DIY Densitometer?

[laroygreen]

Hi all,

I need a transmission densitometer, but the prices online for 30 year old "untested" gear has me thinking; we have access today to so much technology, couldn't I just build one using a Raspberry PI with a camera module, a light source and software to capture a calibration image for comparison against readings? Wouldn't be ideal for researchers that need to share findings against a known standard, but I only need it for my darkroom and my materials. Maybe a project like this exists already and if so, please share.


Full disclosure - I've never seen or used a densitometer. I am working entirely in the realm of how I think it works, so please correct me!

 

[Photomultiplier]

You have to consider the dynamic range of the photodetector, whatever sensor type it may be.

The dynamic range is the difference between full brightness and the highest density you want to be able to measure. For slide film, the maximum density (Dmax) can go up to almost 4

https://asset.fujifilm.com/master/emea/files/2020-10/2f3c7f90a0b0c6e605e84f98b7d489c2/films_velvia-100_datasheet_01.pdf

which means 10^4 = 10000:1 contrast.
10^3 would be 1000:1

Negative film is not that dense and goes up to around 2 (100:1)
https://www.ilfordphoto.com/amfile/file/download/file/1919/product/687/

So your sensor and circuit must be able to handle those differences. It must not be driven into saturation at density=0 and the noise floor must be low enough to give
resonable accuracy down in the dense areas.

100:1 would be a bit less then 7 bits which is pretty resonable
1000:1 -> about 10bit (2^10 = 1024)
10000:1 -> about 13.4bit

I also did some experiments (with photodiodes) in the lab, but i need it for slide film. (high Dmax)

If you just need it for negative film, the requirements are more relaxed.

Do you have a step wedge? They can be used to quickly test the performance of photodetectors and the associated optics.


For a darkroom print analyzer, i made a breadboard prototype with some integrated light sensor that puts out a frequency proportional to the light intensity.
If you are interested, i can dig it out and look what type of sensor it was.

Camera module sounds also interesting, but one has to consider stray light and so on.

 

[ic-racer]

I used my spot meter and a light table for 20 years. Works great.

 

[laroygreen]

I would need it specifically for transparency. Didn't consider dynamic range and I am sure such a small sensor would probably struggle with noise in high density. There is also a light detecting resistor, but I haven't looked at the specs. Already sounds like a lot of work with no guaranteed payoff.

It would be cool if a github repo existed (maybe one does?) that could capture all these designs from the experts here where others can adapt and iterate on. As a programmer its the first thing that comes to mind when I think about stuff like this. Reading some of these old photo / darkroom books from the 70s and 80s, and it seems that building things for the darkroom was the absolute norm.

@ic_racer I do have a spot meter but I got the impression that this method wasn't as reliable, but maybe I was misinformed? I need it to calibrate my enlarger as well as for testing separation negatives and pigment-gelatin prints (on transparent base). I am not looking for more work than I need, so if that can work I am all for it.

 

[nmp]

How about using a scanner:

Sign in - Google Accounts

:Niranjan.

 

[fgorga]

Take a look at https://www.dektronics.com/ and the posts here (c.f. https://www.photrio.com/forum/threa...-budget-friendly-densitometer-project.185429/) describing the Printalyzer densitometer. (Although I do note that it is currently listed as out of stock. The first batch must have sold out quickly.)

This device make both transmission and reflective measurements.

I believe that most, if not all, of this project is open source if you want to build it yourself. However, I consider the $250 price tag for a complete, ready-to-use unit very fair.

I ordered one on 12 November and it arrived quickly. I have not made extensive use of it yet but it works as expected.

 

[albada]

Take a look at https://www.dektronics.com/ and the posts here (c.f. https://www.photrio.com/forum/threa...-budget-friendly-densitometer-project.185429/) describing the Printalyzer densitometer. (Although I do note that it is currently listed as out of stock. The first batch must have sold out quickly.)

This device make both transmission and reflective measurements.

I believe that most, if not all, of this project is open source if you want to build it yourself. However, I consider the $250 price tag for a complete, ready-to-use unit very fair.

I ordered one on 12 November and it arrived quickly. I have not made extensive use of it yet but it works as expected.

I bought one as well, and love it.
I suggest the OP place an order and wait for the next batch to be built. This device is better than any DIY kludge.

 

[gone]

Full disclosure - I've never seen or used a densitometer.

Me either, and let's do our best to keep things just as they are. It's like the old Country Western song lyric, "How can I miss you if you won't go away?" Except this is, how can I miss something that I've never had?

 

[albada]

I notice you have an enlarger.
In that case, you can use an easel meter such as this one made by DarkroomAutomation.com which sells for $110. To measure density, turn on the enlarger, put the meter into delta mode, zero the meter to that light, and put a spot on the transparency over the meter's sensor. The meter tells you the relative density of that spot in stops (multiply by 0.301 for log-density).

 

[koraks]

I also did some experiments (with photodiodes) in the lab

Ah, doing it the hard way
It's feasible to get the necessary contrast range with photodiodes. I'd start with making the light source variable in intensity (PWM) and using an integrating amplifier on the photodiode. Since you have control over the brightness of the light source (typically an LED, or several for different color measurements), it's fairly easy to accommodate whatever contrast range desired, within a somewhat reasonable range. For the integrating amplifier the old but still usable IVC102 is convenient.

But at this day and age I'd do what @dkonigs g also did and get one of the I2C-connected sensors from AMS. They have the diodes (an array of them) and integrating transimpedance amplifiers integrated into them with all the necessary temperature compensation etc. I've done some work with the AMS7341, which is just one of the several available options. I'm not sure which sensor Derek ultimately ended up using; a quick peek at the schematic files suggests TLS2591. This one is color blind, but with RGB LEDs as the light source you can still probe different color channels sequentially. I don't think the Printalyzer does that, so for color work, another version could be made. Perhaps Derek is considering this; I don't know.

If a color-blind measurement is acceptable, I'd personally be hesitant to reinvent the wheel and just get a Printalyzer. It's proven technology that's conceptually sound.

 

[laroygreen]

The PrintAnalyzer would have been THE solution but I specifically need to analyze colour when making separation negatives and seems this can only do BW (I should have stated this more clearly, my apologies).

I've never done any of this before, I am simply following my reference books at this point, but seems this is a must. Seems I may have to buy one that is specd to my needs as building one (at least for now) isn't a priority while I am starting to learn a new process.

(from Colour Photography in Practice)

@albada Thank you for sharing. I will get an easel meter but for a different purpose. Might be worth a try but I feel I would struggle with consistency and I would also have to use filters (I think) to read the correct density for each separation.

@nmp I am trying to avoid using a computer and scanner when working in my darkroom, so this wouldn't work in my case. Sure it would be helpful for someone else, so thank you for sharing.

 

[dkonigs]

But at this day and age I'd do what @dkonigs g also did and get one of the I2C-connected sensors from AMS. They have the diodes (an array of them) and integrating transimpedance amplifiers integrated into them with all the necessary temperature compensation etc. I've done some work with the AMS7341, which is just one of the several available options. I'm not sure which sensor Derek ultimately ended up using; a quick peek at the schematic files suggests TLS2591. This one is color blind, but with RGB LEDs as the light source you can still probe different color channels sequentially. I don't think the Printalyzer does that, so for color work, another version could be made. Perhaps Derek is considering this; I don't know.

If a color-blind measurement is acceptable, I'd personally be hesitant to reinvent the wheel and just get a Printalyzer. It's proven technology that's conceptually sound.

The current version of the Printalyzer Densitometer indeed uses the TSL2591, combined with a UV-IR cut filter and a "white" LED. This does a pretty good job of measuring approximately around what's known as the "visual" range (~570nm, somewhat broad width), which seems to be what's commonly used for checking density of B&W materials. (the filter is really just to cut out stray light from outside of the target spectrum, which can throw off readings)

Using that sensor for color density, by changing the light source, probably won't work so well. The wavelengths you need to measure for blue is kinda near its limits on that side. You might have more luck with red, provided you can dial everything in.

Regardless, I am currently trying to crack the problem of doing color density using the AMS AS7343 sensor, but its a much more difficult problem. (the AS7343 is a newer version of the AS7341, which adds a few more channels and has higher gain settings) But what makes that a lot harder is that I first need to select an LED light source that does a much better job at "full spectrum" performance (Yujileds and Seoul Semiconductor make some, but the majority have this really big dip/spike in the middle of the blue range.). Then there's quite a lot of calibration that needs to happen (channel wavelengths, many sensor gain settings, etc). Finally, having good comparison values is far easier said than done, given that a lot of existing off-the-shelf color densitometers seem to go a bit wonky on their blue channels with age. Ultimately, this is going to require more sophisticated lab equipment and building a few prototypes.

In any case, right now priority #1 is getting the next batch of the current Printalyzer Densitometer version build so I can restock inventory. It is being worked on, but obviously there have been a lot of interruptions this time of year. Fingers still crossed on mid-December for that being ready, though.

 

[koraks]

But what makes that a lot harder is that I first need to select an LED light source that does a much better job at "full spectrum"

How about 3 discrete LEDs in suitable wavelengths and a small light mixing solution? This might be as simple as a bit of frosted plexi and a little tube for a makeshift collimator.

 

[Anon Ymous]

How about 3 discrete LEDs in suitable wavelengths and a small light mixing solution? This might be as simple as a bit of frosted plexi and a little tube for a makeshift collimator.

Perhaps 4 leds with the inclusion of an IR one. This could be used to assess bleach effectiveness in C41/E6 films.

 

[koraks]

Yeah, might be a good idea. They don't have to be very high power either; with a little bit of luck a couple of 0603 and LEDs could be mounted right next to each other, making a compact light source with just the right wavelengths. And also pretty good wavelength binning/separation most likely.

 

[laroygreen]

Any recommendations for a new transmission densitometer? The Tobias TBX1000C appears to be what I need, but I can't actually find it for sale anywhere and I have no idea what a new one would cost.

@koraks what densitometer do you use or are you a monochrome carbon transfer printer? I

 

[laroygreen]

I found a website selling the Tobais TBX 1000 for .... $3,105.00 USD with $354.00 USD standard shipping via UPS

 

[koraks]

@koraks what densitometer do you use or are you a monochrome carbon transfer printer?

None. I am in the process of setting up a color workflow, but using digital separations, so no densitometer is required. I do use a photospectrometer.

 

[laroygreen]

None. I am in the process of setting up a color workflow, but using digital separations, so no densitometer is required. I do use a photospectrometer.

Had to watch a video on what that is . Very cool! I am guessing this will enable you to test your glop for consistency, adjusting your pigment load to get consistent absorption across red, green and blue light?

 

[dkonigs]

Any recommendations for a new transmission densitometer? The Tobias TBX1000C appears to be what I need, but I can't actually find it for sale anywhere and I have no idea what a new one would cost.

@koraks what densitometer do you use or are you a monochrome carbon transfer printer? I

As I was interested in the same, to at least give me an "official" reference for my projects, I sent a message to the contact Email address on the Tobias website.

The response I got was that the TBX1000C was $3485 for Status M readings, with an additional $286 for the filter to get Status A. Delivery in 6 weeks from order.

I haven't gone further down that path as of yet, but at some point I might have to. Those old X-Rites are a lot cheaper and easier to get on the used market, but I'm seriously suspicious of the accuracy of their blue channels due to age.

In any case, for those who wonder... Status A is typically used for measuring positives, while Status M is typically used for measuring negatives.

 

[koraks]

I am guessing this will enable you to test your glop for consistency, adjusting your pigment load to get consistent absorption across red, green and blue light?

No, it's mostly for things like determining exposure for different tissues (colors) and linearization of negatives. And later on also color profiling. In that sense it's a sort of equivalent of a TRD in this digital/hybrid workflow. It's a cool piece of kit; I'm using it more than I had expected.

 

[laroygreen]

I'll take a risk and get one off ebay. I'm just starting out with this process and I don't know what I don't know; i.e. any print would be a good print! If the densitometer turns out to be a limiting factor then I'll just have to cross that bridge when I get there. Also I can try some of the other suggestions here, using a spotmeter or easel meter. Shelling out that much for something I need in theory is probably not wise.

 

[Bill Burk]

The Macbeth TR927 (they say TD but you can see it's a TR which means it does transmission and reflection.) on eBay looks like the most capable unit up there right now. But. And it's a big but. It might not even work at all. I doubt they'd honor free shipping outside US. But that's the kind of thing you'd be looking for.

 

[BCM]

Hi all,

I need a transmission densitometer, but the prices online for 30 year old "untested" gear has me thinking; we have access today to so much technology, couldn't I just build one using a Raspberry PI with a camera module, a light source and software to capture a calibration image for comparison against readings? Wouldn't be ideal for researchers that need to share findings against a known standard, but I only need it for my darkroom and my materials. Maybe a project like this exists already and if so, please share.


Full disclosure - I've never seen or used a densitometer. I am working entirely in the realm of how I think it works, so please correct me!

Just buy a used one. I've purchased a color one for $100 and another XRite in perfect condition for $125 in the last year. By the time you write the code, liberalize the results, deal with color issues, etc, you could have bought one and moved on. Trust me. I've designed them for a company years ago and there are lots of little issues to deal with. You can also build one using a spot meter (as noted by others) and as shown in Phil Davis' Beyond the Zone System book.

 

[sasah zib]

The PrintAnalyzer would have been THE solution but I specifically need to analyze colour when making separation negatives and seems this can only do BW (I should have stated this more clearly, my apologies).

I've never done any of this before, I am simply following my reference books at this point, but seems this is a must. Seems I may have to buy one that is specd to my needs as building one (at least for now) isn't a priority while I am starting to learn a new process.

View attachment 323117
(from Colour Photography in Practice)

@albada Thank you for sharing. I will get an easel meter but for a different purpose. Might be worth a try but I feel I would struggle with consistency and I would also have to use filters (I think) to read the correct density for each separation.

@nmp I am trying to avoid using a computer and scanner when working in my darkroom, so this wouldn't work in my case. Sure it would be helpful for someone else, so thank you for sharing.

you do not need a densitometer ... Spencer provides reasons, uses, serves as a guide to theory and improvement; a how to get the pencil sharper.

Your real problem is making tissue! Can you make a magenta, and a cyan tissue consistently?
Can you make an RA4 print --you can proof, make relative comparisons between negatives from prior sessions to the unknown

Can you make BW prints? then you can make ... with seps prior and unknown

Read him closer: look at what he is saying in your quote above -- and page 176 about lab uses.
A suggestion isn't a requirement.
you must first walk through the system before you begin perfecting your steps.

###end