Curves for gum negatives: Am I doing this right?

I'm starting to play around with gum bichromate. I just made a few quick digital negatives to start with and I'm not sure I'm doing it right.

Here is the original negative, note that the histogram is fairly well distributed.



If you invert this you get a negative but apparently maybe too much tonal range for gum.



Note that I don't have Photoshop but I've managed to create Christina Anderson's suggested curve for gum negatives in Lightroom. If I apply her curve (note the shape) to the positive image it looks like this:



So looking at the histogram for the positive it's basically compressed the tones into the shadows more. So if I invert that curve it ends up looking like this:



Looking at the shape of the curve does this make sense? With the negative tones compressed into the highlights (shadows in the print) the negatives are looking pretty thin which has me wondering if the shadows will block up. I wonder if I should try flipping the curve for a more dense negative.

Any thoughts or feedback welcome.

I fiddle with gum printing and although each printer and exposure system will give you differing results my impression is that the curve is inverted and should
be applied to the positive. See my example but please it is not to say this is good for your setup.
Just in case this is confusing on the left is before applying the curve shown, on the right is after applying with the preview box checked!

I use the same curve for Pt Pd as gum for the prints you saw in my space... I like the negative to have a bit of punch as the gum process on rag paper flattens the scene.. I have found the right negative can work for both silver, pt pd and gum with process adaptation.

I don't know how this works in Lightroom, but in Photoshop you must apply the curve, then flatten it before inverting and flipping. Is there an equivalent to this in Lightroom? Never seen the curve getting inverted.

In Lightroom I can only apply one curve to the image and there is no other way to invert the image. Hence my thinking in trying to invert the curve and use it that way. I'm pretty sure that's not the way to go based on my initial experiments last night. The best way to do this would probably be to scan the image as a positive so it's already inverted and then apply the curve in lightroom. My digital negs (as shown above) were definitely too thin and blocked up the shadows when printing

I'm finding that a more contrasty negative seems to work better. I had decent results with a negative that I had enlarged onto X-ray duplicating film which came out with fairly low contrast, but the best results were just from using a well exposed 8x10 negative that had plenty of contrast. The journey begins....

Also Bob, this is 100% your fault. After seeing your prints in person I am now obsessed with trying to make gum prints.

adelorenzo said:

In Lightroom I can only apply one curve to the image and there is no other way to invert the image. Hence my thinking in trying to invert the curve and use it that way. I'm pretty sure that's not the way to go based on my initial experiments last night. The best way to do this would probably be to scan the image as a positive so it's already inverted and then apply the curve in lightroom. My digital negs (as shown above) were definitely too thin and blocked up the shadows when printing

I'm finding that a more contrasty negative seems to work better. I had decent results with a negative that I had enlarged onto X-ray duplicating film which came out with fairly low contrast, but the best results were just from using a well exposed 8x10 negative that had plenty of contrast. The journey begins....

Also Bob, this is 100% your fault. After seeing your prints in person I am now obsessed with trying to make gum prints.

Click to expand...

I think this might your problem (no guarantee though):

If you want to apply a curve to the negative, you should "invert" the curve top-to-bottom NOT left-to-right as you seem to be doing. Take the original curve of C. Anderson and redo it by exchanging the input for output and vice versa. So it will look like a mirror image top-to-bottom. The two end points at (0, 0) and (255, 255) should not change. Then you can apply to the negative and print.

Looks like I might have another reason not to let go my Photoshop....

I work solely in PS for this reason... I make the image to the contrast I like and then use the curve in Print Tool designed by Ron Reeder. When making digital negatives using any ink jet machine one needs to Profile (Mark Nelson Method, Ron Reeder Method, Richard Boutwell Method) what you are basically doing is making sure that along the L range, what you see in your file is being represented on you film and due course represented on your paper of your process.
Without doing this then you are doing a hit and miss procedure and you would need to make a negative , then print on process, see where the problems are(clipped highlights, blocked shadows and curve shaping to control this.

In all cases I find it best to make your file look good on screen, verify the L number or K numbers or RGB numbers, so you are confident your not pushing areas in to clipping or blocking up ranges.
then you would apply the curve that you have made to apply to your file when you are sending to your printer.

I like PS so I can do all of this, I am a lightroom user but for this type of work , I do not think LR is up to the task.

When making secondary negatives for shadow hits in multiple prints... or splitting channels to make colour specific negative PS is where it is at.


Unfortunately there is a bit of technical wizardly movements needed to get you in a good place to start with.. Otherwise you are always making negs and adjusting to problems.

adelorenzo said:

In Lightroom I can only apply one curve to the image and there is no other way to invert the image. Hence my thinking in trying to invert the curve and use it that way. I'm pretty sure that's not the way to go based on my initial experiments last night. The best way to do this would probably be to scan the image as a positive so it's already inverted and then apply the curve in lightroom. My digital negs (as shown above) were definitely too thin and blocked up the shadows when printing

I'm finding that a more contrasty negative seems to work better. I had decent results with a negative that I had enlarged onto X-ray duplicating film which came out with fairly low contrast, but the best results were just from using a well exposed 8x10 negative that had plenty of contrast. The journey begins....

Also Bob, this is 100% your fault. After seeing your prints in person I am now obsessed with trying to make gum prints.

Click to expand...

Yes but when you are completely happy with your printing and prints then I am 100% to blame as well

I have a lot of info on You Tube as well PDF's in my website to help you on your way...making the negs once you are dialed in is the easiest part of the whole process.. I would dial in for Pt PD first as it is a predictable process and then making gum negatives are a breeze.

I've only got a copy of Photoshop Elements that came with my scanner. It doesn't do curves, alas, so it's not helpful. I just made some quick and dirty negatives to start playing with, I don't own a digital printer either so I just printed a few on transparency film at Staples. As mentioned I also have some large format negatives and some negatives I've enlarged in the darkroom. My plan is basically just to dive in and start mucking around, start seeing what works and doesn't and then try to dial things in. If I feel like things are going well I might have to invest in some more digital gear to make consistent negatives.

I have watched all of Bob's videos I would highly recommend them. Also got the Christina Anderson book. (For some reason it was only $19 on Amazon when I ordered it. It has since gone back up to regular price.)

nmp said:

If you want to apply a curve to the negative, you should "invert" the curve top-to-bottom NOT left-to-right as you seem to be doing. Take the original curve of C. Anderson and redo it by exchanging the input for output and vice versa. So it will look like a mirror image top-to-bottom. The two end points at (0, 0) and (255, 255) should not change. Then you can apply to the negative and print.

Click to expand...

That's what I was thinking as well. I want to have the tones compressed into the lower ranges rather than the higher ranges.Will give it a shot and see what happens.

adelorenzo said:

That's what I was thinking as well. I want to have the tones compressed into the lower ranges rather than the higher ranges.Will give it a shot and see what happens.

Click to expand...

On a second thought, never mind my "mirror" analogy. It is not quite right. Also I realize the input/ouput exchange method will not give an accurate result in comparison what you should get in Photoshop doing inversion after flattening. Hopefully it will be closer than the way Lightroom inverts the curve. Looks like this requires more investigation. As usual. there is more to it than meets the eye when it comes to digital negative...

nmp said:

On a second thought, never mind my "mirror" analogy. It is not quite right. Also I realize the input/ouput exchange method will not give an accurate result in comparison what you should get in Photoshop doing inversion after flattening. Hopefully it will be closer than the way Lightroom inverts the curve. Looks like this requires more investigation. As usual. there is more to it than meets the eye when it comes to digital negative...

Click to expand...

Looks like I shot from the hip (so what's new) in answering OP's question/concern and gave kind of first approximation answer which was, as it turns out, far from being mathematically accurate. On further investigation, here is what I came up with. (HINT: It looks like what OP is doing is correct after all.)

OP's original intent is to be able to take a known correction curve and use it in Lightroom where the capability to “flatten and invert” is non-existent. However, there is a capability ot “invert” a curve. The more familiar way is using Photoshop to first apply a Correction Curve on your positive image, flatten the image (very important) and then invert the image which becomes the corrected negative.

Let's say A and B are the input and output coordinates of a select point on the given Correction Curve. Then the Photoshop method will be apply the curve first – so the image value of A is converted to B. Next it is inverted which is nothing but substraction from 255. The inverted value then is (255-B). Thus A becomes (255-B) on the negative.

Now Lightroom process is to invert the curve and apply it to the positive image. If I understand correctly from the attached figures the inverted curve is nothing but a plot of A vs (255-B). The outcome is same as above.

Hope this does not require eating any more crow....

Debunk or validate?

:Niranjan.

I think this is probably a better curve, I have basically flipped and inverted it and now it looks like a lot better negative to try to print. This is still the same shape as the Christina Anderson curve shown in image #2 above.

Here is the curve file should anyone else want to use it