Transmission densitometers and the importance of emulsion orientation

This is a subject I don't see mentioned much, so I'd like to finally kick off the discussion to bring some visibility.

As many of us are using an increasingly long tail of densitometers of various generations and models, and are turning to mismatched and/or 3rd party sources of calibration materials for them, this is something that's becoming all too easy to get mixed up and confused on.

Here are the top-level bullet points:

• Every densitometer has a preferred orientation for the film emulsion to face when taking measurements.
• Film emulsion direction is sometimes mentioned in the manual, but not always
• Calibration references are labeled for the densitometer they're intended to be used with, which may not match the densitometer you are using them with

So how do you know which direction the emulsion is supposed to face?

• If its in the product documentation, then go with that. It might be a few pages into the manual, and easy to miss.
• If the densitometer has a polished opal diffuser, then the emulsion should face that diffuser.
• If the densitometer does not have a polished opal diffuser that you can see, then the emulsion should probably face the sensor.

In my own survey of densitometers I've paid close attention to, here's what I've found:

Densitometer	Emulsion Direction	Is it in the manual?	Notes
X-Rite 810	Facing down	No	810 and 820 are the same optical mechanism, polished opal diffuser and sensor elements in the base
X-Rite 820	Facing down	No
X-Rite 361T	Facing up	No
X-Rite 301	Facing up	Yes
Tobias TBX1000C	Facing up	Yes	Polished opal diffuser and sensor elements in the head
Heiland TRD-2	Facing down	Yes
Dektronics Printalyzer Densitometer	Facing up	Yes	Acrylic diffuser in the base, sensor in the head, but calibrated to compensate for measurement artifacts.
Dektronics Printalyzer UV/VIS Densitometer	Facing down	Yes	Polished opal diffuser in the base, sensor elements in the head.

One example of a common way to mix this up is when calibrating your unit with a reference strip designed for a different densitometer. For example, the X-Rite 301-27 reference is easy to find new, and clearly labeled to be measured emulsion-side-up. But if you're using it to calibrate an emulsion-side-down instrument, you need to measure it upside-down. If you don't, you will get slightly incorrect readings.

Now to make all this clearer, I decided to go ahead and do a little date collection experiment. Using a fresh X-Rite 301-27 calibration strip, a Stouffer T2120 step wedge, and an X-Rite 810 densitometer, I tested 4 configurations:

• Calibrate emulsion up, Measure emulsion down
• Calibrate emulsion up, Measure emulsion up
• Calibrate emulsion down, Measure emulsion down (The correct configuration)
• Calibrate emulsion down, Measure emulsion up

If you plot the raw density values for the step wedge measurements, things appear somewhat close:


But if you plot the differences between the 4 configurations, then the differences become quite obvious:

(This graph is showing the differences between the "correct" configuration and the other 3.)

The manual for my Gretag D200-II doesn't mention film direction for any of the measurement modes, but I've always placed the film emulsion down.

For the Macbeth TR-524 it is emulsion facing UP, on page 1-6 (j).

Warning - pedantry alert!
Don't you mean emulsion orientation, rather than emulsion direction?
To me, direction implies movement.

Interesting. That might explain some odd behaviour with my old Photronix Delta III (light in the base, sensor in the arm above). Needless to say, it came my way years ago without instructions!

MattKing said:

Warning - pedantry alert!
Don't you mean emulsion orientation, rather than emulsion direction?
To me, direction implies movement.

Click to expand...

Pedantry accepted, within reason.
Thread subject changed, but I didn't upload new graph screenshots.

grahamp said:

Interesting. That might explain some odd behaviour with my old Photronix Delta III (light in the base, sensor in the arm above). Needless to say, it came my way years ago without instructions!

Click to expand...

Even when there are instructions, I think a lot of people miss this part.

One of the test users for my new densitometer model got this wrong on some of his early tests, and suggested that I mention the emulsion orientation in the included quick start guide.

It is mentioned in the quick start guide, and the regular user manual, and even on a slip of colored paper included in the box for extra visibility.

dkonigs said:

Pedantry accepted, within reason.
Thread subject changed, but I didn't upload new graph screenshots.

Click to expand...

The graph screenshots are fine as they are.
And I'm not sure why, but the thread seems so much better now

dkonigs said:

Pedantry accepted, within reason.
Thread subject changed, but I didn't upload new graph screenshots.

Click to expand...

What you meant was obvious to everyone except the pedant.

Milpool said:

What you meant was obvious to everyone except the pedant.

Click to expand...

You seem to forget that this is an international site, where many people do not use English as a first language, and a reasonable percentage of people actually have to use computer translation in order to understand every single English post they read.
Not to mention the people who are just learning the first things they have ever learned about densitometers.
My first reading of the thread title made me think that it referred to the direction that the film was inserted, not whether the emulsion was up or down. It was easy for me to realize quickly that that wasn't what was meant - I have enough familiarity with densitometers to have been able to figure that out - but why would you assume every other person reading this thread would be coming from the same point of understanding?
Pedantry does, sometimes, lead to better understanding and clarity you know.
Not posted as a Moderator per se, but I have been observing how people interact here for a long time, and I have certainly seen a fair few misunderstandings during that time too.
I've even helped straighten a few of those misunderstandings out!

I ran my Stouffer 21 step wedge on my old Photronics Delta III. Flipping the wedge over made no difference. I tried some Delta 100 and got the same result - emulsion and film base gave the same readings.

Now, the best I can get is 0.01 units on this device, which uses interpolation of a needle on a non-linear scale - and the scale does not have a parallax correction mirror either. It measures 7 stops, topping out around 2.20 on an upper range scale maximum of 3.0. So either this machine is not sensitive to the surface, or it is just not able to resolve the difference.

I got it as part of a job lot of gear. It is good enough for my use - internal comparisons with my own materials and methods. It seems to be reasonably linear (user error permitting), but I am not sure about the basic calibration. It gives 0.06 for the step wedge patch 1, to 2.20 for patch 15.

Worth noting Todd-Zakia in Photographic Sensitometry names three types of density: 1. Specular; 2. Diffuse; 3. Doubly Diffuse
and three further classifications related to the nature of the receptor: 1. Visual; 2. Photoelectric; 3. Printing

I looked through the manual for my Marshall Studios Densitometers and they don't say. I always use it emulsion oriented up towards the sensor (my eye).
I label my readings from that densitometer as "Diffuse Visual." Most Kodak material, if you look at the graphs, mark their densitometry as "Diffuse Visual" but I wonder if they really have engineers using a visual device because it's hard work. I suspect they just forgot what it meant and put the words on the graphs because that's what they always did.

The Macbeth TR524 also doesn't say in its manual. I always use it emulsion oriented up towards the photomultiplier tube sensor in the head.
I label my readings from that densitometer as "Diffuse Photoelectric."

Bill Burk said:

Worth noting Todd-Zakia in Photographic Sensitometry names three types of density: 1. Specular; 2. Diffuse; 3. Doubly Diffuse
and three further classifications related to the nature of the receptor: 1. Visual; 2. Photoelectric; 3. Printing

Click to expand...

And the moment you try comparing readings across densitometers, or across lab instruments capable of measuring transmission density (e.g. spectrometers), you'll quickly notice that the nitpicky peculiarities of light diffusion make a very noticeable difference in how well all of your results match up with each other.

Most desktop densitometers are diffuse (either influx or efflux). I think the setup that NIST uses for the 38120C (a.k.a. SRM 1008) might be doubly diffuse, but its not clearly stated . They have a whole document (NIST SP 250-99) covering the specifics of their setup in excruciating detail, and that was my takeaway from some of the figures.

Ultimately, the setup used for "the master standard" and the setup used by your actual densitometer might not match perfectly. But if things are calibrated correctly, and you're generally measuring the same sorts of materials, it might not matter.

The biggest issue is when you do things like mismatching film orientation between calibration and measurement. In that case, you get this low density "hump" in the results.

I think it would be good to check using a Doubly Diffuse Printing method: Make a test print of a calibrated step wedge on paper, and measure the deviation between reflection density of a step wedge created with open air light, with the reflection density of the printed step wedge. For the open air light step wedge I might use the Haynes Grayscaler, which simply asks you to spin a sector wheel on timed intervals over a sheet of printing paper.

Or, I guess, make a contact print of the test step wedge or the photographic negative you want to analyze next to a contact print of the calibrated step wedge and compare the deviation in reflection densities of different points.

All to see how well your chosen densitometer correlates to your print method.

I don't know if I would go to the trouble of 'correcting' the readings from a photoelectric densitometer to agree with the 'effective expected' density of printing.

I'd just alter my aims (or mindset) to fit the chosen densitometer into my workflow.

dkonigs said:

And the moment you try comparing readings across densitometers, or across lab instruments capable of measuring transmission density (e.g. spectrometers), you'll quickly notice that the nitpicky peculiarities of light diffusion make a very noticeable difference in how well all of your results match up with each other.

Most desktop densitometers are diffuse (either influx or efflux). I think the setup that NIST uses for the 38120C (a.k.a. SRM 1008) might be doubly diffuse, but its not clearly stated . They have a whole document (NIST SP 250-99) covering the specifics of their setup in excruciating detail, and that was my takeaway from some of the figures.

Ultimately, the setup used for "the master standard" and the setup used by your actual densitometer might not match perfectly. But if things are calibrated correctly, and you're generally measuring the same sorts of materials, it might not matter.

The biggest issue is when you do things like mismatching film orientation between calibration and measurement. In that case, you get this low density "hump" in the results.

Click to expand...

I looked at light integrating spheres on eBay. Don't think I'll be buying one of those. They're pretty pricey.

Maybe paint a ping-pong ball black and poke a couple holes in it.

Years ago I worked for a company that repaired and calibrated Macbeth densitometers. They were outside my technical speciality but I got to play with several of them during my "zone system obsession" years.
Not knowing any better I always took density readings with the film emulsion facing the sensor. I figured that my enlarging lens would be looking at the emulsion side of the film so the densitometer should be looking at
that side of the film too. Results were always within the generous error bars I worked between. I should have studied more when I had the chance.

A good rule of thumb, for everything, is "emulsion to emulsion".
Unless the instructions say otherwise (as @dkonigs said at the beginning).
For example, in graphic arts, sometimes you have to do emulsion-to-base because the number of steps are even, and you of course need the image to be right-reading in the end.
Or sometimes base-to-base because you are doing spreads or chokes (and you want light to get in around the edges).

To add another instrument to the list, the manual for the X-Rite 311 specifies that the emulsion side should be facing up when calibrated or measured.

I don't have it available at the moment, but recall that there was a substantive difference for each orientation and a difference in the resulting HD-LD measurements from an Ilford Control Strip.

And coincidentally, it also works quite well using NiMH batteries in place of the ancient NiCd that it was originally spec'd for.