Making a Lightroom plugin to simulate darkroom techniques of a scanned negative

[cakescakescakes]

Hello Photrio,

I am considering building a lightroom plugin that can take a scanned negative, and then simulate what a B+W print would look like at each contrast grade. Before I begin, I would like to address a lot of the common responses I feel this might get:

1. Why do this? Lightroom offers everything you can do in the darkroom and more, with a touch of a button. Lightroom also was already designed with the darkroom in mind. Isn't this going full circle?
The purpose of this is more to practice workflows and darkroom thinking. Instead of being able to adjust the curve, you will have to play with exposure, grade, split grade, dodge/burn, toning, etc.

2. Why not just do it all in the darkroom the old fashioned way? If you are scanning it, whats the point?
The reality is that I, and many people, do not have permanent or free and unlimited access to a darkroom. Community darkrooms are expensive, and when paying for time, mistakes add up. Additionally if developing at home, you can get very comfortable overdeveloping or underexposing because you can get a good image with it in lightroom, whereas in the darkroom it could be challenging or impossible to get a decent print out of it. With this, you could get fast, quality, and free feedback.

3. I can make a contact sheet in one go, then based on that guess exposures and grades for all my prints. Maybe just get good?
That is wonderful for you .

4. This all depends on your enlarger, light source, paper, water ph, the phase of the moon, etc.
Assume that I will be standardizing everything on my enlarger (condensor) and a single paper type (ilford MG RC).

Ok, now onto the rest of the post. What I am envisioning is that from a scan of the negative, you can simulate what it would look like at each grade at different exposure levels. You can manipulate an exposure slider to see the shadows come in, then midtones, highlights etc, all the way to a way overexposed print. This way, you can easily see how your exposure and development was and will allow you to plan how to spend your time in the darkroom.

My main question is, what kind of knowledge will this take to accomplish? Will it require an understanding of densitometry? How do things like base fog come into play (and how do you measure it in a scan?) Will I need to include sprockets in the scan to get the pure light source? I'm guessing I will need to make standardized reference prints at all the different grades to base my conversions on, and then map that to the values of the scanned negative.

Any guidance on where to start would be appreciated. I'm a software engineer, so the programming will not be a problem. Thanks!

 

[loccdor]

What I like about copystand camera digitization is I can set a single shutter speed, aperture, ISO setting for a whole roll or multiple rolls of the same film. If the light source is constant that means I can make direct exposure comparisons without real variance.

Depending on how you're getting the scanner input you could find that the scanned image doesn't have a constant exposure. That would be the first big problem I would try to solve in this.

Fog, film base color can have an impact. When I do the digitization I like to find a blank frame on the roll and snap it at the exact same settings. Then I can blur it a little and overlay it on the other images with the "divide" filter. This gets rid of the base and also helps with unevenness in the light source.

 

[koraks]

My main question is, what kind of knowledge will this take to accomplish?

Just start with the curves in the datasheet of the paper of your choice. Using this, you should be able to get pretty far.

The main issue with your idea is what @loccdor also hints at above: you'll need to digitize the negatives and then output them with an absolute tonal scale - i.e. without any contrast adjustments. Apps like VueScan have provisions for this. Some brand-specific scanning software, such as Epson Scan, have something similar as well. You could also scan each negative along with a step tabled (Stouffer T2115 etc.) so that you have an absolute benchmark to start with.

The rationale is that if you apply the curves of the paper to absolute density scale of the negatives, you should be able to do a reasonably decent simulation of what an actual print would look like.

I'm guessing I will need to make standardized reference prints at all the different grades to base my conversions on

Not necessarily; you could do the above in 'theoretical space' all the way and get reasonably close. By means of validation, I would suggest doing some test prints of a step wedge at different grades and then compare that to the output of your to-be-made plugin/app.

Sounds like a nice project to play with; have fun!

PS: Welcome to Photrio!

 

[cakescakescakes]

What I like about copystand camera digitization is I can set a single shutter speed, aperture, ISO setting for a whole roll or multiple rolls of the same film. If the light source is constant that means I can make direct exposure comparisons without real variance.

Depending on how you're getting the scanner input you could find that the scanned image doesn't have a constant exposure. That would be the first big problem I would try to solve in this.

Fog, film base color can have an impact. When I do the digitization I like to find a blank frame on the roll and snap it at the exact same settings. Then I can blur it a little and overlay it on the other images with the "divide" filter. This gets rid of the base and also helps with unevenness in the light source.

Thanks for the reply! I plan on doing digital camera scanning with the valoi easy35 kit, which should make for consistent exposures. But the idea would be that anyone can use this, without having my specific camera scanning flow. This would pose a problem. Perhaps your method of using the film base can help, even for people doing scans with say, an epson flatbed?

 

[cakescakescakes]

Just start with the curves in the datasheet of the paper of your choice. Using this, you should be able to get pretty far.

The main issue with your idea is what @loccdor also hints at above: you'll need to digitize the negatives and then output them with an absolute tonal scale - i.e. without any contrast adjustments. Apps like VueScan have provisions for this. Some brand-specific scanning software, such as Epson Scan, have something similar as well. You could also scan each negative along with a step tabled (Stouffer T2115 etc.) so that you have an absolute benchmark to start with.

The rationale is that if you apply the curves of the paper to absolute density scale of the negatives, you should be able to do a reasonably decent simulation of what an actual print would look like.



Not necessarily; you could do the above in 'theoretical space' all the way and get reasonably close. By means of validation, I would suggest doing some test prints of a step wedge at different grades and then compare that to the output of your to-be-made plugin/app.

Sounds like a nice project to play with; have fun!

PS: Welcome to Photrio!

Thanks for the reply! When you refer to "the curves in the datasheet of the paper of your choice", I assume you are referring to the characteristic curves given here: https://www.ilfordphoto.com/wp/wp-content/uploads/2021/01/MULTIGRADE-RC-Papers-J20.pdf.

But I see a few problems with this:
1. For grades 00 to 3, I can't even tell which line is which on the left hand side
2. How do I map the density value given (what is the unit for the y axis also?) to the luminance value in lightroom?

 

[koraks]

I assume you are referring to the characteristic curves given here: https://www.ilfordphoto.com/wp/wp-content/uploads/2021/01/MULTIGRADE-RC-Papers-J20.pdf.

That sort of thing, yes.

1. For grades 00 to 3, I can't even tell which line is which on the left hand side

You can work that out on the basis of the assumption that the curves will not be significantly kinked in the midpoint where they intersect.
If those particular curves don't work, you could always make your own using a scanner, reflection densitometer or photospectrometer and a step tablet.

2. How do I map the density value given (what is the unit for the y axis also?) to the luminance value in lightroom?

Note that the horizontal axis is defined as 'relative log exposure'. I.e. each (ca.) 0.3 unit change on the horizontal axis corresponds with one stop of more/less exposure. You can then correlate this to the density range of your negative (or step tablet). In doing so, you'll have to contend with problems like the Callier effect, but I'd suggest ignoring this for now and see if you can get a basic approach to work. Then dress it up with nuances as you see fit.

 

[jeffreyg]

What you are discussing is way above me but you may want to check out On 1 software. They have a number of presets that may already do some of the things you plan.

 

[loccdor]

Yes, if you want it to work for users of various input devices you'll need an initial calibration step of some kind. The scanned image you're working from will have to be relationally mapped to the density of the negative. An f(x)=y function. If you for example varied the light source power in your scanning setup while keeping all other things equal, this function would need to recalibrate. But this recalibration could then be applied across all images with this difference.

And if differences in enlargers are considered, you may need a step to calibrate to the enlarger as well. I don't know all that much about them, I only ever used 1 model and in a relatively simple way.

If I were going to try this I would start with some practical tests by making some graphs for the f( x )=f( y )=z mapping while recording all settings I used. X can be the density of a point on the negative, y can be the value of a pixel on the scanned negative, z can be the value of a pixel on the scanned print. Keeping all variables constant, other than your x, and the contrast grade, would be very important.

Koraks knows much more about the photographic science side of things. I have a computer science degree, and took an image processing class, and have worked with a few image processing libraries a while back. What I like about programming projects of this type is they teach you how the thing works while you code it - for example there were huge gaps in my knowledge about how a basic 3d engine works until making one. But now that I did, I see it all boils down to our earliest discoveries regarding the triangle and the circle.

 

[cakescakescakes]

Thanks everyone! I think I have enough to get started. I'll post an update when (and if) I have one.

 

[Sirius Glass]

Welcome to Photrio!

LightRoom Classic does not put your negatives on the cloud. That allows one to maintain control of their negatives.