my Linearization of Pd80Pt20 on Arches. I think it’s pretty close and I am happy with the test print results. It’s funny even before I read the print with the spyder I could visually tell the print was much more linear. Just the way the test image blocks are arranged. The other factor that I have observed that has been really helpful is pressing down pretty firmly on the SpyerPrint head and to have a few sheets of paper beneath the test to not allow any stray light to throw-off the measurements.

Last question, do you use the ICC profiles as suggested in the manual to bring down the shadows a bit?

Cheers and thanks again.


#4

So sorry for not replying! I’ve been moving a massive amount of equipment this week and didn’t see this post.

The ICC will attempt to add contrast to your print to match the higher contrast of your monitor. Some really like it, some do not. I’ve been leaning towards linear (not ICC) myself of late because I feel like the resulting print is closer to the way a normal PtPd print behaves tonally. (It’s also they default Piezography workflow too which exists to replicate the tonal longevity of Platinum.)

 

best,

Walker


#5

I usually get email notices when new topics post on this forum, but didn’t get one for this. I did write a reply yesterday, but when I clicked submit it went to the login screen and my work was lost.

I agree that looks pretty damn good. (Excuse the technical jargon. ;)) It might be worthwhile for you to do a 2nd iteration just to see what happens for yourself. It’s good practice if nothing else.

I don’t use the the ICC workflow personally. Walker compares it to matching the higher contrast of the monitor. Being more of a darkroom than a digital printer, I think of it as more approximating the typical contrast profile of a 20th century silver print. Some strive for that look even in a Pt/Pd print. I like the inherent characteristics of Pt/Pd and especially of Pd alone. If anything, I would go for lower contrast, not higher. But I prefer to let the linearization be linear and prepare my files to print the way I want them to based on that.

That being said, there are ways to make custom shaped curves or ICCs to accomplish whatever end result you want. I’ve played with that a bit, but still prefer to do it in the file.


#6

I would add a word of caution about multiple relinearisations, always against the most recent linearisation. I have been doing this for some time without problems. However, after my most recent relinearisation I started to get horrible gradations (not banding) in light sky areas.

I initially thought this must be down to old inks, but a test print against an earlier quad removed the banding. Reviewing the series of quads, the smoothness of the individual ink profiles gradually deteriorated and some became distinctly ‘wobbly’.

I have now started again from scratch and intend to always relinearise against a ‘master’ quad. Perhaps it was my bad luck, or working practices, but it is something to consider if you see similar problems.


#7

I agree that it doesn’t make a lot of sense and that it probably would introduce errors in the math. I agree with Keith’s idea of printing multiple targets over a few days or weeks and then averaging the results then trying to run the linearization software if you really want to get it done right. So much happens in the “Wet” side of the process. I was recently testing my PdNa2 (2.5%) on Arches Paper combination and noticed that since I was using 15drops of Pd and only 1 drop of Na2 that my results were all over the place depending on the size of the drop. I went back and mixed a larger quality of PdNa2 at the same ratio but pipetted out quantities and the results are much better.


#8

This very thing is going to be addressed in a new update to Pro Tools and PiezoDN in near future. aka, how to “rehabilitate” curves.

Can’t say anything now, but soon.

W


#9

Walker, I should have known you would have it covered. It would be useful (I think) if the ‘rehabilitation’ could be applied at an early stage in the linearisation process - I have resorted to using excel to manually remove some small bumps in the ink profiles before finalising a master quad. I have another question about the profiles - I will start another topic.

On the subject of drop size, I suspected this was a source of poor repeatability so I purchased a dirt-cheap electronic balance (200g x 0.01g). I don’t know how good the absolute measurements are, but the repeatability of the balance seems to be excellent. This showed big variations in drop volume from day-to-day and in particular between new and used pipettes and new and old chemistry. I still measure volumes by number of drops, but with the beaker on the scales - it is quite easy to squeeze-out half drops to get consistent volumes.

Oh, and I do not use the standard Piezo Pt/Pd ICC profile as I find it way too strong. I guess it depends on the screen quality and the viewing conditions of the screen and print. As I am using Windows, so cannot use the ICC within QTR, I created a photoshop acv curve and apply it to the image before printing. The curve gives half the correction of the ICC throughout the tonal range and, for me, seems right for the majority of prints - still too strong for some, though.