Choosing Colours and Curves: The HSL Array

[mkochsch]

To Change the Negative, or the Chemistry?

Yes different printers and inks are going to produce different results. Add to that different light sources. Seems that printing digital negatives runs a close second in its individuality to gum printing in general. It's possible (what you're inferring) that a particular manufacturer puts "a little something" into its yellow ink but not into its cyan ink that inadvertently blocks more UV. Many UV blockers used in industry have no apparent colour at all. The yellow-green-red-blocking-more "theory" may be just anecdotal -- but if it is just anecdotal the evidence is pretty strong as many people report that blues are also the weakest blockers...when you see your own tests you can draw your own conclusions. I think that for practical purposes what's more important is to document (read: calibrate) what your printer is doing and move forward from there and not get too bogged down with the physics of the phenomena.

The last paragraph I simply meant to choose a blocking colour that errs on the dense side rather than the thin side of the process. That way you would always be assured you had highlight rather than underexposing or changing chemistry to clear the highlight. When the negative more labourious to produce (a more fixed object that you printed around) it was not so much a variable. Now we have to comes to terms with when do we chnage the negative and when do we change the chemistry.

 

[Katharine Thayer]

Contact!

Okay, I think we're getting there. Thanks for clarification; that makes sense.

But sorry, another question. I've never heard of the "yellow-green-red-blocking-more theory." Is there some ranking implied in the theory, or is it just that yellow green and red all tend to block better than blue? I would have agreed with this off the top of my head, because I remember that when I did Clay's ternary triangle, the blue corner was the one that didn't give me any blocking to speak of, but when I went down to the darkroom and found and unpacked my box of test prints from my tests with the 2-color arrays of color density patches on a lot of different pigments, I didn't find this to be true. As I've reported before, it was a strong red with a little blue or a little green that was the best blocker for me with all pigments at stock solution (very saturated). For all, the green and blue rows printed dark, and pretty much equally dark, in other words, green was not a better blocker than blue in my tests. One of the green rows and one of the blue rows (the ones with red in them) blocked better than the other ones, but still printed way darker than you'd want a highlight to be.

The only exception was PR 209 which was inadvertently way underpigmented (about the same dilution as the sample you showed) because I had already dumped the last of the last tube of paint into 30 ml gum before I realized there was no backup tube available. That underpigmented mix gave me a blocking color of a strong green with a little red in it. I'm inclined to think that the reason this produced a different blocking color is because the color was way undersaturated compared to the other pigments, but as I said the other day, I'll need to redo that with a properly pigmented mix to be sure.

I'm just adding this to your other anecdotal evidence, I'm not drawing any conclusions other than for my gum emulsions, for my printer and substrate (and light, yes) red rocks.

Thanks again for your patient help.

Katharine

Yes different printers and inks are going to produce different results. Add to that different light sources. Seems that printing digital negatives runs a close second in its individuality to gum printing in general. It's possible (what you're inferring) that a particular manufacturer puts "a little something" into its yellow ink but not into its cyan ink that inadvertently blocks more UV. Many UV blockers used in industry have no apparent colour at all. The yellow-green-red-blocking-more "theory" may be just anecdotal -- but if it is just anecdotal the evidence is pretty strong as many people report that blues are also the weakest blockers...when you see your own tests you can draw your own conclusions. I think that for practical purposes what's more important is to document (read: calibrate) what your printer is doing and move forward from there and not get too bogged down with the physics of the phenomena.

The last paragraph I simply meant to choose a blocking colour that errs on the dense side rather than the thin side of the process. That way you would always be assured you had highlight rather than underexposing or changing chemistry to clear the highlight. When the negative more labourious to produce (a more fixed object that you printed around) it was not so much a variable. Now we have to comes to terms with when do we chnage the negative and when do we change the chemistry.

 

[mkochsch]

Switching to VC Silver w/ Grade #5

View attachment 97

Compare and contrast. That's what they taught me in school. So here's some Agfa MultiGrade paper I shot last night. I used a #5 filter so I could attempt to use density colours from "above" the HSL colour equator. It doesn't happen. Not with my inks and printer. Luckily, I still have some choices below the equatorial line.
It's a little troubling in that the area on the left doesn't really give a good clean white does it? This relates to a problem Loris has noticed in another thread on the 9180 printer in his attempts to calibrate to his current system (psst Loris, join the dark side....).

Another thing seeing all these HSL Arrays is opening my eyes to is that we can classify the Array into three general areas.
View attachment 95
At the top there is usually the least amount of effective density. My gum test with a range or .7 logD didn't even read this high up. This is where the curve spends a lot of energy and ink filling in contrast. The middle area is where we seem to see the most change occurring. As exposure sets in you can track how it sweeps across certain colour areas or not. And, like I already mentioned...you can be tricked into thinking you're getting white...rather trick yourself but in truth you're getting "off-white" and you're in the weeds. Look at that grey area with no real contrast change -- what does that curve look like? The equator represents the swatch of colours present in RGB density wedges (i.e. PDN's CDRP).
Then, finally, near the bottom we finally get to the point where actual black ink is kicking in supplied from the driver. This is the "densest" area. Probably useful for me when I want to print the iron and noble alternative processes. I'm at the upper fringes of this area with my Silver paper coming in around 1.2 logD. Different emulsions of course will behave differently depending on their characteristics (i.e. self masking or not).
My choice for a blocking colour is based on avoiding the "weeds" the grey area and probably heading for either the 130 degree column around the low 40 mark (RGB 0,102,26) where the contrast seems to be constantly progressing, or to the 0 (zero) degrees column around RGB 255,0,0 where it's white but a little uncertain. If I take my first choice I know I'll get a good white, I might need to do a second iteration like I did with my gum test and the 101 step wedge but at least I know I'm avoiding a problem area.

~m

 

[mkochsch]

Ah ha I've figured it out...only two images per post. Can we change this??

View attachment 98

This shows what happens when I give my Array about one stop to two-thirds of a stop more exposure. My colour choices are all near the bottom of the Array now. But I still have lots of options.

~m

 

[mkochsch]

View attachment 99 View attachment 100

The resulting uncurved chart on the left and curved chart on the right.
Note: the curve was calculated by hand as I felt ChartThrob was having issues in the shadows. It still needs a little tweak around 40 per cent but it's close to perfect otherwise.

~m

 

[mkochsch]

The Curve File

If your interested in looking at curve...bit of a lifted toe but all in all not too tough a curve.
View attachment 101

~m

 

[Katharine Thayer]

"Density range" for gum.

The use of a particular blocking colour is for the sake of the negative and used to find an equal to the emulsion's highlight property that's all. e.g. Emulsion XYZed prints highlight at density 1.05, ink combo R255, G0, B64 also prints UV highlight at 1.05....therefore they match. The technique of choosing a colour to so that you don't make negative better suited to pt/pd than gum.
~m

Loris: You're right. Red -- has been giving me problems. My range is barely .70 logD. The attached sample was shot at 6 minutes (the one in the sink from last night is 3:30 seconds. I'll post it after breakfast ) To compare this to another colour, I've been shooting my yellow plate at 14 minutes and getting 1.1 logD, but the red was problematic to say the least. Thx.

~m

Sorry, another annoyingly stupid question coming up.

I understand that the density range referred to is the density range of the negative, not the density range of the print (it took Michael a while to get that through my head several months ago) but I'm still having some trouble with the concept of a "density range" for gum, when what you're really talking about is the highlights. Yes, yellow will always print with lighter highlights than a darker color, and a very diluted pigment will always print with very light highlights, but neither of those will give you much contrast (DMax) in your print, regardless of the amount of exposure. So, with a very light/diluted color, you'll get a high DR, if the DR is purely a measure of the highlight density, as the examples above and your discussion seem to indicate, but almost by definition, as a result of the short range of gum, an emulsion that gives you a very light highlight will not give you a dark DMax. So what's the object in aiming for a high DR?
kt

 

[Anonymous]

Katharine, I think that part of the confusion is one of the use of certain terms in different contexts. DR is normally considered to be the density range of a negative material. The DR is capable of producing the optimal print when it matches the ES (exposure scale) of the print material or printing process.

Dmax is the term that is used in describing the measurement of a reflective densitometer reading of a print deepest black (in the case of a black and white image). Dmin is the term that is used to describe the reflective densitometric measurement of the lightest value of a print material or process.

If one were to measure the dmin and the dmax of a print, one would be able to determiine the ES and from that determine the DR that a negative would best exhibit to print on that material or in that process.

I hope that this helps.

 

[Katharine Thayer]

Katharine, I think that part of the confusion is one of the use of certain terms in different contexts. DR is normally considered to be the density range of a negative material. The DR is capable of producing the optimal print when it matches the ES (exposure scale) of the print material or printing process.

Dmax is the term that is used in describing the measurement of a reflective densitometer reading of a print deepest black (in the case of a black and white image). Dmin is the term that is used to describe the reflective densitometric measurement of the lightest value of a print material or process.

If one were to measure the dmin and the dmax of a print, one would be able to determiine the ES and from that determine the DR that a negative would best exhibit to print on that material or in that process.

I hope that this helps.

Well, yes, that's what I had always understood, and that helps to clarify some of the confusion, but using this terminology doesn't answer the question, just puts it in different words. If one were to measure the dmin and dmax of a gum print containing one of a selection of yellow pigments, for example, one would find that the range of tones (the exposure scale, yes?) would be very short, since the DMax would be very light in absolute value, very close to the dmin. From the discussion above, I gathered that the blocking color (which is said to be matched to density range) is chosen to match the density of the highlight, without regard to the other end of the exposure scale; that's what confused me. Thanks for setting the terminology straight.
Katharine

 

[mkochsch]

... If one were to measure the dmin and dmax of a gum print containing one of a selection of yellow pigments, for example, one would find that the range of tones (the exposure scale, yes?) would be very short, since the DMax would be very light in absolute value, very close to the dmin.
Katharine

If you were analysing the output by converting it with a monochrome reflective densitometer this might be true but if you were reading it using the correct tool, a colorimeter, you would see something else. You can observe this by using the colour channels in Photoshop Levels.
View attachment 102
Remember with gum we're more often dealing in colour. Yellow is hard to see with human eyes but it marches the scales in gum just like the other colours.
The above swatch is joined at step 3 and 4 but can be observed stepping all the way from 5 to step 11. 12-13 and onward are white again. You need to switch to Levels and hold the alt/opt key and "scrub" the triangles to see it, also try switching to the Blue channel in Levels. Now, considering lowly I can get .7 logD from a negative to show up in print think what an experienced gummy like yourself should be able to do. BTW, that gum test you posted a few days ago looked really good as far as saturation and distribution of tones, you're obviously doing something right despite sometimes claiming the opposite.

~m

 

[Katharine Thayer]

If you were analysing the output by converting it with a monochrome reflective densitometer this might be true but if you were reading it using the correct tool, a colorimeter, you would see something else. You can observe this by using the colour channels in Photoshop Levels.
View attachment 102
Remember with gum we're more often dealing in colour. Yellow is hard to see with human eyes but it marches the scales in gum just like the other colours.
The above swatch is joined at step 3 and 4 but can be observed stepping all the way from 5 to step 11. 12-13 and onward are white again. You need to switch to Levels and hold the alt/opt key and "scrub" the triangles to see it, also try switching to the Blue channel in Levels. Now, considering lowly I can get .7 logD from a negative to show up in print think what an experienced gummy like yourself should be able to do. BTW, that gum test you posted a few days ago looked really good as far as saturation and distribution of tones, you're obviously doing something right despite sometimes claiming the opposite.

~m

Goodness, I certainly hope I haven't given the impression I'm claiming I don't know what I'm doing when it comes to gum. I KNOW how to print gum. I know gum forward and backward, upside down, with my eyes closed, in my sleep. I can find a way to work with gum to make it do anything I want it to do, and I've wanted to do a whole heck of a lot of different things with it, from strong high-contrast images, to monochrome images of normal tonality range, to coarse images like the early pictorialist gum printers, to tricolor, to the work I did for several years after 9/11, which was very quiet and high key with no contrast at all, as a sort of understated protest against violence; the image was barely visible on the paper. (Most photographers never "got" what I was doing with that body of work, someone on the alt-photo list even called it "crap" in a veiled remark, but the people who collect my work understood it completely, and bought it, and I personally consider it some of my best work, although in the last year or two I've started craving contrast and gone back to printing stronger images). But I digress....at any rate, no modesty from me, false or otherwise, about my ability to print gum, in any tonal range of my own choosing, using only printing variables as control.

But back to the digital negatives, about which I know very little and am quite willing to admit ignorance: well, sure, you can print as many steps with a lighter pigment (in most cases more steps) than with a darker pigment; all I was saying is that the darkest step, as read by a densitometer, won't be very far in absolute tone below the lightest step. I gather that's not an issue, so I guess I'll quit puzzling about it... ??

 

[sanking]

If one were to measure the dmin and the dmax of a print, one would be able to determiine the ES and from that determine the DR that a negative would best exhibit to print on that material or in that process.

I hope that this helps.

Donald,

If I understand you correctly I believe this part is not right. The ES of pure palladium is on the order of 2.2, but if you were to measure the difference in reflective values between Dmin and Dmax you will find a difference of no greater than about log 1.5. Thus, palladium has a very low Dmax compared to silver papers, but a very long ES.

The ES of a print process is measured by printing a step wedge and then figuring the log range between maximum density (or 90% of Dmax) and minimum density or paper white. Or maybe that is what you said and I just misunderstood you?

Sandy King

 

[Katharine Thayer]

Donald,

If I understand you correctly I believe this part is not right. The ES of pure palladium is on the order of 2.2, but if you were to measure the difference in reflective values between Dmin and Dmax you will find a difference of no greater than about log 1.5. Thus, palladium has a very low Dmax compared to silver papers, but a very long ES.

The ES of a print process is measured by printing a step wedge and then figuring the log range between maximum density (or 90% of Dmax) and minimum density or paper white. Or maybe that is what you said and I just misunderstood you?

Sandy King

This was the crux of my earlier puzzlement, how an emulsion could have a very short absolute distance between DMax and Dmin (a short tonal range) but generate a high DR for the negative, if the DR is figured from the tonal range. From what Michael said, and from this further explanation, I gather that the ES is figured not from the tonal range, but from the number of steps printed, regardless of the tonal range of the print. Okay, I can grock that, I guess.

On reflection, I think I wasn't being fair to myself when I said I know very little about digital negatives. Actually I've been printing gum from digital negatives for 15 years, using first a primitive method I came across in a magazine somewhere, then after Dan's book was published, using Dan Burkholder's method. That was all pretty straightforward and made perfect sense to me. You print the negative, then you print the negative on your emulsion, then you compare the values you get to the values you want, change the curve to match, repeat until satisfied, and you're done. No complicated charts and palettes and jargon, just straightforward and easy. So it's not digital negatives I have a hard time making sense of, it's this way of approaching digital negatives.
Katharine

 

[Anonymous]

Donald,

If I understand you correctly I believe this part is not right. The ES of pure palladium is on the order of 2.2, but if you were to measure the difference in reflective values between Dmin and Dmax you will find a difference of no greater than about log 1.5. Thus, palladium has a very low Dmax compared to silver papers, but a very long ES.

The ES of a print process is measured by printing a step wedge and then figuring the log range between maximum density (or 90% of Dmax) and minimum density or paper white. Or maybe that is what you said and I just misunderstood you?

Sandy King

Sandy, That is what I should have said. What I said was not correct.

 

[Katharine Thayer]

Red will not allow much UV to reach the bottom of the emulsion layer because it filters it very effectively (think of rubylith masking tape - it's red!) and therefore it will increase the contrast of the coating solution.

Loris, do you have practical empirical data to back up the above statement?

As has been discussed earlier in this thread, there's not a clear relationship between negative color and UV blocking, and I assume that would apply also to the color of the pigment in the emulsion. But any data that bears on the question one way or the other would be useful, thanks...
kt

 

[sanking]

Loris, do you have practical empirical data to back up the above statement?

As has been discussed earlier in this thread, there's not a clear relationship between negative color and UV blocking, and I assume that would apply also to the color of the pigment in the emulsion. But any data that bears on the question one way or the other would be useful, thanks...
kt

Katharine,

I accept the fact that there is not an absolute relationship between negative color and UV blocking with inkjet printers. However, on the whole my experience with several printers indicates that theoretical models based on expected color blocking are usually accurate.

In other words, if you start in the UV and violet of the spectrum, and go throug the blue, green, and red, you will generally find that the filter that blocks the most is the one most distant from UV. Red blocks UV light more than green and green blocks more than blue. The various pigment inks used do not agree with this 100%, but they generally agree.

So yes, by my own tests red pigment inks generally offer the most UV blocking.

Sandy

 

[clay]

FWIW, the new K3 Epson inks don't adhere to this generalization absolutely. In particular, the magenta is a pretty weak UV blocker, and is weaker than the cyan, which is a bit counter-intuitive to me. Attached is a spreadsheet showing the UV transmission density of each ink of the epson 7800 at various ink percentages. Make of it what you will.

Katharine,

I accept the fact that there is not an absolute relationship between negative color and UV blocking with inkjet printers. However, on the whole my experience with several printers indicates that theoretical models based on expected color blocking are usually accurate.

In other words, if you start in the UV and violet of the spectrum, and go throug the blue, green, and red, you will generally find that the filter that blocks the most is the one most distant from UV. Red blocks UV light more than green and green blocks more than blue. The various pigment inks used do not agree with this 100%, but they generally agree.

So yes, by my own tests red pigment inks generally offer the most UV blocking.

Sandy

 

[mkochsch]

FWIW, the new K3 Epson inks don't adhere to this generalization absolutely. In particular, the magenta is a pretty weak UV blocker, and is weaker than the cyan, which is a bit counter-intuitive to me. Attached is a spreadsheet showing the UV transmission density of each ink of the epson 7800 at various ink percentages. Make of it what you will.

Can you print the HSL-Array using your standard driver settings for your printer and post that as well? I'm interested to see if Yellow stays at the top or if the greens and reds produced by the Yellow-Cyans and Yellow-Magentas perform better or worse than expected compared to just the individual inks.

 

[mkochsch]

Onward: to to refine the blocking colour

I adjusted my initial blocking colour of 140H,100S,40B by examining and determining the output print that step 12 (12%) was actually where white started. I then went back to the 140,100,40 colourised chart and sample step 12 using the eye dropper. Its colour was 140H,79S,46B (RGB25,117,56). I then printed the a new ChartThrob using this as my blocking colour.
View attachment 105
Now Step Zero has been set to produce a very accurate white. I decided I wanted to get a better read on what was happening to the data so I scanned and hand curved the data. I used Excel to record the value from the tablet in one column and the value that printed in the next column. Then I opened up a new curve file in Photoshop and plotted the points. The first column of data is the Y-coordinate, is put in the "Output" box in Photoshop Curve and the second column (the readings obtained by sampling the scanned print) are the X-coordinate or the "Input" box in photoshop. I used "Smart Blur" to average the steps in the scanned tablet and a 5x5 sample in the eye-dropper to read the values, this just removes the odd bit of bad data caused by dust, chemical variance in developing etc. I also, set the white points and black point to 0 and 100 prior to sampling. This can/should be done because we know we won't be changing the blocking colour (step 0 dMin) and we won't be changing the exposure time (step 100 DMax.). I also noted that Step 30 in the uncurved tablet printed 50 per cent on paper. This is another important bit of information because when I'm building my curve I can set an anchor handle at Input 50, Output 30 in Photoshop and I know that this point shouldn't change or change very little.
The curve was produced and applied to the ChartThrob tablet again.
View attachment 106
With my mid-point still anchored at 50,30 I needed to get a touch more darkness in the print at around 80 to 90 per cent. I selected multiple anchors by holding the shift key and used the up arrows to shift that part of the curve up. I also needed to get a little less density in the negative around 10 to 20 per cent. Again I select anchors in this area and used the down arrow to bump them down a point or two. So, shift the curve up = removing density from the negative (print becomes darker), shift the curve down adds density to the negative (print becomes lighter). Next...the results.

 

[mkochsch]

Original Photo and Digital Neg Side by Side

The results speak for themselves. On the left the original of my daughter Isabel, on the right the print produced by the digital negative to Agfa MC paper using Pictorico OHP (I've got some Pictorico High Gloss White Film, I should give that a try too) developed in Tektol Standard by Silvergrain. The digital negative is somewhat softer but I attribute this to the fact I don't have a vacuum frame and use an old 8x10 contact printing frame.
View attachment 107 View attachment 108
~m

 

[mkochsch]

The Final Curve

And finally the curve. Notice it starts and finishes in or very close to the corners. Also 50 per cent is at 30 on the X-axis.
View attachment 109

~m

 

[Katharine Thayer]

The use of a particular blocking colour is for the sake of the negative and used to find an equal to the emulsion's highlight property that's all. e.g. Emulsion XYZed prints highlight at density 1.05, ink combo R255, G0, B64 also prints UV highlight at 1.05....therefore they match.

~m

Michael, I've deduced, by scratching my head a while, that you arrive at "emulsion xwz prints highlight at density 1.05" by counting the steps printed and multiplying by 0.15? So if the emulsion prints 7 steps, regardless of where those steps fall in the range of absolute print density, then the DR for the negative will be 1.05, which is another way of saying the emulsion "prints highight at density 1.05." Is that correct?

As before, I keep wanting to bring the actual print density in there somewhere; it seems like it needs to be part of the equation. Bear with me while I come to terms with the idea that the print density has nothing to do with anything. Well, of course the print density, the tonal range of the print in absolute terms, has a great deal to do with how the print looks, but I mean as far as generating the negative, the print density has nothing to do with anything; only the number of steps printed matters. Yes?
Katharine

 

[Katharine Thayer]

Edit

Since I don't have the "edit" function available to me, I'll edit this way. What I wanted to edit in the last post is to substitute the phrase "print tonal range" for "print density," which is ambiguous if not outright confusing when applied to gum, since the density of hardened gum is related to tonal range only in a relative sense within each print, not in an absolute sense across emulsions.

kt

Michael, I've deduced, by scratching my head a while, that you arrive at "emulsion xwz prints highlight at density 1.05" by counting the steps printed and multiplying by 0.15? So if the emulsion prints 7 steps, regardless of where those steps fall in the range of absolute print density, then the DR for the negative will be 1.05, which is another way of saying the emulsion "prints highight at density 1.05." Is that correct?

As before, I keep wanting to bring the actual print density in there somewhere; it seems like it needs to be part of the equation. Bear with me while I come to terms with the idea that the print density has nothing to do with anything. Well, of course the print density, the tonal range of the print in absolute terms, has a great deal to do with how the print looks, but I mean as far as generating the negative, the print density has nothing to do with anything; only the number of steps printed matters. Yes?
Katharine

 

[mkochsch]

Since I don't have the "edit" function available to me, I'll edit this way. What I wanted to edit in the last post is to substitute the phrase "print tonal range" for "print density," which is ambiguous if not outright confusing when applied to gum, since the density of hardened gum is related to tonal range only in a relative sense within each print, not in an absolute sense across emulsions.

kt

Print Tonal Range works for me. We could also just say "Print Values". It hard to use one term because of the differences between colour (where we need to discuss hue and saturation) and Black & White (where we just are concerned mostly with brightness).

~m

 

[mkochsch]

Updated link

Someone has pointed out that the link to the HSL-Array was pointing to the wrong file. Please note it has been updated on the page:

Dead Link Removed

~m