Early impressions of Pt/Pd, diginegs and curves

[poliweb]

I have only just begun my journey of discovery in both Pt/Pd printing and digital negatives. Going into this I had high hopes and expectations of both aspects. I can see from posts that you have all been where I am now, and many of you have pioneered and continually improve the techniques I am trying. I'd like to share with this forum a few of my initial impressions and observations and my reactions to them in the (rather selfish) hope that you can help accelerate me along my journey - or, in some cases, perhaps reset my expectations.

I will admit up front that I am somewhat of a sharpness addict - having originally jumped into large format photography by the lure of the quality it can produce. I now work all digitally (current Pt/Pd printing excepted) and still strive for a similar "look". From an aesthetic perspective, I started in black and white (silver) and found that my style was more suited to color. I started in landscapes and found my passion lay with architecture. However, my current photography tends towards very subdued colors with many of my images having a rather monochromatic look. As a result I have found myself lured back to black and white, and have used that as a springboard for me to try Pt/Pd printing.

From a techincal perspective I am working with Pt/Pd in a 1:3 ratio with just a couple of drops of FO #2 in the emulsion to eliminate a slight fog I found without restrainer. My only paper right now is Cranes Weston Diploma. I am using straight Ammonium Citrate for the developer and EDTA for clearing (all from a B&S kit.) I have used the HSL-RNP array and ChartThrob to select a color and produce a curve. I am printing diginegs on Picotrico OHP with an HP B9180.

Here are some of my observations and questions:

My first print disappointed me in a couple of areas - it was not as sharp as I would have liked and I could detect mottling (from the inkjet dots) in the gray areas. The first issue seemed to be related to the paper (Cranes Weston Diploma) - although fairly smooth out of the box, it gained some texture after processing. This leads to my first set of questions:

• Are there smoother papers than Cranes Weston Diploma available?
• Is Bergger Cot 320 smoother?
• Is there anything at approaches the finish of traditional fiber-based B&W paper?

The second issue seems more complex. At first I thought that the mottling might be a function of printing in negative - meaning that the dot pattern of the printer that is designed to build up density for a dark area of the print was now being used to hold back UV light to produce highlights. Perhaps this was exacerbating the issue. However, some scans of positive and negative images printed on OHP convinced me that the differences we not that significant. My experiments did, however, lead me to one source of the problem - the curve.

Below are two images: on the left a section my digineg, with curve applied, scanned and inverted and then adjusted to provide full contrast.; on the right is the same section, printed using the same blocking color, but with no curve applied, then scanned and inverted:

View attachment 196

As you can see, the effect the curve has had on the image quality is huge. The compression and subsequent re-expansion of the tones has done great damage. For reference, here is the smoothed curve I used for my blocking color of (255,64,0):

View attachment 197

I am still experimenting with trying to find colors that have a longer scale. This leads me to my second set of questions:

• Has anyone seen similar issues with negative quality due to the curve?
• Have others found colors that require less dramatic cuves?
• Would going pure Pd help?

I have tried some greens and black only, both composite and gray inks only, but with no better results in terms of curves.

Any suggestions, observations and advice would be very welcome,

Thanks

Richard

 

[sanking]

• Has anyone seen similar issues with negative quality due to the curve?
• Have others found colors that require less dramatic cuves?
• Would going pure Pd help?

I have tried some greens and black only, both composite and gray inks only, but with no better results in terms of curves.

Any suggestions, observations and advice would be very welcome,

Thanks

Richard

1. Sure, the shape of the curve is essential in balancing input and output. Did you develop your own curve or adopt one from someone else?

2. The best color is one that 1) matches the ES of the process, and 2) with a given printer gives a linear progression of densities from Step 1 to 255.

3. Going to pure palladium would simplify your life. If necessary you could increase contrast and get rid of fog by adding a very small amount of dichromate to the developer, or by adding some Na2.

What we normally due in developing digital negatives is first determine the UV blocking color that matches the ES of the process, then devleop the curve with a method such as PDN or ChartThrob. Choosing the correct color will allow the least abrupt or dramatic curve.

With the HP 9180, assuming you are using pure palladium, the best color layer IMO is G=255, B=20. This color will work for pure palladium where the ES is about log 2.5 or slightly higher. I don't recommend the use of black inks with this printer since the UV transmission densities are much less linear than with green.

Sandy King

 

[poliweb]

1. Sure, the shape of the curve is essential in balancing input and output. Did you develop your own curve or adopt one from someone else?

2. The best color is one that 1) matches the ES of the process, and 2) with a given printer gives a linear progression of densities from Step 1 to 255.

3. Going to pure palladium would simplify your life. If necessary you could increase contrast and get rid of fog by adding a very small amount of dichromate to the developer, or by adding some Na2.

What we normally due in developing digital negatives is first determine the UV blocking color that matches the ES of the process, then devleop the curve with a method such as PDN or ChartThrob. Choosing the correct color will allow the least abrupt or dramatic curve.

With the HP 9180, assuming you are using pure palladium, the best color layer IMO is G=255, B=20. This color will work for pure palladium where the ES is about log 2.5 or slightly higher. I don't recommend the use of black inks with this printer since the UV transmission densities are much less linear than with green.

Sandy King

Sandy,

Thanks for the response.

1) The curve was derived using ChartThrob.

2) I guess that's the root of my problem right now - finding a color that matches my ES in a linear way - so that I can get a more linear curve.

3) I will try some pure Pd - I don't have any dichromate, so I can only use a bit of FO #2 in the emulsion until I get some. Same goes for Na2. I'll try your green with it and see what I get.

So does your curve from ChartThrob look much less dramatic?

Thanks again,

Richard

 

[donbga]

Sandy,

Thanks for the response.

1) The curve was derived using ChartThrob.

2) I guess that's the root of my problem right now - finding a color that matches my ES in a linear way - so that I can get a more linear curve.

3) I will try some pure Pd - I don't have any dichromate, so I can only use a bit of FO #2 in the emulsion until I get some. Same goes for Na2. I'll try your green with it and see what I get.

So does your curve from ChartThrob look much less dramatic?

Thanks again,

Richard

Richard,

To answer your question about paper, COT 320 is a smoother paper though the Cranes Diploma Parchment is a very economical paper to use when learning to build curves. How are you coating the paper?

There aren't any papers I have used or know of that have the apperance of air dried FB paper for alt. process printing though Craig Koshick <sp?> has printed on fixed out FB papers sucessfully.

If possible don't use FO#2, use NA2 instead.

Hope this helps,

Don

 

[sanking]

Sandy,


So does your curve from ChartThrob look much less dramatic?

Thanks again,

Richard

Richard,

Attached is a curve for palladium I developed for the HP B9180 using Green = 255, B = 20. The curve should be applied to a positive image.


Sandy

 

[sanking]

Richard,

BTW, it looks to me that your curve may be reversed. If you made it with ChartThrob the curve should be applied to a positive screen image, but this looks like it should be applied to a negative.

Have a look at the attached curve which was made with the PDN system, in which the curve is applied to a negative screen image.

Sandy King

 

[poliweb]

Richard,

BTW, it looks to me that your curve may be reversed. If you made it with ChartThrob the curve should be applied to a positive screen image, but this looks like it should be applied to a negative.

Have a look at the attached curve which was made with the PDN system, in which the curve is applied to a negative screen image.

Sandy King

Sandy,

My curve is applied to the positive - if you look at the axes you will see that mine are reversed from yours - I have the black near the origin. So it's correct.

Your curve does look less extreme though - I can see that it would cause fewer problems than mine. I'll have to see if I can get there.

Richard

 

[poliweb]

Richard,

To answer your question about paper, COT 320 is a smoother paper though the Cranes Diploma Parchment is a very economical paper to use when learning to build curves. How are you coating the paper?

There aren't any papers I have used or know of that have the apperance of air dried FB paper for alt. process printing though Craig Koshick <sp?> has printed on fixed out FB papers sucessfully.

If possible don't use FO#2, use NA2 instead.

Hope this helps,

Don

Don,

Thanks for the comments. Fixed out FB paper does sound interesting - maybe worth a try.

I am coating my paper with a puddle pusher from B&S. Though I just got a Grumbaucher Gold Edge brush to try.

And I promise that my next order from B&S will have Na2 in it and no more FO #2. The feedback on that one has been unanimous so far!

Richard

 

[donbga]

Richard,

BTW, it looks to me that your curve may be reversed. If you made it with ChartThrob the curve should be applied to a positive screen image, but this looks like it should be applied to a negative.

Have a look at the attached curve which was made with the PDN system, in which the curve is applied to a negative screen image.

Sandy King

Sandy,

The way the curve is displayed depends on the orientation of the highlight/shadow values. I use mine set opposite of yours and thus the curve lies below the diagonal when applied to a negative image.

Don

 

[sanking]

Sandy,

The way the curve is displayed depends on the orientation of the highlight/shadow values. I use mine set opposite of yours and thus the curve lies below the diagonal when applied to a negative image.

Don

Don,

Yes, I see that now.

Thanks,

Sandy

 

[mkochsch]

View attachment 200

I tend to agree with you that the curve is too steep and we could do better. The uncurved example has a lot of play in the highlight end and is probably "floating" a bit. We're only seeing a small portion of it but I'm guessing the shadows are a bit on the muddy side. It's sure nice and smooth though. I'm also seeing what looks like head banding in the curved sample. I'm not sure how the curve could cause that though.
I went back and looked at your original HSL test and that area of colour chosen rapidly goes from white to grey into the first two blocks. I think there are better areas where your driver and inks feather the scale out which you can use to your advantage. Note the area I circled from the original post. It's a transition area (green) from high (red) to low (blue) blocking. The scale sweeps "diagonally" in this area. Count the number of blocks between the white and the black as they progress vertically upward. Maybe 8 in your choice, but at least double that in the green colour. I guestimated an initial colour of 16, 180, 0 which is, not so strangely, in the ball park to what Sandy has also found.
Try that colour (or possibly even a darker colour such as 0,90,30) as your first fill and then do the second iteration to find the exact blocking colour on this colour path. I might be a little on the dense side but I can't say exactly from the sample you initially posted so bear with me. I think the key point here is finding a colour hue that really has a long scale flattens out the curve and makes it less drastic.
~m

 

[poliweb]

View attachment 200

I tend to agree with you that the curve is too steep and we could do better. The uncurved example has a lot of play in the highlight end and is probably "floating" a bit. We're only seeing a small portion of it but I'm guessing the shadows are a bit on the muddy side. It's sure nice and smooth though. I'm also seeing what looks like head banding in the curved sample. I'm not sure how the curve could cause that though.
I went back and looked at your original HSL test and that area of colour chosen rapidly goes from white to grey into the first two blocks. I think there are better areas where your driver and inks feather the scale out which you can use to your advantage. Note the area I circled from the original post. It's a transition area (green) from high (red) to low (blue) blocking. The scale sweeps "diagonally" in this area. Count the number of blocks between the white and the black as they progress vertically upward. Maybe 8 in your choice, but at least double that in the green colour. I guestimated an initial colour of 16, 180, 0 which is, not so strangely, in the ball park to what Sandy has also found.
Try that colour (or possibly even a darker colour such as 0,90,30) as your first fill and then do the second iteration to find the exact blocking colour on this colour path. I might be a little on the dense side but I can't say exactly from the sample you initially posted so bear with me. I think the key point here is finding a colour hue that really has a long scale flattens out the curve and makes it less drastic.
~m

Just to reiterate - the uncurved sample was just a scan of an uncurved negative - I never printed it. But it does give an idea of how much damage the curve is doing - and motivation for me to get it more linear.

There is some banding on the curved example - not sure how that go there either.

I have done a new HSL test on my (mostly) unrestrained Pd/Pt emulsion. It's essentially the same, with slightly longer transitions from white to black. The area you point out is filled with a "mushy" light gray - no definite squares, just a general fog. It doesn't show up too well on the scan. I avoided this area because of that - but maybe I should give it a try. I would have to go darker to reach a color not in the "mush". But I will give a dark green a try.

Yes - finding a color with a long scale seems to be the holy grail.

Richard