Sensitometer Calibration: Synchronizing In Camera Data with Sensitometer Data

[ic-racer]

I made some modifications to my film H&D curve analyzing spreadsheet. The idea is to find out how the set the ISO/ASA dial on the camera. Isn't that the question we all want to know?


In the camera, a series of exposures at Zone I are made with various exposure index settings of the camera's ISO/ASA dial. For example 400, 320, 250, 200, etc.

There are hundreds of ways to do this. In this test I metered at EV9 on a calibrated light source and exposed at the EV5 setting on the light source.

This data was then graphically superimposed over the density readings from a sensitometer exposure on the same strip of film. The in-camera data is represented by the GREEN circles.

The "Estimated Sensitometer Millilux-sec Output" is tweaked until the two curves are aligned.

The spreadsheet then gives the number to which the camera's ISO/ASA would be set to make the best printing negative.

 

[ic-racer]

I added these extra cells to the spread sheet. The green column of Exposure Indexes (800, 650, etc) are converted to millilux seconds for plotting with the H&D curve (gray column).
The white column contains the log density readings from the processed film.

 

[ic-racer]

The green, in-camera data, does not align with the sensitometer data. The "Estimated Millilux-sec" output of the sensitometer is not matching the film camera's meter/shutter/diaphragm response. The sensitometer's H&D curve needs to be shifted to the left. Simple addition and subtraction due to the fact the millilux-sec scale is a log scale.

 

[ic-racer]

Again, here the data are not aligned. The H&D curve needs to be shifted to the right. Again, in my spreadsheet I'm altering my "Estimated Millilux-sec" sensitometer output value to do this.

 

[ic-racer]

Now the two curves (black and green) are overlapping. In this case I can read off the calculated exposure index from the spreadsheet for my HP5 sample:

X-intercept based speed (W-Speed) = 280
Fixed Density based speed (Delta-X) = 245

 

[Bill Burk]

Have you aligned your estimated illumination at the film plane of the sensitometer with the film's nominal rated speed?

 

[ic-racer]

Have you aligned your estimated illumination at the film plane of the sensitometer with the film's nominal rated speed?

Bill, I'm glad you saw the thread.

Here is the back story:

Since building the Serhiy Rozum tester, I now have both a calibrated light source and a way to measure shutter speeds. Knowing the shutter speed and light output, one can then know a specific millilux-second exposure.

So, I used a Nikon F2 as a sensitometer. The camera has been adjusted to a known shutter speed, and it's DP-12 finder has been adjusted to my calibrated light source. The adjustments for accuracy have been made in stop-down mode to eliminate any aperture coupling error.

The F2 made the 4 exposures represented by the green circles. These exposures are therefore the 'calibration standard.' The H&D curve was adjusted to match.

Getting to your question about box film speed. What I found is that in my darkroom, I can't get brand-name film to perform to its stated box speed. So, using that technique (of calibrating the sensitometer to film), I was still getting under-exposure in my photographs.

So in the example 400 should be 400 and the film does not reach this (so the calibrated Nikon F2 says).

 

[ic-racer]

The test is not perfectly matched for these reasons, but otherwise I like basing exposure on what the camera 'sees':

Wejex exposure 1 sec -- Nikon Exposure 1/30 sec
Wejex Lamp = incandescent + 80A -- Nikon Exposure = LED 3200 + 80A

 

[BCM]

How are you accounting for the myriad of other variables in the darkroom, variability in the shutter speed and your densitometer fluctuations? I designed a few densitometers in my lifetime and it is very, very common to have a 1/3 stop variation from reading to reading just based on the electronics. What is your end goal and expected variation? I'm sure you already know that light meters will respond to different light intensities and colors (and in heat and cold) resulting in non-linear measurements. Testing as you have described it will only be (sort of) accurate for those specific test parameters.

 

[ic-racer]

Thank you for the inquiry.

No variability that I can detect in my densitometer. Redings are rock solid. That, and the step wedges are probably the most stable components of the system.

Light meter test comforms to ISO 2720. Though rather an use the Ammonium cobalt sulphate hexhydrate and coper sulphate pentahydrate filter, I'm using off the shelf 80A filters.

If not mentioned, the in-camera exposures and sensitometer esposures are on the same piece of 35mm film.

 

[Bill Burk]

If you didn’t already, I’d repeat Wejex exposures and the four film exposures a few times on the roll. Maybe average the readings. Develop close as you can to ASA parameters per the Wejex plot.

I’d plot the Wejex curve and then superimpose the camera density readings on that curve.

You have to assume that identical densities are created by the same exposures. So the camera exposures have to be on the Wejex curve.

Then you can try to reverse engineer the x-axis and maybe decide a discrepancy might be the camera f/stop or the camera calibration.

 

[Bill Burk]

Tell more about how you suspect you never got full film speed. I always assume that manufacturers speed is easily confirmed. Maybe it’s easier for me with an EG&G because the flash is close to daylight, while the Wejex and blue filter might not have as much “actinic” light.

 

[Stephen Benskin]

Could you also be more specific on the breakdown?

"I metered at EV9 on a calibrated light source and exposed at the EV5 setting on the light source," sounds like you stopped down 4 stops from the metered exposure, but it could also mean you opened up 4 stops, or you used a light source that metered at EV5 to expose the test but set the camera for EV9.. I would rather not assume.

It also sounds like you made a sensitometric exposure using a sensitometer and a step tablet and then did a camera exposure test. Could you explain how you determined the sensitometric exposure? It would be easier for me to understand what's going on if you use the exposure at the film plane and use lxs.

 

[ic-racer]

The camera’s meter was zeroed ( stop down metering) to EV 9 on the light source, then the light source was set to EV 5 for the exposure. This was done at varying exposure indexes. Forming the toe of a H&D curve.

That partial curve was superimposed on the uncalibrated step wedge exposure to complete the H&D curve.

 

[ic-racer]

Tell more about how you suspect you never got full film speed. I always assume that manufacturers speed is easily confirmed. Maybe it’s easier for me with an EG&G because the flash is close to daylight, while the Wejex and blue filter might not have as much “actinic” light.

None of my cameras have ever produced zone I, 0.1 at box speed (With Zone I criteria). Not in the 1970s, 1980s, 1990s, etc. Not even now that I have calibration equipment for meters and shutters.

 

[Bill Burk]

None of my cameras have ever produced zone I, 0.1 at box speed. Not in the 1970s, 1980s, 1990s, etc. Not even now that I have calibration equipment for meters and shutters.

Is that at 1/10 the illumination of the meter recommended exposure?

 

[ic-racer]

Is that at 1/10 the illumination of the meter recommended exposure?

No, I believe 4 stops down is 16.

You are alluding to a discrepancy somewhere.

 

[Milpool]

None of my cameras have ever produced zone I, 0.1 at box speed. Not in the 1970s, 1980s, 1990s, etc. Not even now that I have calibration equipment for meters and shutters.

Isn’t that basically because you’re using the Zone System criterion (ie targeting the 0.1 density 4 stops below the metered exposure) whereas the ISO speed point Hm is 3 1/3 stops below the metered exposure?

In my own work on this stuff I always found I got the ISO speed for Kodak and Ilford films I used - in other words a Zone System EI ~2/3 stops less than the ISO speed.

On the other hand I might have misunderstood what you meant. I’m not entirely clear on your methodology.

 

[ic-racer]

Isn’t that basically because you’re using the Zone System criterion (ie targeting the 0.1 density 4 stops below the metered exposure) whereas the ISO speed point Hm is 3 1/3 stops below the metered exposure?

In my own work on this stuff I always found I got the ISO speed for Kodak and Ilford films I used - in other words a Zone System EI ~2/3 stops less than the ISO speed.

On the other hand I might have misunderstood what you meant. I’m not entirely clear on your methodology.

Yes, I think that is what Bill was alluding to above.

Again, the methodology here to adjust the sensitometer response to match the 'old fashioned' zone I test which has been working for me since the 1970s. The idea is to consistently produce exposures that print well for me rather than setup an ISO testing lab.

 

[Bill Burk]

ic-racer I thought you taught me this stuff.

Yes the meter wanting to put its reading 10x the 0.10 speed point is what I was alluding to.

This was the closest demonstration I made, the picture of my dog with a Sekonic multi-spot gray card.

I used an OM-4 and after metering on gray card 18% with its spotmeter, I selected the recommended exposure as close as possible and trimmed with ND filters. (Olympus only gives me full stop f/stops and shutter speeds so to get 0.1 increments I used 0.1 and 0.2 Wratten 96 ND filters to get as precise an exposure as possible).

The exposure that I put on the film at the 18% gray where I metered would have been an exposure of -1.10 log MCS and what I found was density corresponding to -1.17 log MCS.

I considered this to be 1/6 stop away from expected results, but I was trying to show the meter point is 10x the 0.1 speed point. For 100 speed film where 0.1 speed point is expected to fall with -2.1 log MCS, the meter point is expected to be at -1.1 log MCS.

 

[Stephen Benskin]

ic-racer I thought you taught me this stuff.

Yes the meter wanting to put its reading 10x the 0.10 speed point is what I was alluding to.

This was the closest demonstration I made, the picture of my dog with a Sekonic multi-spot gray card.

I used an OM-4 and after metering on gray card 18% with its spotmeter, I selected the recommended exposure as close as possible and trimmed with ND filters. (Olympus only gives me full stop f/stops and shutter speeds so to get 0.1 increments I used 0.1 and 0.2 Wratten 96 ND filters to get as precise an exposure as possible).

The exposure that I put on the film at the 18% gray where I metered would have been an exposure of -1.10 log MCS and what I found was density corresponding to -1.17 log MCS.

I considered this to be 1/6 stop away from expected results, but I was trying to show the meter point is 10x the 0.1 speed point. For 100 speed film where 0.1 speed point is expected to fall with -2.1 log MCS, the meter point is expected to be at -1.1 log MCS.



View attachment 394263

Let's not forget the constant, 0.80, is used because of the sensitivity of black and white film to the additional blue of the simulated daylight used as the spectral value of the illuminating light source. The increase of blue in the color temperature effectively makes the film appear 1/3 stop faster than it is. The constant adjusts the speed to reflect the point of the curve Δ 0.10 log-H to the left of the speed point. This can introduce a potentially indeterminant 1/3 stop variable into the film speed results depending on the spectral characteristics of the light source and whether a constant is or isn't incorporated.

 

[Stephen Benskin]

Yes, I think that is what Bill was alluding to above.

Again, the methodology here to adjust the sensitometer response to match the 'old fashioned' zone I test which has been working for me since the 1970s. The idea is to consistently produce exposures that print well for me rather than setup an ISO testing lab.

I'm a little confused why someone at your level is doing serious testing with Zone System speeds to find a working EI? I can understand intellectual curiosity. I know you understand the discrepancy in the ratios between Hg and Hm. Why not just find the Delta-X point to determine the actual speed (0.3G) and incorporate a higher constant to create an EI that works for you? This is effectively how speed determination actually works. A meaningful point on the film curve is defined based on image quality and then the camera exposure through the choice of EI is determined based on the average Luminance range. Or why not just factor the ratio discrepancy into an EI derived from the ISO speed?

 

[ic-racer]

All points understood and well recieved.

This experiment is somewhat an example of forensic exposure research, pure pragmatism, coupled with some new features (scaleable/shifting baseline) added to my spreadsheet.

What was the most direct connection between my well exposed negatives of 70s, 80s, 90s? It was using the zone I = 0.1 criteria. So I just tried shifting everything else around that.

More of the back story: I'm way behind in printing. Looks like some time around 3-5 years ago negatives go thin. Probably same time I gave up on "New T-max" and shifted totally to HP5 exposing at 400.

 

[Bill Burk]

Ah good story.

I thought your Great Sensitometer Shootout would have led you to prove the film is 400 and our discussions about “Zone System is always half box speed” would lead you to set the EI on your spotmeter to 250, when you intend to meter by Zone System.

I haven’t figured out where the fallacy breaks, somewhere near Zone V it doesn’t matter if you use 250 or 400.

 

[ic-racer]

The story is a little more convoluted too, because near that same time I got a bunch of late-model Nikons with Matrix metering. That 'black box' system appears to add extra exposure in difficult situations. This DOES seem to give good negatives in my darkroom with HP5 at 400.

So, nothing to suggest the ISO rating is wrong, just with the new added features of my spreadsheet, (shifting the curve) turned out to be a very easy way to calibrate the system to match my old negatives and allow for similar interpretations of unknown films (specifically 200ft of Svema I just got) without any in-camera testing. I see nothing wrong with the mathematic provisions suggested by Stephen above, just that my spreasheet isn't configured to shift things that way.

Those are limitations from doing my spreadsheet in Apple Numbers. Which brings up another part of the back story: What ever happend to Aparat's software? I actually had given up on my own spreadsheet a few years ago, waiting for his software to appear. When I got the 200ft of Svema I needed to evaluate it and had to re-created my spreadsheet from scratch in Apple Numbers.