DIY densitometer

radiant · Jan 6, 2020

So here we go again.. As @koraks suggested here https://www.photrio.com/forum/threads/diy-light-meter-for-darkroom.169398/page-2#post-2204322 that I should make a densitometer.. Film photography is interesting hobby; sometime you wonder where to use a densitometer and then in next picture one is actually designing one and measuring with it. Oh well..

So I build one. A DIY densitometer, currently working only on film so no reflective light measurement - altough I have prepared for it in the parts. The prototype is 3D printed. Under the sensor sits a light box (and don't ask me why instead of just a sharp led light) with a small, maybe 1mm or a bit under aperture punched in 120 backing paper for the light to enter the film. Upper part contains TLS2691 lux sensor (can measure fractions of lux easily) and software runs on ESP32 with OLED display.

So I did some tests on underexposed negative that is difficult or impossible to print. There are some other values as well for funny reference. And log2 and log10 already calculated, dunno which I should use for "official" values.

Then here is a much more dense negative for comparison - haven't tried to print but visibly much better better (Tri-X exposed at 3200 and developed by digitaltruth times on xtol 1+1):

.. so what next? What can we learn from these values? When trying to adjust development for printing, what should the values look like? I got approximately 3 lux difference out of 5 (pure light) so I think I have enough resolution in the system or do I?

Chan Tran · Jan 6, 2020

Use a step wedge to test your densitometer. Densitometer you should use log10.

radiant · Jan 6, 2020

Chan Tran said:
Use a step wedge to test your densitometer. Densitometer you should use log10.

Good idea. Would Stouffer T2115 do it? It could be also used on printing!

Bill Burk · Jan 6, 2020

For what you are doing... T2115 C, the calibrated one, will make you more confident in the results of your design.

koraks · Jan 6, 2020

Nice job! I should do this as well sometime (when I find the time). How do you like the esp32?
I've got a few 1" oleds lying around waiting for some projects. I've also looked at various light sensors but should order a few to play around with them.
I think the "diffuser" light source may actually be a good idea; the spread of a naked led is always a bit uneven, although it may be acceptable in the very center of the beam.

radiant · Jan 7, 2020

Bill Burk said:
For what you are doing... T2115 C, the calibrated one, will make you more confident in the results of your design.

Do I really need the calibrated one for this? I've understood that it is only for commercial/certified stuff where you need to prove your results. Meaning the same data is available but there isn't certificate included?

koraks said:
Nice job! I should do this as well sometime (when I find the time). How do you like the esp32?
I've got a few 1" oleds lying around waiting for some projects. I've also looked at various light sensors but should order a few to play around with them.
I think the "diffuser" light source may actually be a good idea; the spread of a naked led is always a bit uneven, although it may be acceptable in the very center of the beam.

Thanks koraks! ESP32 is great and the combo of ESP32+OLED+18650+chager is a killer in features and in price (6.2 euros, whaat!), I'm using similar to this: https://www.ebay.com/itm/0-96-inch-...063347?hash=item5d9a435873:g:CHIAAOSwzZFeDB0i - just solder the sensor and everything else is connected. Programming it with Arduino with ESP32 toolchain, which is nowadays stable.

Yes I made a pretty good diffuser actually; it is approximately 6cm x 6cm 6cm 3D printed box. The inside walls are covered with aluminium tape and I have a led strip in the bottom facing invards (attached to the "walls" - not to the bottom), going around few times. Then I have 4mm thick opaque acrylic plate on top. The light seems to be really even, I cannot find any "hotspots" on the surface at all. The distance to the LED strip seems to be enough and the light is bouncing around in the box well, I think.

radiant · Jan 7, 2020

So what about the values. The underexposed / badly printing negative has contrast difference on 0,28 in log10. The visually much better Tri-X negative has 0,72 difference.

So the ISO R values for these are 28 (underexposed) and 72 (tri-x). We can easily see why the underexposed negative is difficult to brint, it goes out of roof from grade 5. For example Ilford V paper at 5 grade is ISO R 50. The Tri-X negative is very close to perfect, I guess; it should be printed with grade 3 on Ilford V paper.

Question is what grade should be the target, or is this so simple? If I've understood correctly the grade 2 would be more ideal to target which would mean that we need ISO R of 110, which is in log10 1.1 - correct me freely on this! The Tri-X was heavily pushed which might caused a bit thinner negative than the target is. Is the correct way to get closer to ISO R 110 by overexposing or overdeveloping?

Have I stepped correctly into the world of negative densities? What I really know now that I never ever want negatives with low density because those are too impossible to print. Why I have been exposing the negatives incorrectly is another story..

jisner · Jul 7, 2020

vedostuu said:
Have I stepped correctly into the world of negative densities? What I really know now that I never ever want negatives with low density because those are too impossible to print. Why I have been exposing the negatives incorrectly is another story.

I just noticed this discussion, almost a year too late. I think your film densitometer would be of interest to people who make digital negatives for alternative photographic processes. In our world, the exposure scale of the process (similar to a paper grade) is given. Say it's 2.4 log base-ten. Then we must make a digital negative (on an inkjet printer) having a density range of 2.4. It's a hard problem with no perfect solution (that's my opinion). A film densitometer would make the problem easy.

Do you have any photos of your DIY densitometer? Do you have any plans to commercialize it?

radiant · Jul 8, 2020

The thing is that I haven't used my densitometer almost at all.

The biggest problem is that is really difficult to aim the measuring point to certain place on the negative (135 or even 120 film). I mostly just panned the film around while measuring max/min densities. For example if the density I want to measure is on the edge of the negative, I have to be really careful not to measure the base+fog area.

Much easier way is to do this measurement on enlarger with my DIY darkroom meter - as one can imagine. Precise point measurement is doable on enlarged image however sometimes it might be a bit tricky on that too if the measured area is really small on the negative. For digital negatives this could be somewhat helpful of course since the negative and details are much larger.

I don't have photo and maybe it is best that way since the device is quite a hack, not beautiful in any means. There is small light table (5cm x 5cm) and the sensor is above that on pivoted arm. The sensor is in the end of maybe 1cm long tube, diameter under 1mm. And worth mentioning if someone is planning to do the same, the light table was a bad idea. It causes light to enter from too large area to the sensor - that is why I just added some backing paper on top of the light table and punctured small hole straight under the sensor tube.

And no I haven't even considered commercialize it Tempting thought - of course - but it is too far stretch.

Lachlan Young · Jul 8, 2020

vedostuu said:
So what about the values. The underexposed / badly printing negative has contrast difference on 0,28 in log10. The visually much better Tri-X negative has 0,72 difference.

So the ISO R values for these are 28 (underexposed) and 72 (tri-x). We can easily see why the underexposed negative is difficult to brint, it goes out of roof from grade 5. For example Ilford V paper at 5 grade is ISO R 50. The Tri-X negative is very close to perfect, I guess; it should be printed with grade 3 on Ilford V paper.

Question is what grade should be the target, or is this so simple? If I've understood correctly the grade 2 would be more ideal to target which would mean that we need ISO R of 110, which is in log10 1.1 - correct me freely on this! The Tri-X was heavily pushed which might caused a bit thinner negative than the target is. Is the correct way to get closer to ISO R 110 by overexposing or overdeveloping?

Have I stepped correctly into the world of negative densities? What I really know now that I never ever want negatives with low density because those are too impossible to print. Why I have been exposing the negatives incorrectly is another story..

The problem with densitometry is that in many cases it's a bit like somebody being able to draw a map without being able to actually read a map. It can show you the toe, midtone shape, and shoulder of a film or paper - and the speed point in your system relative to ISO, but unless you understand it systemically, people often end up with harder to print negatives - because they ignore the importance of the toe and shoulder shape relative to the exposure and processing they should use - let alone the paper they subsequently print on.

radiant · Jul 8, 2020

@Lachlan Young what you wrote actually works in many other things we measure

I basically needed the densitometer for just to see that the overall contrast is OK. Maybe after year I think differently but for now I cannot see any way that I could control toe&shoulder too.. No way.

ic-racer · Jul 8, 2020

The novice should concentrate on speed and contrast and ignore shoulder and toe shape until later. Speed and contrast can be measured with a densitometer (home-made or purchased).

DREW WILEY · Jul 8, 2020

Don't forget old school visual densitometry. You take some black card stock and a simple single-hole paper punch. Punch two holes toward the middle of the card but a few inches apart. Atop a lightbox, use one hole to select the portion of the negative you want to read, and compare it to the most closely matching step on the wedge using the other hole. It's good enough for beginner results like Ice Racer is recommending. For actual curve plotting, you need something serious.

jisner · Jul 8, 2020

vedostuu said:
The thing is that I haven't used my densitometer almost at all.

The biggest problem is that is really difficult to aim the measuring point to certain place on the negative (135 or even 120 film). I mostly just panned the film around while measuring max/min densities. For example if the density I want to measure is on the edge of the negative, I have to be really careful not to measure the base+fog area.

Thank you for your reply. In Photoshop, you have no control over the amount of ink the printer uses to make "black." It can only be controlled in the printer settings. On professional Epson inkjet printers, the amount of ink is controlled by a printer setting that you can adjust from -50% to +50%. For each value of this setting (say, in 5% increments), we would like to know the negative density range it produces. The results could be recorded in a table. With the table, we can choose the printer setting to match the required exposure scale of each alternative photographic process (for that specific printer/ink/transparency material). To build the table, we need only two density measurements for each table entry: base + fog and black. Their difference gives the negative density range. These measurements would only require a tiny piece of negative material (say 2cm x 2cm) to be measured by the densitometer.

I can see that the biggest challenge of this project will be the physical construction of the densitometer body/measuring platform, not the programming, electronics, and sensors.

DREW WILEY · Jul 8, 2020

That's actually a huge area to measure. Most densitometers can read 1 or 2 mm diameter so they can home in on a specific spot on the negative.

jisner · Jul 8, 2020

I was referring to the size of the sample to be measured. It's just a small piece of transparency printed black. It wouldn't require a large platform to maneuver it under the read head. I would take a few "point" readings on the sample and average them.

RalphLambrecht · Jul 8, 2020

vedostuu said:
So here we go again.. As @koraks suggested here https://www.photrio.com/forum/threads/diy-light-meter-for-darkroom.169398/page-2#post-2204322 that I should make a densitometer.. Film photography is interesting hobby; sometime you wonder where to use a densitometer and then in next picture one is actually designing one and measuring with it. Oh well..

So I build one. A DIY densitometer, currently working only on film so no reflective light measurement - altough I have prepared for it in the parts. The prototype is 3D printed. Under the sensor sits a light box (and don't ask me why instead of just a sharp led light) with a small, maybe 1mm or a bit under aperture punched in 120 backing paper for the light to enter the film. Upper part contains TLS2691 lux sensor (can measure fractions of lux easily) and software runs on ESP32 with OLED display.

So I did some tests on underexposed negative that is difficult or impossible to print. There are some other values as well for funny reference. And log2 and log10 already calculated, dunno which I should use for "official" values.

View attachment 237664

Then here is a much more dense negative for comparison - haven't tried to print but visibly much better better (Tri-X exposed at 3200 and developed by digitaltruth times on xtol 1+1):

View attachment 237665

.. so what next? What can we learn from these values? When trying to adjust development for printing, what should the values look like? I got approximately 3 lux difference out of 5 (pure light) so I think I have enough resolution in the system or do I?

My hat goes off to you that's quite a task I wouldn't know where to start.

radiant · Jul 9, 2020

jisner said:
I can see that the biggest challenge of this project will be the physical construction of the densitometer body/measuring platform, not the programming, electronics, and sensors.

I had the same challenge. Altough I have a 3D printer and can design parts, it can still end in non-wanted results. We are talking about a bit optical device here..

I think there are alternatives, for example flatbed scanner can be used as densitometer, see here: https://www.scotty-elmslie.com/uplo...tical_guide_to_film_characteristic_curves.pdf

Other possibility is to use digital camera and aim the printed material so that it covers whole frame. Cameras in-built-meter shows with a fixed shutter speed & ISO the aperture change on different areas.

RalphLambrecht said:
My hat goes off to you that's quite a task I wouldn't know where to start.

This is quite an achievement since your book is just next to me!

radiant · Jul 9, 2020

@jisner and one more idea: if you are talking about 2cm x 2cm you could easily just press the sensor on PCB against your printed material on a light table and get pretty decent reading?

jisner · Jul 9, 2020

Ralph Lambrecht! I knew that name sounded familiar. I just bought a copy of Way Beyond Monochrome and have been recommending it to all my photographer friends. I don't have a darkroom so I can't do film, but I am fascinated reading about it, doing my best to map ideas from silver gelatin film and paper into the alt process world of digital negatives and hand-coated papers.

jisner · Jul 9, 2020

vedostuu said:
@jisner and one more idea: if you are talking about 2cm x 2cm you could easily just press the sensor on PCB against your printed material on a light table and get pretty decent reading?

It would need to be a UV light table. All our processes are exposed under UV light. But thanks to this discussion (and to the references you gave me), I think I have an easy solution that will cost $0 and require very little DIY effort. We already have the three necessary ingredients: (1) a UV exposure unit (2) a UV dose meter and (3) a Stouffer 4x5 21 step wedge.

Place the UV sensor at a fixed distance under the UV lights, pointing up toward the lights and tape it down so it can't move. Place the sample negative on top of the sensor (this would require some DIY hacking to secure it, prevent flare, etc.) so that the sample sits between the sensor and the light source. Expose for a fixed dose, say 1000 units. Record the time in seconds. Let's say we get 180 seconds. We have now replaced an unknown density by a known time. To get from time to density, we can calibrate with a Stouffer. Fortunately, this only needs to be done once. Repeat the above procedure for each step of the Stouffer's 21 steps. This will give 21 time values. Each of the time values corresponds to a known density, from 0.05 to 3.05. From the Stouffer measurements, make a table of time vs. density. Look up the sample time (180) in the table and interpolate the density. I think the result should be good within 1/2 stop.

DREW WILEY · Jul 9, 2020

Affordable UV-simulating XRite transmission densitometers exist, but are hard to get because they've all apparently been snatched up by UV printmakers already. But in relation to this topic, it's quite important to use a clean modern step tablet. Many of the older kind have acquired a distinct yellowish cast which will give somewhat false results due to the spectral blockage of the yellow. Something to think about if shopping for a used one. New ones aren't all that expensive anyway.

jisner · Jul 9, 2020

I wasn't aware that step wedges were susceptible to that problem, but it's good to know for future reference. I have a Stouffer 4x5 21-step wedge that's in excellent shape. The more steps the better, but the 4x5's go up in price exponentially with the number of steps. A true UV transmission densitometer would be ideal, but I can't justify the price, especially if I can cobble something together like what I described.

Bill Burk · Jul 10, 2020

jisner said:
It would need to be a UV light table. All our processes are exposed under UV light. But thanks to this discussion (and to the references you gave me), I think I have an easy solution that will cost $0 and require very little DIY effort. We already have the three necessary ingredients: (1) a UV exposure unit (2) a UV dose meter and (3) a Stouffer 4x5 21 step wedge.

Place the UV sensor at a fixed distance under the UV lights, pointing up toward the lights and tape it down so it can't move. Place the sample negative on top of the sensor (this would require some DIY hacking to secure it, prevent flare, etc.) so that the sample sits between the sensor and the light source. Expose for a fixed dose, say 1000 units. Record the time in seconds. Let's say we get 180 seconds. We have now replaced an unknown density by a known time. To get from time to density, we can calibrate with a Stouffer. Fortunately, this only needs to be done once. Repeat the above procedure for each step of the Stouffer's 21 steps. This will give 21 time values. Each of the time values corresponds to a known density, from 0.05 to 3.05. From the Stouffer measurements, make a table of time vs. density. Look up the sample time (180) in the table and interpolate the density. I think the result should be good within 1/2 stop.

Tomorrow when the sun is out I will try a UV meter and Stouffer scale. There’s a chance it will block UV

radiant · Jul 10, 2020

Just a side note; there is quite capable UV-A/B sensor available https://www.ebay.com/itm/I2C-UVA-UV...964824?hash=item3f6f3948d8:g:oOkAAOSwr2RYKWDA

(edit: adafruit one was being discontinued)

DIY densitometer

Moderator

Privacy & Transparency

Privacy & Transparency