PDA

View Full Version : New darkroom calibration software.



Pere Casals
14-Jan-2019, 12:23
https://c1.staticflickr.com/5/4872/39776268813_073b8f6f3e_o.jpg


I'd request opinions and suggestions for a new software I'm developing (to be relesed as open source/freeware), intended to speed up darkroom technical procedures.

Motivation: a) To plot curves is boring, for DIY emulsion, BW slide reversal calibration, carbon calibration, paper calibration, paper contrast calibration for a RGB LED illumination of the enlarger, SCIM/CRM and calibrations for contrast color mask with Alan Ross method. b) Practicing a bit GDI+ I'll have to use, this was also a motivation.

Features:

1) Densitometer:

A pictorial Negative is scanned alongside the stouffer wedge, the red cells are dragged (mouse) on the stouffer, densities of the steps are in a configuration so the software can compare what gray levels are what densities, by moving the mouse to any spot on the negative we see what density is.

2) Curve families for Film calibration:

We place (wyswyg) the green cells on the contact copies osf the stouffer, we enter the exposure in LUX.S, then from densities in the stouffer we know exposure in each green cell, so the software knows the exposure in Lux·S and density for each step of each curve, a text file is generated with a data column for each wedge (All have same exoisure but each has a different development, see BTZS)

This file is imported in Exel so we can plot all curves instantly. So a film calibration requires scanning the contact copies of the stouffer alongside the stouffer itself, adjusting position of the step areas by dragging with mouse, and a click in the calibration button, and then data rows are in Excel, no densitometer, no manual data handling to make a calibration, just making an scan.

Until here it works nice yet, next is planned:

3) Paper calibrations.

4) Proofing silver gelatin paper result by selecting a paper calibration, a negative scanned alognside with a stouffer, and LUX.s exposure reading with lux meter on the easel without the negative. We can determine aproximate paper grade and exposure, this development step is necessari to to to the 5).

5) Generating a 3D LUT to be used in Ps, that will allow proofing the printing of a negative sandwiched with contrast color mask. In one Ps layer we have the image, with calibrated densities thanks to the stouffer that has been scanned with the negative. In the next layer we have the color mask for VC papers that we draw on the BW image. Nex is a applied the calibrated 3D LUT generated by this software that is to make the proofing, so wswyg the effect of the contrast mask.

The target is to make easy crafting a good contrast mask, sure some adjustments would be required in the darkroom when we see the real result on paper, but 90% of the work would be made without wasting paper.


Welcomed:

Any kind of critique, constructive or destructive. Ideas. Opinions about what components would be useful, possible additions. Volunteers to test the software while is developed to see how the thing goes.

I don't ask help for coding the software but that it would also be welcomed, PM for that.

IMHO a tool that boosts control in the darkroom would make optical printing more competitive, being easier to obtain sound results that now are only possible with hybrid system.

Well, perhaps this is also kind of hybrid process, because be make the contrast mask from digital, but at least we have an authentic optic projection crafting the image, helping with complex prints, but at same time allowing for adding any ammount of manual crafting, burning/dodging and the like.

Larry Gebhardt
14-Jan-2019, 13:19
Sounds interesting. A few suggestions:

- make work on mac, linux and windows
- put the code on github to enable collaboration
- support multiple step wedges
- support multiple step wedge layouts and density ranges

Pere Casals
14-Jan-2019, 14:00
- support multiple step wedges
- support multiple step wedge layouts and density ranges

It suports any wedge, you load a text configuration file of any wedge, just two rows, step number and calibrated density. When loading a Wedge the step count is updated, but it can we modified if we want:

https://live.staticflickr.com/4810/31802249967_be1a1e47a9_o.jpg

It also supports any pattern, with mouse we place the first and last position, and we can also change the rectangle XY area size where we average the reading:

https://live.staticflickr.com/4834/39778583383_e331286586_o.jpg



make work on mac, linux and windows
- put the code on github to enable collaboration


I'm a native windows programmer, perhaps in on github someone may help porting to mac/linux or making a cross platform app, I'm encapsulating in classes so it would be easier, perhaps dependence to gdi+ should be removed.

Tin Can
14-Jan-2019, 15:19
I would need a youtube tutorial.

Thanks!

neil poulsen
14-Jan-2019, 15:20
This is really interesting. Here are some of the calibrations in which I'd be interested. (Plus, recommended methodology to execute the calibration.)

A proper way for testing to use aperture adjustments to compensate for black and white film reciprocity failure, so that one need not rely on mfg graphs. (These are often not very complete.)

Testing for contrast grades on paper using a reflection densitometer.

If it helps, I use a particular routine in SYSTAT statistical software to plot my film curves; it works perfectly and doesn't assume any particular model. It's called "double-weighted least squares" analysis. All one needs to do to use it is enter two columns of numbers (zone number as the horizontal axis and densities measured using a transmission densitometer on the vertical axis) and say go. It does the rest. I've not taken the time to determine how this method relates to regular least squares analysis, but as I indicated, it works really well for these and many other types of curves.

Jac@stafford.net
14-Jan-2019, 15:33
This is really interesting. Here are some of the calibrations in which I'd be interested. (Plus, recommended methodology to execute the calibration.)

A proper way for testing to use aperture adjustments to compensate for black and white film reciprocity failure, so that one need not rely on mfg graphs. (These are often not very complete.)

Testing for contrast grades on paper using a reflection densitometer.

What ever happened to test strips or personal experience?

While not perfect what about Ilford's variable contrast heads and controllers for the dimwits?

Pere Casals
14-Jan-2019, 16:28
A proper way for testing to use aperture adjustments to compensate for black and white film reciprocity failure

This is supported yet, you know, this is a regular film calibration but we use long exposures with lower LUX for the contact copy. A way would be (with same development) exposing several contact strips with different intensity but with the proper shorter or longer time to have the same lux seconds, in the family of curves each would show what happens with different exposure time. We would have a graphs for each development showing different LIRF.

Another approach would be making graphs with a single exposure time in the graph but with each curve in the graph having a different development.



Testing for contrast grades on paper using a reflection densitometer.


This is the next that it will be added, there will be two wedges, the transmission one tells (from Lux.Second reaching the wedge) what exposure received each step in the paper. A second reflective wedge that would be scanned with the contact copied paper strips would be the reference to calculate the densities in the paper strips. Not much work from present development status.



routine in SYSTAT statistical software to plot my film curves;

Calibration output is a text file, (also for the moment there is a basic curve display working yet).

https://live.staticflickr.com/7843/46020506384_68f2ba665c_o.jpg

So you can import in Excel, in SYSTAT or many other. In this way one can plot and print curves with scales, colors etc in the way he likes, decimal separators can also be comma or colon.

Coulmns are Step ID, exposure in Lux·Second, H (Log exposure) and then there is a column for each strip densities. I should add a user custom label for each column for easy identification.




I would need a youtube tutorial.

Randy, beyond that, don't worry, if necessary we can use Teamviewer or Anydesk to solve it.

Bill Burk
14-Jan-2019, 19:04
I’d like to be able to read on my densitometer and enter values for each reading. (Support a ‘debugging’ mode if you would like to call it that).

In some cases users don’t know their Lux seconds. So it would be nice if you work backwards to it. The way I work backwards is to find the curve that met ISO/ASA and from the 0.1 speed point of that curve (and the box speed) deduce the Lux seconds. Also assume users will be a little inconsistent, don’t allow curves to cross but instead slide the exposure axis of misfits and report the exposure deviation.

Encourage users to spend the extra money on T2115C calibrated scales.

Bill Burk
14-Jan-2019, 19:08
Draw the ASA triangle. Show the selected points for contrast measures and support Kodak Contrast Index as well as other gamma, average gradient measures. Show a Time/CI curve.

Bill Burk
14-Jan-2019, 19:09
As well as support the Zone System N’s

Bill Burk
14-Jan-2019, 19:12
Import manufacturer’s data sheets and overlay expected vs actual results.

I hope this doesn’t sound like trolling, because I think these are all neat things to add and it seems like you are at the stage where you are fleshing out features.

Pere Casals
14-Jan-2019, 20:24
I think these are all neat things to add and it seems like you are at the stage where you are fleshing out features.

Bill, yes, you are rigth... at this stage I find important to realize what potential needs can be there, this would make easier to keep development open for what matters.



I’d like to be able to read on my densitometer and enter values for each reading. (Support a ‘debugging’ mode if you would like to call it that).

This is possible yet, as the tool is intended (for the moment) to source data colums to Excel (or other) then the values in the tables tables can be modified there. I find convenient (for the moment) making the tool extract the data and leaving complete freedom to user to organize data in Excel, each Excel tab may contain a different film for example with several graphs in it...

Densities can be checked in the tool itself, because by pointing any spot with the tool we see the calculated density.

https://c2.staticflickr.com/8/7835/45831378235_ce65d627e9_c.jpg

(Curves in this sample are arbitrary, I copied several times the same image of a single strip, simulating being 4 different, just for development, I made tests and density calculation by referencing to the stouffer is really fine.)




In some cases users don’t know their Lux seconds. So it would be nice if you work backwards to it. The way I work backwards is to find the curve that met ISO/ASA and from the 0.1 speed point of that curve (and the box speed) deduce the Lux seconds. Also assume users will be a little inconsistent, don’t allow curves to cross but instead slide the exposure axis of misfits and report the exposure deviation.


Well, as the LUX*Second exposure is introduced then by changing it all the graphs shifts left-rigt, so that could be done by selecting a point of the Normal curve to tell what's the meter point. OK, I understood that...

lux meters today are $20, but if not having it then that way will allows something better than the "arbitrary units".



Also assume users will be a little inconsistent, don’t allow curves to cross but instead slide the exposure axis of misfits and report the exposure deviation.

For the moments all strips share the same exposure to simplify interface, but in the data structure each strip has its own exposure, it would easy to allows to edit particular exposure for each strip, in a sort of advanced mode.

ok, a wyswyg edition to shift the curves...



Draw the ASA triangle. Show the selected points for contrast measures and support Kodak Contrast Index as well as other gamma, average gradient measures. Show a Time/CI curve.

As well as support the Zone System N’s

Ok... good ideas.



Import manufacturer’s data sheets and overlay expected vs actual results.

What would be easy is importing an image (sceenshot), to be scaled and placed under the graph, ok.



Thanks for te ideas pointing where development should target.

interneg
18-Jan-2019, 15:51
What ever happened to test strips or personal experience?


It's the same old belief/ delusion that has persisted for many, many decades that art in photography can be reductively turned into a form of applied maths that will hence produce a 'perfect' print. Which while it is fine (great, even) for industrial printmaking, it just doesn't understand the way that making a great print is about far more than sensitometric calculations.

Fundamentally, it's a worldview that denies the acquisition of aesthetic skill (art) & knowledge (science) & the practice thereof in favour of a techno-fetishistic belief that with all this software, gubbins etc you too can become a great printmaker...

Pere Casals
18-Jan-2019, 17:19
Which while it is fine (great, even) for industrial printmaking, it just doesn't understand the way that making a great print is about far more than sensitometric calculations.

Interneg, this is not the aim of this software. Since several years ago I'm addicted to Sally Mann and Denise Ross ways.

I'm thinking in this tool as a way to assist those (me the first) wanting learn advanced techniques and to also be useful to speed up boring technical tasks.

The last thing I want is being intrusive in the optical path. I'm a radical beliver in the purity of the optical crafting.

There is something that I don't like: positive discrimination for the optical process. To me it's wrong to give a higher value to an optical work because "it's more difficult". What need a positive discrimination are digital works, IMHO.

...but optical crafting of a print can be exhausting when a situation is complex.

I don't plan making great prints from great sensitometric calculations, at all. What I plan is making sensitometry easy, to reach the good point easier, faster and by wasting 1/4 of the paper.

What the print has to be it's in the photographer's mind. This tool it's not about the photographer's mind. This tool will be about making easier to craft in a print what's in the photograper's mind.

I hope that if you follow the evolution of the development then you may find that this is the intention.

faberryman
18-Jan-2019, 17:24
I am not sure what you plan to do with all this sensitometric data to help you print a negative. Is it to select paper grade?

Pere Casals
18-Jan-2019, 18:02
I am not sure what you plan to do with all this sensitometric data to help you print a negative. Is it to select paper grade?

Yes, we can load:

> a negative, calibrated with the stouffer, we know densities
> a film calibration, with several curves
> a mask (crm/scim) that is obtained from the negative + particular curve of the film calibration + exposure setting
> we also have the negative + mask sandwich like if it was a new negative
> a paper calibration, with different grades, a particular calibration is for grades from a color contrast masking of the Alan Ross process
> a color contrast mask, Alan Ross type
> a preflash layer

This is to make a proofing of how the print of a negative will be with a particular BW masking + Color masking + paper curve.

I don't plan that the result has to be exact, just it will show the path we should follow, allowing for example to know how to expose an scim mask to have more or less the effect we want, delegating in the manual crafting the refined adjustments.

BTW it plots fim/paper calibrations, and tells how to expose/process BW slides.

Most of the structural work is done, user interface development is slow and boring. In some two months it may be ready for testing, if someone wants.

For the moment you scan a calibration alongside the stouffer and it delivers data to Excel to plot the film curves, and works like a densitometer if scanning a negative with the stoffer. This weekend paper calibrations will be up. A paper calibration has 2 wedges, one is the trasmission one to say the exposure each step on paper received, and an additional reflective wedge scanned alongside the paper strips calibrates the resulting densities in the steps.

jp
18-Jan-2019, 18:44
Take my money!

(especially if it's freeware) I like the idea of this replacing a densitometer. People derided phones being replacement for incident meters and it has happened for many people. Using what people have is appealing compared to getting extra instruments.

The film testing part I mostly interested in testing different development and this would quantify those differences with very little work on the end users' part.

Sounds like #5 could be useful to quantify the effect of pyro stains.

From the screenshots, it seems like someone could run it on linux with wine if needed. I use Linux for work and surfing but it's not as practical for photo editing and scanning where Windows/Mac are the more versatile tools.

Pere Casals
18-Jan-2019, 18:51
The densitometer works perfectly yet, you only need an stouffer or similar. If you want to test yet it just PM.

Bill Burk
20-Jan-2019, 09:05
I am not sure what you plan to do with all this sensitometric data to help you print a negative. Is it to select paper grade?

It’s to help you evaluate whether or not you are getting the quality you want from your negatives.

And if not, what direction and how much to change development and/or exposure...to make them better next time.

Pere Casals
20-Jan-2019, 09:35
It’s to help you evaluate whether or not you are getting the quality you want from your negatives.

And if not, what direction and how much to change development and/or exposure...to make them better next time.

Bill, yes, these are the primary objectives, helping to be aware about how we are using the medium, and helping to get a solid criterion about what to do to refine our process, so we would spare energy to focus in on the subject or on the print aesthetics, of course one may not need that help from a software... anyway if wanting curves for something it does it with a few clicks and no densitometer

Bill Burk
20-Jan-2019, 10:12
You could make the software WYSIWYG by tying a scanned negative to sliders... and from the result (after user picks appearance they are going for) feedback new exposure and development times.

Bill Burk
20-Jan-2019, 10:15
Or go the other way to make it a What-if illustrator... have user slide their exposure and development times, print exposure and grade... and make it an exploratory experience.

Pere Casals
20-Jan-2019, 10:30
Or go the other way to make it a What-if illustrator... have user slide their exposure and development times, print exposure and grade... and make it an exploratory experience.

This is the way I find easy to implement.

As we have density of each spot in the negative, by entering a Lux.Second exposure and selecting a paper curve then we have the proofing, so we see what that paper grade/exposure does, we see the image in the monitor, what we also can check what predicted density will have any spot we point with mouse, to overcome a bit the monitor vs print look.

...but as we also have the film calibration the software also allows to place a scim/crm/etc mask or (Alan Ross contrast color mask) in the middle, so it helps to design the mask in a WYSWYG way, just it should allow to invest the effort in the refined adjustements without much waste of materials, replacing the try/error cycle by a digital proofing exploration of the possibilities we have.

As a subproduct it calibrates the BW slides reversal process to nail exposure/contrast, DIY emulsions, carbon and Pt/Pd...

Carbon has an easy solution with a digital negative made with black or white spots, but I like Jim's way, this is formulating the carbon content to adjust contrast as if it was paper grades.

The idea is to help the optical process from digital proofing, rather from digital manipulation. In that way a pure optic process would be competitive in refinement terms compared to an hybrid of digital way.

At least, it would help to refine process. Adjusting for example the own carbon traditional process would take a remarkable effort, because each try is very time consuming, while with an straight calibration/proofing we may speed up a lot our progression.

Tin Can
20-Jan-2019, 10:34
An line active version would be nice, like DOF Master (http://www.dofmaster.com/dofjs.html) which I have been using lately a lot.

Pere Casals
20-Jan-2019, 10:57
Randy, I was considering that, I also think it's best way. Problem is my capability as software developer, while I'm a proficient & really fast native code developer for RT and image procesing, I'm not good enough for the other. If this ends in useful tool then it would be possible the reuse most of the code for a PHP web project.

interneg
20-Jan-2019, 14:04
The aspects that amount to essentially an automated comparison densitometer are potentially useful as a means of refining exposure & processing, the rest is more questionable in terms of usefulness of making prints that are actually 'good' as opposed to 'accurate' - and this is my biggest criticism of the BTZS & Plotter/Matcher approach - it tells you a great deal of the 'what' in a scene but it seems to be often mistaken as telling the photographer 'how' in terms of providing a seemingly easy approach to making a print through technological means. In my experience it seems to have resulted in a great many technically perfect and utterly boring images. Understanding how to push around materials to creative ends is not difficult (if you put the hours in) & unless you have a database of analysed examples of where you want to end up, the system you describe is not going to be terribly useful in actually making spectacular prints. Same with generating masks. Making a really good print from LF isn't difficult, despite the excessive complexities & linguistics that some people like to dress it up in.

faberryman
20-Jan-2019, 14:14
It seems to me that futzing around writing programs, and using a scanner and computer on your negatives, just takes away time from making images and becoming a better printer through experience, if that is even the goal.

Pere Casals
20-Jan-2019, 14:46
the system you describe is not going to be terribly useful in actually making spectacular prints. Same with generating masks.

Interneg, I agree with you completely.

But let me reiterate that this tool is not intended to make spectacular prints. This tool is only intended to assist the photographer or the printer in the crafting.

Let me ask you something, have you ever adjusted process for BW reversal ?




It seems to me that futzing around writing programs, and using a scanner and computer on your negatives, just takes away time from making images and becoming a better printer through experience, if that is even the goal.

Frank, may be you are right. Karsh's printer had no computer, I guess. But they processed 250,000 sheets...

But I've several projects that require calibrations, some related to DIY emulsion for dry plates. Well, the software makes calibrations yet from a bare scan and a click, not that bad !