I am getting DMAX at 230 units on Nuarc 26-1K, but my step wedge 0% comes out as 18%

[Zero_Equals_Infinity]

Getting dmax calibration on the Nuarc 26-1K w using QTR was pretty easy, but now how do I bet a 0% dmin instead of 18%?

Here are my settings:

1) For 25 sq inch or Arches Platine (.250 ml Lithium PD, .250 ml FeO, .010 ml NA2 @2.5%)
Exposure = 230 units for max black. Vacuum reading = 6; Exposure = 230 units.


Any thoughts are gratiously appreciated.


Nick.

 

[pschwart]

Getting dmax calibration on the Nuarc 26-1K w using QTR was pretty easy, but now how do I bet a 0% dmin instead of 18%?

Here are my settings:

1) For 25 sq inch or Arches Platine (.250 ml Lithium PD, .250 ml FeO, .010 ml NA2 @2.5%)
Exposure = 230 units for max black. Vacuum reading = 6; Exposure = 230 units.


Any thoughts are gratiously appreciated.


Nick.

What you have done so far is determine the base exposure, which is the minimum time required to achieve maximum black. I assume you are using digital negatives since you mentioned QTR, but the base exposure has nothing to do with QTR since you are printing through clear OHP. This is the easy part The next are steps are more challenging:

- create a QTR profile for your printer and inkset that can achieve paper white at the base exposure.
Note that QTR is just one of many methods of doing this.
- then create a correction curve that will linearize all the tones between dmin and dmax

 

[Zero_Equals_Infinity]

That summarizes my position. I am at the beginning of the second step, and was wondering now that I have the base exposure, what are a good set of numbers to begin with. Here is what I am currently using, (and it seemed on the surface that I was already pushing the ink output, but then I am new to this, so I could be out to lunch.)


#
# PD/PT curve for Arches Platine on Inkpress Transparency film using
# ratios of Lithium PD, Ferric Oxalate, and NO2 - 2.5% of 24:24:1
# and a Nuarc 26-1K with Mylar sheet at Vacuum = 6
#


PRINTER=Quad3880
CALIBRATION=no
CURVE_NAME=PD-profile-Arches-Platine
GRAPH_CURVE=YES

N_OF_INKS=8
DEFAULT_INK_LIMIT=100

BOOST_K=50
LIMIT_K=25
LIMIT_C=25
LIMIT_M=25
LIMIT_Y=25
LIMIT_LC=40
LIMIT_LM=0
LIMIT_LK=40
LIMIT_LLK=40

N_OF_GRAY_PARTS=2
GRAY_INK_1=K
GRAY_VAL_1=100

GRAY_INK_2=LK
GRAY_VAL_2=30

GRAY_HIGHLIGHT=0
GRAY_SHADOW=0
GRAY_OVERLAP=
GRAY_GAMMA=0.70

GRAY_CURVE=

COPY_CURVE_Y=K
COPY_CURVE_C=K
COPY_CURVE_M=K

COPY_CURVE_LC=LK
COPY_CURVE_LLK=LK


The GRAY_GAMMA value was suggested in an R3800 profile which I used as a starting point. My understanding is that decreasing gamma will act much as pushing the centre slider in a levels layer in Photoshop, but that does not change the end points. I could keep pushing the Limits on the K/Y/C/M and BOOST_K, but it looks as though I am already pushing those values.

Any thoughts?

Thanks again,

Nick.

 

[pschwart]

If you are new to QTR profiling for digital negatives, Ron Reeder has published a helpful little guide:

A New Book on QTR and Digital Negatives

 

[Zero_Equals_Infinity]

Thanks ... that is a useful resource.

Nick.

 

[gmikol]

...it seemed on the surface that I was already pushing the ink output, but then I am new to this, so I could be out to lunch.

Does this mean you're starting to see a little bit of puddling or other artifacts of too much ink on your transparency, or is it just a "gut" feeling.

2 things that could cut down your ink usage:
-- Magenta inks have virtually 0 UV blocking ability. Set both M and LM to 0.
-- You also don't want your LLK to be the same as your LK. partition your K/LK/LLK properly, and you won't be laying down so much useless (LLK) ink in the midones.

If you choose to use cyan or yellow inks, set them as "toners" so they can have independent curves (but they don't have to). Black has 3 parts, cyan has 2, yellow only 1...they probably shouldn't use the same curves.

But I would try just making a negative with just K/LK/LLK and see where that gets you. I realize that PT is expensive...perhaps you could use a different medium for learning how to manipulate QTR digital negative curves. I just used RC photopaper to learn about curve creation. But cyano or VDB might be a good choice, too.

Here's a curve I used with some carbon a while back (never did get it linearized before I changed the recipe.)

#Notes Initial Ink Limits Targeted at 37.5PK / 25C / 20Y
#Notes After first set of measurements, limits increased
#Notes to 39PK / 26C / 21Y.
#Notes
#Notes Gamma of 1.4 added to all ink curves based on
#Notes my graphing tool.
PRINTER=Quad3880_GM
CURVE_NAME=Carbon_Batch_6
GRAPH_CURVE=YES
N_OF_INKS=8
DEFAULT_INK_LIMIT=
BOOST_K=
LIMIT_K=39
LIMIT_C=26
LIMIT_M=5
LIMIT_Y=21
LIMIT_LC=20
LIMIT_LM=5
LIMIT_LK=35
LIMIT_LLK=35
N_OF_GRAY_PARTS=3
GRAY_INK_1=K
GRAY_VAL_1=100
GRAY_INK_2=LK
GRAY_VAL_2=22.5
GRAY_INK_3=LLK
GRAY_VAL_3=9
GRAY_HIGHLIGHT=0
GRAY_SHADOW=0
GRAY_OVERLAP=
GRAY_GAMMA=1.4
GRAY_CURVE=
N_OF_TONER_PARTS=2
TONER_INK_1=C
TONER_VAL_1=100
TONER_INK_2=LC
TONER_VAL_2=8
TONER_HIGHLIGHT=0
TONER_SHADOW=0
TONER_GAMMA=1.4
TONER_CURVE=
N_OF_TONER_2_PARTS=1
TONER_2_INK_1=Y
TONER_2_VAL_1=100
TONER_2_HIGHLIGHT=0
TONER_2_SHADOW=0
TONER_2_GAMMA=1.4
TONER_2_CURVE=
N_OF_UNUSED=0
COPY_CURVE_M=K
COPY_CURVE_LM=K

Here's a graph of the inks...tried to attach it, but the server won't let me.



I'll note that I am running a tiny bit of magenta...just to keep the nozzles flowing.

Hope that helps.

--Greg

 

[Zero_Equals_Infinity]

Thanks Greg, it is a bit of gut feeling more than anything else. I was using the curve that Ron Reeder had specified in his book for his R3800 printer as a starting point and had pushed the values significantly, (about 50%). I use a Nuarc, and he uses something else, and the R3800 and R3880 have somewhat different inksets, so I may have more latitude than I think. Perhaps the Inkpress transparency film takes in more than the Pictorico, (which is what I used in a workshop last year.)

I have reduced the size of my step wedges to allow me to create 4 per 8.5 x 11 sheet of transparency film. (Yes, I create one per sheet, and cut it out, and then I scale the formula for the emulsion accordingly.) That stretches the chemistry a lot, and I bit the bullet and bought adjustable pipettes to enable precise measurements ... and that helps constrain the chemistry costs.

I will take your advice on the Magentas and LLK, and boost other areas, then once I get a good white, play with the gamma to get a decent curve.

I really wish that I could take the best UV blocking ink and load that into all 8 cartridges so as to have uniform distribution (360 x 8 = 2880 DPI). Then the whole thing would be simple as pie: (Exposure for DMAX, create simple curve, tweek for minor linearization issues.)

I will head back into the darkroom on the weekend to give it another go.

Thanks,

Nick.

 

[gmikol]

Keep in mind that depending on your chemistry, you might have a longer exposure scale than what was used in the example in the book, requiring that you use more ink to get a higher density.

I would guess that Inkpress and Pictorico are similar in their ink holding ability...you'll know when you've pushed them too far.

I'll also note that if you haven't printed the calibration page from QTR in your process, you might find it quite informative...you might discover that even the cyan inks aren't doing anything for you. Can anyone comment on the peak wavelengths for NA2 Pd?

--Greg

 

[Zero_Equals_Infinity]

Yes, I had expected, (naively), that the UV characteristics of the inks had not changed dramatically. I will either have to print the calibration page or find a reference to it online for the inkset used in the Ultrachrome K3 inkset. I was probably just trying to save some time and materials, (falsely ... duh.)

My understanding from Ron Reeder is that the use of a tad of NA2 is to inhibit highlight fogging, (though it can also be used to increase contrast in larger amounts.) I am not expecting the small amount added to have a significant effect, and there are cheaper non-chemical ways to alter contrast, (i.e. QTR curves and linearity corrections). As to the peak wavelengths, I will do a search in case it should be kept in mind while changing the curves.

I must thank-you once again for your helpful advice. When doing this for the first time, (one year after the workshop), it does have a slightly daunting feel to it. I am sure that once I have done this a few times it will feel as familiar as a favourite shirt.

After this is nailed, I will create a dichromate-gum curve. Adding a single burned umbra layer as we did in Ron's workshop increases the Dmax, and gives a 19th century feel to a print. He does this a fair bit, and if you have ever seen his prints it is quite a lovely effect for many subjects. (By the way he is an extraordinarily helpful and kind-spirited gentleman with just enough eccentricity to make him a joy to have leading a workshop. The playfulness in some of his prints expresses that aspect of his personality. A remarkable person, he has generously made the process of digital negatives accessible to people like me who would have been daunted by the old methods. If you are reading this Ron, my many thanks to you.)

Nick.

 

[gmikol]

I am not expecting the small amount added to have a significant effect, and there are cheaper non-chemical ways to alter contrast, (i.e. QTR curves and linearity corrections).

You're using contrast in 2 different ways here...

NA2, dichromate, chlorate alter the exposure scale of the process. i.e. you need a more- or less-dense negative to achieve paper white.
"Digital" contrast adjustments keep the white and black points the same, typically, but change the curve shape, as would a curves adjustment layer or curve shaping controls in QTR.
QTR ink limits don't change the contrast of the process, either...but it allows you do put down more ink (get more density) to accommodate the exposure scale of the process. If you can't lay down enough ink, then you need to try a different substrate or add some contrast agents to your chemistry. It seems like Pictorico can handle about 200% "total ink", don't know about Inkpress. The profile you posted has 125% total ink. (50K + 25Y + 25C + 25M, LC, LM, LK, LLK are not printing at the far end of the curve.

If you don't want to go back to start all the way from the beginning on your curves (i.e. from the Calibration page), I've also had success with just scaling up all the inks by some large amount (25 to 50%, that is, from an ink limit of 50 to 50*1.25=62.5, from 20 to 25, etc.) The intention is to overshoot how much ink you need, and then by looking at the step wedge you printed, you can figure out how much to pull it back. If you see paper white at 90% on the step wedge, then you can then scale all those inks back by a factor of 0.9, from the example above 62.5 would go down to 56.2, 25 down to 22.5. Now you're ready to make a print for linearization.

--Greg

 

[Zero_Equals_Infinity]

It sounds like I want to boost the ink load up to the saturation point for the Inkpress substrate, and then use the inks which will give max blocking, and only then consider chemical means if that does not yield a pure white. (I would like to avoid the chemical approach if possible.) Thanks for the idea of scaling. That makes sense, and I will try this first as it has the virtue of simplicity. I will assume that a 200% load is saturation, and start scaling towards that. As you point out, the blocking factor should be linear, so calculating the correct setting if I get white at 90% is pretty simple. Assuming I get my 100% pure white, I will then do some curve shifting (Gamma changing) to get a reasonable curve, and then make linearity adjustments.

I am indebted for the time you have spent responding, and I will keep you appraised of progress. I think this will work.

As an aside: I wonder if a refill system for these printers would accept one pigment / colour only in all cartridges, (see InkRepublic - R2880 CIS, 3880 CIS, 3800 CIS, R1900 CIS, R2400 CIS, R1800 CIS, R800 CIS, R280 CIS, 1400 CIS, C120 CIS), to give an ideal setup for digital negatives. The process would require the usual Dmax exposure iteration, followed by determining the density for pure white, and then creating the linearization curve. While no steps from the traditional method are dropped, the number of tests required would be fewer, and any corrections less extreme and with more predictable results. I am tempted to contact them to see if there is a technical reason that the same ink could not be applied across all nozzles, as it strikes me that the ultimate solution would be to use the best blocker in every cartridge. It should yield a simple curve given the linear nature of UV blocking and density. (I may have to buy an old cheap Epson with small cartridges to try out this possibility. If it worked it would be a great and easy way to provide smooth tonal transition, and a more consistent and easy method to calibrate ink for a substrate / emulsion / paper.

I may open a thread to see if anyone has explored this idea.

 

[pschwart]

As an aside: I wonder if a refill system for these printers would accept one pigment / colour only in all cartridges,

Sure it can -- the printer relies on the cartridge chip to ensure that the correct color ink is installed in each slot. As long as you have the right chip on the cartridge, you can fill the cartridge with any color you like. But read my post about why you might want to reconsider this strategy: http://www.dpug.org/forums/viewpost.php?p=21654

 

[Zero_Equals_Infinity]

More fun ... long learning curve ... and the joys of min White.

Sometimes it feels like two steps forward, one step back.

Below is what I plan to try next time in the darkroom, (after a not terribly productive day yesterday.)



The two major problems I have had are getting a min white (100% on curve), and a decent transition in the deep shadows (0 - 10% on curve.)
In my previous step curve the use of LK is pretty much as shown, (a very steep transition between 0 - 10% should have resulted in a significant gradient, or so I thought, and yes I have verified that I am using the correct exposure for Dmax black based on the boundary where a transition from clear film to no film is not noticeable.)

The other problem is dmin (pure white), which I never seem to get closer to than 8% on the step wedge. (My previous curve had 100% PK and 18% Yellow. at 100% on the curve.) I have now changed that to use Yellow as my base which has a fairly flat transition at the 90 - 100% portion of the curve, and boosting the PK to control rate of change in the 90 - 100% of the curve. If I am still not hitting dmin, I have enough latitude to push Yellow as high as 70% to raise the overall amount of UV blocking in this range while using PK to control the transition.

If I get that under control, I can play around a bit with LK density and limit to tune the mid-tones before doing linearization.

Below is the curve I was using but which proved unsatisfactory (as described), and any criticism or thoughts are invited. Thank-you

 

[gmikol]

The other problem is dmin (pure white), which I never seem to get closer to than 8% on the step wedge. (My previous curve had 100% PK and 18% Yellow. at 100% on the curve.) I have now changed that to use Yellow as my base which has a fairly flat transition at the 90 - 100% portion of the curve, and boosting the PK to control rate of change in the 90 - 100% of the curve. If I am still not hitting dmin, I have enough latitude to push Yellow as high as 70% to raise the overall amount of UV blocking in this range while using PK to control the transition.

Not quite clear what you're saying here...Are you hitting paper white at 8% on the step wedge? In which case you have too much density (ink) and need to back off slightly. If everything from 8% and below is some shade of gray darker than the paper, and there's no differentiation, then you might be getting some fogging or staining. If there is differentiation below 8%, but 0 is not paper white, then you need to add more density.

--Greg

 

[Zero_Equals_Infinity]

When I measure what I want to show as pure white (0% on the wedge), it measures as 8%. The curve is also not showing much change from 10% down. I am attributing this lack of change to the fact that the PK curve is flat from roughly 85% on.

Since PK has a steep slope from 10% to 45%, I am going to try using it to show gradation in the far right (85 - 100%) to both get me my pure white, and to show differences in this range of the curve. I am hoping that this will work if Y has sufficient blocking to add with PK to be >= 100%.

I will make up a new batch of clearing bath to insure fogging is not an issue.

I had read in another thread that using the very poor blocking inks (LLK, M, LM, C) has some sort of "catalytic" affect to improve blocking and reduce banding issues. Do you have any experience with this, as it seems both wasteful of ink and useless to use the very poor blockers unless there is some other advantage of which I am unaware. As it sits, I am going to allow tiny amounts of these poor blockers and use K, Y for the heavy lifting, and LK on the left of the curve, to provide gradation from 0 - 10%, and to fine tune the midtones by changing the density to shift the curve, or limit for the slope and peak. (It requires training myself to think differently from what I am used to, to imagine the effects of the curves, to test them and adjust in a fuzzy way towards an acceptable curve.) The funny think is I am actually enjoying playing with curves, and will probably learn more about Pt/Pd from this, than printing real images. Don't get me wrong, I want to print real images, and then add a dark umbra gum layer for shadows, but I need to learn to walk before running. (I'll get there with a little help from my friends.)

Thanks,

Nick.

 

[gmikol]

When I measure what I want to show as pure white (0% on the wedge), it measures as 8%. The curve is also not showing much change from 10% down. I am attributing this lack of change to the fact that the PK curve is flat from roughly 85% on.

Wait...Are you using the scanner method of measuring for linearization?

I don't use that method, but I believe it requires you to set the white and black points in photoshop before you measure (or in your scanning software, if it allows it). A scanner will only see something that perfectly reflects all light as "white", everything else, including white paper, will render as some very light shade of grey. So you might not have any problem at all, just make sure you're doing the measurement process correctly.

I had read in another thread that using the very poor blocking inks (LLK, M, LM, C) has some sort of "catalytic" affect to improve blocking and reduce banding issues. Do you have any experience with this, as it seems both wasteful of ink and useless to use the very poor blockers unless there is some other advantage of which I am unaware.

I've never heard of this, and the physics really doesn't bear it out...there is AFAIK no "super-additive" effect of using these inks. Can you provide the link where you read this?

If you've made a print of the calibration page from QTR, then you already know the answer to which inks work well for blocking, and which work poorly.

I believe I've posted this before:
http://www.albertonovo.it/scan/epson_inks.html

The magenta inks have relatively low absorbance in the UV region. Depending on your light source and chemistry, C and LC may be beneficial, as it does have an absorbance that starts to rise very quickly below 400nm (this is good for dichromated processes at 360-370nm, not sure about others). You should absolutely use LLK as a third gray ink, with appropriate partitioning. It may not reduce banding, but it should reduce the visibility of halftone dots.

--Greg

 

[Zero_Equals_Infinity]

Greg
Wait...Are you using the scanner method of measuring for linearization?

I don't use that method, but I believe it requires you to set the white and black points in photoshop before you measure (or in your scanning software, if it allows it). A scanner will only see something that perfectly reflects all light as "white", everything else, including white paper, will render as some very light shade of grey. So you might not have any problem at all, just make sure you're doing the measurement process correctly.

Yes that is the method I am using. I set the white and black points using the eye dropper on the V700 scanning software. I use the white dropper on an area where there is no emulsion, and use the black dropper on an area with emulsion that was not covered by film. That sets the input points, and the output points are set to 0 and FF / 100% respectively.

Greg
I've never heard of this, and the physics really doesn't bear it out...there is AFAIK no "super-additive" effect of using these inks. Can you provide the link where you read this?

If you've made a print of the calibration page from QTR, then you already know the answer to which inks work well for blocking, and which work poorly.

I believe I've posted this before:
Epson inks

I got it from a link provided on another thread to a document in which an article pointed to and written by Ron Reeder describes some woes he had using just K, and LK.

Ron Reeder article
Once I had all inks contributing, I printed out a test step tablet with the default ink limit set rather high (75%) and got a surprise. The darkest step of the tablet had an optical density in the UV of 4.6. This is the highest UV density I have ever seen on a digital negative. Apparently combining all the inks can lead to extremely high UV densities. The UV density that is needed to print pure white on a palladium emulsion is only about 3.2. So I kept dialling the default ink limit down until I reached a setting of 44% -- which gave me a UV density in the darkest step of just over 3.2. This QTR profile prints negatives that yield the smoothest tones on a palladium emulsion that I have ever been able to achieve. I conclude that an additional reason why this profile prints such smooth tones is that no one ink is used to a very high amount. But together they easily achieve the needed UV density.

This indicated to me that there was some synergistic effect, though as you say, the physics, (at least as I understand it), will not work. On the other hand K flattens at around 40 to 45%, and Y starts to contribute substantially in that range. So he may have just been describing the normal additive effect, and the fact that black tops out pretty early supports that interpretation. Hence, the increase likely comes from Y, and smooth tonal transitions from a more even distribution from more inks being used.

Greg
The magenta inks have relatively low absorbance in the UV region. Depending on your light source and chemistry, C and LC may be beneficial, as it does have an absorbance that starts to rise very quickly below 400nm (this is good for dichromated processes at 360-370nm, not sure about others). You should absolutely use LLK as a third gray ink, with appropriate partitioning. It may not reduce banding, but it should the visibility of halftone dots.

--Greg

I will keep that in mind and play around with it. Thanks again, (and I played darkroom hooky today to go and shoot mating dragonflies and moving water.)

 

[gmikol]

Nick...3 points:

1) I think you need to check to rule out fogging before you try increasing your ink density.

2) If it turns out you don't have fogging, I've found that a sure-fire way to find my ink limits is to scale everything up by some big factor, like 25% or more so a limit of 40 becomes 50, 20 becomes 25 for 25% increase. Then print and see on your step wedge where you hit paper white. If you hit at 10% (90% ink), then your 50 becomes 45, your 25 becomes 22.5, etc. I personally think it's easier to overshoot and come back than to try and creep up on the right ink limit.

3) Keep in mind that Ron Reeder wrote that article in regards to an Epson 4000...a fairly old printer. It did not have a 3rd black ink, and it used a substantially different inkset than the current K3 or K3VM inks. So his observations are really only relevant to that printer model. It's possible that they early Ultrachrome C and M inks had more UV absorbance than the new K3 C and M inks do.

--Greg

 

[Zero_Equals_Infinity]

Thanks for your help Greg. I checked for fogging today, and "fortunately" it was not fogging. (I placed a piece of thick opaque paper over part of the emulsion to block UV completely, and when developed and cleared, it measured zero as hoped for.)

The max inkload for Inkpress film is 120%, (and that is taking it right to the edge.) I played around with the yellow and black curves, and finally found the sweetspot. I am now at 2% for pure white, and will see if there is latitude to reduce the exposure from 230 to 225 units. Max black measured at 230 initially, but I notice with a higher emulsion load on the paper the dmax seems to hit at around 99% on the step wedge, so I think I have it. As a bonus the midtone portion of the curve was very close to linear, and requires the smallest of linearization correction.

Finally, I am within spitting distance of printing real images. Thanks to all who have contributed with suggestions. (Now if I can only get rid of the occasional pizza wheels.)