Digital negative tutorial

[tnp651]

I've posted a YouTube video with my technique for making a curve and printing a digital negative. I'd be interested in your comments.
https://youtu.be/nPoZt4-yzy4
Tom Nelson

 

[Dan Pavel]

Very nice and easy to understand tutorial. Thanks for sharing it.
My only comment is that in your final adjustments stages it would be easier to compare prints of the chart with the curve applied instead of prints of the actual image. In this way you'll have images with all the gray levels already marked to compare.

 

[Alan Edward Klein]

Tom, Thanks for your effort. You teach very clearly. I don;t do home developing but I am shooting medium format BW Tmax 100 negative and Velvia 50 color chrome film. I haven't printed yet but am interested in digital printing in a lab. Do you have any recommendations or tutorials for processing and scanning to get the digital file up to snuff so an outside lab could print in the best possible way? Thanks.

 

[tnp651]

Tom, Thanks for your effort. You teach very clearly. I don;t do home developing but I am shooting medium format BW Tmax 100 negative and Velvia 50 color chrome film. I haven't printed yet but am interested in digital printing in a lab. Do you have any recommendations or tutorials for processing and scanning to get the digital file up to snuff so an outside lab could print in the best possible way? Thanks.

Hi Alan, Photoshop has the tools you need. Make sure the scan captures the whole range of tones ( or scan light & dark versions and combine them). A pro-level lab might have a “Shirley” - a digital image they give away with a portrait reference and color patches. You can compare your image to the Shirley for brightness & saturation. You could also make your own test image.
The most important thing is to find the limits of tone that holds texture - light and dark.

(added later) There’s a big difference in the prints from consumer and pro labs. Consumer prints are contrastier and more saturated. That’s not necessarily a good thing. They’re more likely to blow out highlights or block up shadows, and to produce garish color.
Tom

 

[tnp651]

Very nice and easy to understand tutorial. Thanks for sharing it.
My only comment is that in your final adjustments stages it would be easier to compare prints of the chart with the curve applied instead of prints of the actual image. In this way you'll have images with all the gray levels already marked to compare.

I see your point, Dan. The problem with the second curve was that I picked the wrong area for the deep shadow. But I still like to see the effect on on actual image. I find it’s easier to see what the problems are. But others may differ.
Tom

 

[Dan Pavel]

But I still like to see the effect on on actual image.

Me too, Tom. That's why I am using my "Soft Proofing" method for my DNs.

For me a question still remains regarding the adjustment curves : why, even when you have your monitor/scanner properly calibrated, the adjustment curve generated by scanning a chart will not always give the best possible results? Why is often necessary to make additional adjustments? The way the printer uses different inks may be an answer, but even with Quad Tone Rip you have to manually readjust the curve. The problem may not be so big with some processes (like Palladium, Platinum, VDB) but it becomes annoying wit Gum prints, for example. None of the adjustment curves produced by different scripts give satisfactory results for Gum prints. Probably paper inconsistency or the difficulty in keeping the process perfectly reproducible... All these can't be seen in a 16 nodes curve, as is the PS curve, but look how a 48 nodes response curve of the Cyan layer of a CMYK Gum print really looks in detail:

The magenta regions of the curve are descending. No descending regions should appear in a response curve, otherwise the adjustment curve is not properly computable. All the scripts for computing the adj. curve are, in fact, using all the gray levels of the chart (resulting in an 101 values response curve, in the case of ChartThrobe) for computing the 14-16 nodes of the PS adj. curve. None can produce accurate results if they start from a response curve with descendant regions.

 

[nmp]

Me too, Tom. That's why I am using my "Soft Proofing" method for my DNs.

For me a question still remains regarding the adjustment curves : why, even when you have your monitor/scanner properly calibrated, the adjustment curve generated by scanning a chart will not always give the best possible results? Why is often necessary to make additional adjustments? The way the printer uses different inks may be an answer, but even with Quad Tone Rip you have to manually readjust the curve. The problem may not be so big with some processes (like Palladium, Platinum, VDB) but it becomes annoying wit Gum prints, for example. None of the adjustment curves produced by different scripts give satisfactory results for Gum prints. Probably paper inconsistency or the difficulty in keeping the process perfectly reproducible... All these can't be seen in a 16 nodes curve, as is the PS curve, but look how a 48 nodes response curve of the Cyan layer of a CMYK Gum print really looks in detail:
View attachment 235542
The magenta regions of the curve are descending. No descending regions should appear in a response curve, otherwise the adjustment curve is not properly computable. All the scripts for computing the adj. curve are, in fact, using all the gray levels of the chart (resulting in an 101 values response curve, in the case of ChartThrobe) for computing the 14-16 nodes of the PS adj. curve. None can produce accurate results if they start from a response curve with descendant regions.

One additional source of error is the scanner itself. In particular, there is an effect on the reading of a spot by the cross-reflections around it. So depending on the nature of the space around a particular block, the measurement will be different - by as much as 10% in L* value. Check this thread:

https://forum.luminous-landscape.com/index.php?topic=124885.0

Just in case you need to complicate this further....

:Niranjan.

 

[Dan Pavel]

Very interesting thread, Niranjan. Thanks for sharing.
Indeed, searching deeper makes things even more complicated.
How have you shaped your step-wedge in the end? Have you tried the scanner fixup program provided by Doug Gray?

By the way, a variant of your pre-curve method proved, in my tests, the only reliable method to get good (or at least satisfactory) curves for CMYK gum prints.

 

[nmp]

Very interesting thread, Niranjan. Thanks for sharing.
Indeed, searching deeper makes things even more complicated.
How have you shaped your step-wedge in the end? Have you tried the scanner fixup program provided by Doug Gray?

By the way, a variant of your pre-curve method proved, in my tests, the only reliable method to get good (or at least satisfactory) curves for CMYK gum prints.

I didn't delve into doing any post-scan corrections, but I did re-design my step wedge with each step embedded in a larger block of black. This gave me somewhat smoother curves. For machine-made papers like the Centennial POP I was working with at the time or the silver-gelatin process, this extra level of accuracy might be visibly beneficial. However, I figured for hand-coated processes, given that there are so many other sources of print-to-print variation, perhaps this might not be as important so I have gone back to using the generic step wedge. I am also playing with using a spectrophotometer to measure L*ab's and then converting to RGB and using those to calculate the correction curve.

To me, monochromes are more forgiving to less than perfect correction curve. On the other hand, I can see if you are doing 4 layers in CMYK gum prints, the colors can be quite a bit off if one or two layers are not accurate enough - like getting a cyan sky instead of blue....

:Niranjan.

 

[tnp651]

I just posted a video with a modified (and , I hope, easier to understand) technique:

 

[nmp]

I guess my question is why does ammonium dichromate need to be shaken? 5% solution has a plenty of solubility headroom (almost 35% at RT.) If it is falling out of the solution, may be something else is going on, no? Does it have a shelf life?

 

[tnp651]

Good question, Naranjan, but it's the only variable I could think of to explain my problem. The solution does have a somewhat grainy consistency.

 

[nmp]

Good question, Naranjan, but it's the only variable I could think of to explain my problem. The solution does have a somewhat grainy consistency.

Never having used pure dichromate myself, I wouldn't know quitehow a new "good" solution would look like, but yours in the video does look a little weird like a orange paint rather than a clear solution. If it works with shaking, then it's OK I guess. As an aside, given that you are using digital negatives, do you have to use the dichromate for contrast in kallitypes - can you not compensate for it in the negative itself?

:Niranjan.

 

[tnp651]

Hi Niranjan,
Negatives using inkjet on film have difficulty producing enough difference in density to produce a full-tone print. One solution is to add a "blocking color" to the dark parts, which will be the highlights. This is a color which is rather opaque to ultraviolet. Combined with the black ink, it adds density. But it's still a struggle. Adding dichromate to the emulsion adds contrast and allows the use of thinner negatives. I haven't tested omitting it but I suspect the result would be disappointing.

I tried doubling the amount of dichromate to see if it deepened the blacks. It didn't.

The ammonium dichromate I'm using is professionally-made by Bostick and Sullivan and is presumably the way it should look. I'm still working on my original bottle of the stuff, one drop at a time.
Tom

 

[nmp]

Hi Niranjan,
Negatives using inkjet on film have difficulty producing enough difference in density to produce a full-tone print. One solution is to add a "blocking color" to the dark parts, which will be the highlights. This is a color which is rather opaque to ultraviolet. Combined with the black ink, it adds density. But it's still a struggle. Adding dichromate to the emulsion adds contrast and allows the use of thinner negatives. I haven't tested omitting it but I suspect the result would be disappointing.

I tried doubling the amount of dichromate to see if it deepened the blacks. It didn't.

The ammonium dichromate I'm using is professionally-made by Bostick and Sullivan and is presumably the way it should look. I'm still working on my original bottle of the stuff, one drop at a time.
Tom

Hi, Tom:

Thanks for the explanation. Points well taken....it's a struggle I am well aware of. Another thing I can think of in favor of using a chemical "blocking" agent as opposed denser negative is inhibition of the "dark" reaction ( and the resulting background fog) which the negative, however dense it may be, have no influence on.

:Niranjan.

 

[tnp651]

BTW Naranjan, your prints on printing-out paper are lovely, and the technique masterful. Kudos!
Tom

 

[nmp]

BTW Naranjan, your prints on printing-out paper are lovely, and the technique masterful. Kudos!
Tom

Thanks Tom....I appreciate it. Feels like that was beginner's luck though. Lately I have not been as productive as I like to be so I can have some new material to display. Lots of learning and experimenting in different directions though, may be one of these days I will focus on one thing...like you do with kallitypes, which I admire a lot too.

:Niranjan.