The Printalyzer Densitometer Project

The Printalyzer Densitometer project is my attempt at building a new budget-friendly densitometer to make available to the resurging analog photography community. My ultimate goal with this project is to eliminate the cost and convenience barriers that currently stand in the way of many, who might want to try using such a device.

Its actually a project idea I've been kicking around for a long time, but initially sidelined due to concerns over optics and calibration. However, I've recently become motivated to revisit the idea.

What is a densitometer?

Simply put, a densitometer is a device that lets you measure the amount of light that is reflected off of a surface, or transmitted through a transparent material.

Reflection densitometers typically have a light source and a sensor positioned to bounce the light off the surface at a 45 degree angle. They then measure the ratio of light reflected off that surface, to the light that would be measured off a perfect reflection diffuser. This ratio is known as the "reflectance factor."

Transmission densitometers have a light source positioned in-line with the sensor, where the material being measured is placed in the path between the two. They then measure the ratio of light that passed through the material, to the light that would be detected if the material were not present. This ratio is known as the "transmittance factor."

In both cases, this "factor" is displayed as a density value on a logarithmic scale. (FYI, density is typically measured on a log10 scale, while camera exposure is typically measured on a log2 scale.)

Another thing that's important to note is that these reflectance or transmittance measurements are done at specific wavelengths, often depending on what sort of material is being measured. These are typically described as follows:

• Visual (570nm) - For B&W film and prints
  
• Status A (440nm, 530nm, 620nm) - For color positives
• Status M (450nm, 540nm, 640nm) - For color negatives
  

For a much deeper dive into everything about how these devices are supposed to work, there's an ISO standard that is extremely useful. Its known as ISO 5:2009 "Photography and graphic technology - Density measurements" and is broken out into 4 parts:

• ISO 5-1:2009 - Geometry and functional notation
• ISO 5-2:2009 - Geometric conditions for transmittance
  density
• ISO 5-3:2009 - Spectral conditions
• ISO 5-4:2009 - Geometric conditions for reflection density
  

Unfortunately, ISO standards are typically not free, but its still worth looking around to see what you can find out there. If you simply want to understand the wavelength distributions and density calculations, ISO 5-3 (and its accompanying data tables) is the most useful of these documents.

Two footnotes of interest here:

• The spectrum for "Visual" density is different from the spectrum used by lux meters. Lux meters use something known as the photopic luminous efficiency function, which peaks around 555nm.
• ISO 5-3 also defines something called "type 2 printing density (silver halide)", which appears to be of interest. However, I've tested my X-Rite 810's behavior and it does not appear to measure this alternate spectrum in its "Visual" reflection mode.
  

Why am I building this?

While working on the Printalyzer Enlarging Timer & Exposure Meter project, it became clear to me that anyone using the device would need a densitometer to build reasonably accurate profiles for their printing paper. Furthermore, it didn't seem like I could expect most people to actually have one of these instruments just laying around.

For the purposes of analog photography, there are currently two viable options for getting oneself a densitometer:

1. If you want new, there's the Heiland TRD-2. This is a very expensive option ($800+) and probably hard to justify for most people. It is also a very basic instrument (B&W only, not sure how well it follows ISO 5-3), and likely only commands such a high price due to the niche nature of the product.
2. There's the used market, with plenty of choices. The most common of these is something like the X-Rite 810/811/820. You can commonly find these for $300-500 on the used market. The problem is that you may need an independent source for calibration materials, and replacement bulbs can be extremely expensive and/or hard to find. However, these are very capable instruments. They'll do Visual, as well as Status A/M, so they're good all-around devices.

So it occurred to me... why does a densitometer have to be so clunky and expensive? All you need is a light source and sensor in the proper arrangement, and some electronics to drive them. It used to be expensive to do all of this back in 1987, when microcontrollers cost a pretty penny and decent ADCs were bespoke technology. But now its 2021, microcontrollers are cheap, we have full spectrum LEDs, and low cost highly integrated light sensors are readily available.

X-Rite 810 uses a complex ring of individually-filtered photodiodes

X-Rite 810 uses all this just to digitize analog values from the reflection sensors

Wouldn't it be nice if someone could make a densitometer cheap enough that every analog darkroom photographer could justify having one?

What are my project goals?

In the beginning, I was thinking about building the cheapest possible reflection-only densitometer. The idea was to make a simple "peripheral" that I could throw in the box with my Printalyzer Timer. (The likely-defunct DLG EMT project seems to have had this idea too.) While this could really get the cost down, it would unfortunately be useless as a standalone device. You'd still need something external to act as the user-interface, and use cases would be limited.

As I got deeper into the main Printalyzer Timer project, I also realized that bringing it to market would be quite a tall order. Its a complex device, would have a fair number of manufacturing and certification hurdles to overcome, and might be a little too much for a first swing.

However, developing the densitometer idea into a fully standalone product would be a lot easier. Sure, adding transmission capabilities and an on-device UI would add cost, but I could still keep it reasonable.

So anyways, here's what I'm currently trying to build:

• Reflection and Transmission capable densitometer, useful for measuring B&W film and paper
• Attempt to follow ISO 5:2009 (Visual spectrum) as closely as possible, and regularly test my device against the other market options
• Size the device so that its possible to measure any part of piece of film or printing paper, up to 5x7" in size. (This is smaller than other densitometers, but if you're measuring the middle of full 8x10"+ sheets then budget is likely less of a concern.)
• Use 3D Printing technologies to make it easier to design and build the sensor head, and possibly the complete device.
  
• Have buttons and a display on the top of the device, so it can be used entirely standalone. (For both calibration and measurement.)
• Use USB for both power and external control/interfacing. This way I don't need a custom or unique power brick, and it'll be easier to use with the Printalyzer Timer in the future. (The X-Rite has an RS-232 port that can be used for this, and the Heiland has an expensive extra option for a USB port as well.)
  

 

The Research Phase

The project started with me launching head-first into a fair amount of R&D. I really wanted to see how various light sensors compared, and whether or not I could actually build something capable of accurately measuring reflection density.

This meant building a series of development and test devices with different LEDs and sensors, collecting a lot of data, and seeing how they compared to each other and to my "reference" densitometers (X-Rite 810 and Heiland TRD-2).

These development and test devices included:

• A microcontroller with USB capabilities, similar to what I'd want to use in a real device
• Four LEDs with a constant current driver, to act as the reflection light source
• Two buttons for various purposes
• A 3D printed "shell" that would allow me to "snoot" the LED illumination at a 45 degree angle and put various diffusion/filter materials in front of the sensor
  

Circuit board from one of the development devices

Development devices and measurement references

I also built a specialized test rig that allowed for the use of the official "Evaluation Kit" boards for all of these AMS sensors, for an even broader range of testing.

Evaluation Kit Test Rig

The bulk of this research used two different sensors: The AMS TCS3472, and the AMS AS7341. The former is an RGB sensor that I figured could be tuned to give me equivalent "Visual" spectrum data, and the latter is a spectral sensor that I figured might be useful if I wanted to try the color statuses. (I've since given up on color for now, since its hard to get enough spectrum accuracy on a budget.)

My test materials consisted of an IT8/7.2 (Kodak Q-60) color reflection target and various strips of B&W paper (Stouffer tablet, step wedge exposures, various solid patch test strips). I also had my X-Rite Reflection Standard as a calibration reference.

From all of these experiments, I learned a couple important things:

• At lower densities, the sensors are less precise, but on a logarithmic scale this is an almost undetectable issue.
• At higher densities, the sensors are very precise but the overall device tends to be less accurate. The logarithmic scale tends to amplify this, so a lot of work went into refining the optical path to try and minimize this issue.
• Most reflection densitometers claim to work up to a density of D=2.50, but it is extremely hard to actually find any B&W papers with a density close to this. (Most top out around a Dmax of D=2.00, give or take a bit, including calibrated reference materials.)
• Color paper does seem to have a Dmax in that upper range. The darkest patch on my IT8 target measures around D=2.33 on my X-Rite.
  
• Color and B&W paper tend to have different spectral reflectance characteristics, especially into the infrared, which absolutely do affect measurement accuracy at the higher densities.

On the tail end of these experiments, I was finally taking a closer look at the sensitivity spectrum of different sensors and came to the realization that UV-IR cut filters were actually something I could buy on the cheap from the right sources. (Also, much credit is due to Engauge Digitizer for making it possible to actually bring datasheet graphs into spreadsheet-land for useful comparisons.)

Putting this together, it occurred to me that the simpler AMS TSL2591 might actually be a better choice for the project.

 

If I took that sensor's raw sensitivity curve, subtracted its second channel from the first, and then applied the UV-IR cut filter data, I'd end up with something that looks quite close to the actual "Visual" density spectrum:

Okay, its not exactly the same, but its pretty darn close. As long as we're not competing with the X-Rite for visual density readings of an overly blue or red target, we should be fine. (I should add that the Heiland TRD-2 will absolutely fail that sort of envelope test.)

The Development Phase

Finally, on to the design and development of the actual device. The design of the actual device was based on those earlier test devices, with a number of additions:

• Increase the length of the overall device
• Increase the LED spacing to increase height and have more room for filters under the sensor
  
• Add an OLED display screen with 4 buttons below it
• Design a new hinged enclosure with a separate transmission LED embedded in it

Full Prototype Circuit Board

Full Prototypes Being Assembled

Cross-Section of the Design Model

With this real device, I'm now writing the actual firmware to make it fully capable of acting like usable densitometer.

Reflection Measurement

Of course there's still plenty of work ahead, before I can declare this project to be "complete."

The Path Forward

So where do I go from here?

In the short term, there's still plenty of work to be done. The firmware still has some rough edges, and the mechanics of the enclosure are still in need of refinement. I also need to do more testing on the transmission side of things. I'm also exploring a variety of options for professional-quality 3D printing of the enclosure.

At some point I'm also going to need to decide on "official" calibration references for the device. Most likely it'll be something from the Stouffer catalog, calibrated either by them or by me.

Like all my projects, I fully intend for the hardware and software to be open source. This will make it possible to anyone in the community to improve the device firmware to their tastes, and to even try and build their own devices from scratch if they are so inclined.

Of course building your own hardware isn't really all that practical for a one-off. So once all the kinks are worked out, I hope to actually start producing these things for sale. I don't yet know what the price will be, but my goal is to make it cheaper than most of the competition (new and used).

Unlike my enlarger timer project, this device goes nowhere near mains AC. So I fully expect the whole regulatory/certification process to be significantly simpler. I'm likely going to feel comfortable sharing prototypes with other people, and using this project as a stepping stone towards finally being able to enter the analog photography marketplace.

There's just one big problem, in the short term... The great microchip shortage of 2021. You've probably heard about it in the news, perhaps in reference to cars or graphics cards. Well, its really a lot bigger than that. It pretty much affects everything in the embedded space. Even this little project uses a few components that, while available 6 months ago, are now out of stock everywhere with obscene lead times. So until that logjam clears, I won't be able to build more than a few prototypes.

 

The full schematics and design details for the project can be found here:

https://github.com/dkonigsberg/printalyzer-densitometer