Interesting observation regarding light meters

[RalphLambrecht]

The majority of light meters are not made in the USA and likely were not in the past.

I agree;It's more likely ,the sun is calibrated to the Gossen standard.

 

[RalphLambrecht]

Agree wholeheartedly!

ending up with a different calibration for every camera-lens combination?

 

[RobC]

since we know different makes of meter use different K factors, they would all need to be using the K factor to meet the standard. Methinks that isn't the case and therefore they are not all calibrated to the Ansi standard. They may use the Ansi standard as their baseline but not comply with it exactly.

 

[Arvee]

Fred,

What happens if you meter off caucasian skin and place that reading on Zone VI.

Does incident light reading of that same scene agree with the reflected light reading placed on Zone VI?

I find that it does, and this makes me think a Zone VI reference placed on Zone VI is a better double-check for incident reading... Than a reading from an 18% gray card placed on Zone V where it does not belong.

Nope, all SBC type meters still recommend 1 stop less exposure. All Japanese meters indicate correct exposure, incident reading matches Caucasian skin reading placed on zone vi.

 

[Arvee]

since we know different makes of meter use different K factors, they would all need to be using the K factor to meet the standard. Methinks that isn't the case and therefore they are not all calibrated to the Ansi standard. They may use the Ansi standard as their baseline but not comply with it exactly.

I believe this is the issue; all the SBC type meters are exactly 1stop under in reflected mode. The offset must be incorporated in the Gossen calibration protocol for that family of meters, since it only appears in the SBC, F, Lux type meters and the early Sixtomat/Luna Pro Digital. All my other meters in the Gossen line (my early CdS Luna Pros and the Digisix) are calibrated differently and match the Japanese meters.

I don't have any of the current models and was hoping some of you folks out there might run a quick check on current production models. (Again, the issue is whether the incident readings match the reflected/gray card readings.)

 

[Arvee]

ending up with a different calibration for every camera-lens combination?

You got it! But, if you are aware of the anomalies, it is simple to compensate.

 

[David Allen]

The majority of light meters are not made in the USA and likely were not in the past.

Quite correct but, as the world’s biggest manufactures at the time, the influence of the standards that Kodak adhered to had a great inpact. Also, it is important to note that the creation of ISO 2720:1974 was the result of harmonising the ASA and DIN systems.

ending up with a different calibration for every camera-lens combination?

Well yes and no. Certainly for each system and possibly for lenses made in different eras.

When my dad started his photography school we discovered early on that, for example with the Olympus OM1 system, tests frequently revealed a one stop variation between different students’ EI for the same make of equipment, despite all other variables being the same. However, across the lens range there were no variables when using prime lenses (i.e if someone got EI 200 in the tests this worked across all prime lenses).

I recently had a student who uses a 1955 Rollei that tested at EI 400 for Tri-X developed in Thornton’s Two-bath. His Nikon FM3 tested at EI 200. I suspect that this may have something to do with the level of lens coating, accuracy of the shutter / aperture, etc. He has 3 prime lenses that all deliver the same results.

However, where one can not assume consistency is with prime lenses from differing eras. Relatively late in his life, James Ravillious (http://www.jamesravilious.com) finally had a little bit of money to buy equipment (a rare exception for him as he was always more poor than the poorest people he photographed). James bought some newer Leica equipment and just couldn't get on with it and so he contacted my father.

It was very quickly determined that the newer lenses were the problem. James had developed a style and exposure technique that worked because the older Leitz lenses had relatively few coatings and therefore were not so sharp and suffered some degree of lens flare that helped boost apparent film speed or let us say gave more light to the shadows. Further tests identified that some benefits from the newer lenses coatings suited James' style (better sharpness and bokeh, etc) but not his way of exposing.

In the end, with the help of the staff at Exeter Camera Exchange, they tested a number of newer (but not current) Leitz lenses and found one that, with the use of my father's own two-bath developer, gave James the results that he was looking for.

Bests,

David.
www.dsallen.de

 

[Arvee]

Here is another thread that discusses the same issue I discovered:

http://www.largeformatphotography.info/forum/archive/index.php/t-15599.html

Between the meters' differences in spectral sensitivity, non-linearity in the meters' response, and completely different calibration procedures at different color temps we are lucky to get a reasonable exposure! Pick a meter, learn its characteristics and don't change! What a mish-mash of variables.

 

[wombat2go]

The old CdS meters that I been refurbishing ( Pentax 3/21 and LunaSix) could not have been very accurate.
The CdS typically has a positive tempco , and its resistance will be about 10% lower at -5 Celsius compared to 35 Celsius.

Moreover the moving coil magnets have a negative tempco meaning that they would indicate higher for a given light level when colder.
The CdS and iron temco errors would be cumulative, and the meters had no attempt at compensation.

The mercuric oxide batteries (I read , but not much tech reference on internet) had anodes selected to provide
a constant voltage with time and temperature.

The scales were typically of 10 or 12 divisions.
Over say a temp range of 45 C , the error would be about 1 division which is almost 1 EV stop.
eg, variation between f/8 and f/11 when used outdoors from winter to summer here in Michigan.

So there is no point trying to use these for fractions of a stop, and I am sure they were not intended for any accuracy.

These meters contained no active electronics. To approximately obtain the logarithmic transfer function for EV,
they just used a series divider which gives a reciprocal function.
Then the meter scales were graduated in non linear increments to try to correct the error between the recip and the log.
They used 2 ranges ( two voltage divider strings) so that the EV graduations are spread across the spans for readability.

I crudely measured the angle of view of the LunaSix sensor in reflective mode to be about 90 degrees at the half EV points.
So to measure a gray card on this meter here, it would need to be 600mm (24 inch) diameter when held at 300mm away from the meter !

 

[Stephen Benskin]

Good information on exposure and exposure meters can be found in this thread - (there was a url link here which no longer exists)

There was a link to some of the seminal papers on exposure, but I'm not sure it's still working. 64.165.113.140/content/benskin/ Bill Burk was hosting it.

 

[Bill Burk]

Sorry about that, had a power failure couple weeks ago and forgot to restart it. http://64.165.113.140/content/benskin/ is online again.

 

[Bill Burk]

I still can't explain your findings but maybe this will help some:

In the wikipedia light meter article a simple formula is given for "Calibrated reflectance"

wikipedia Light_meter article

R = ( pi * K ) / C

Many light meter manuals can be found on Michael Butkus' site.

Butkus Light Meter manuals

Most of your meters don't specify the calibration constants. (Gossen meters explain the standard is 18% but that is counter to your findings).

A couple of your Sekonics do give calibration constants in their specifications

Example Sekonic 308S, Sekonic 428:

K = 12.5

C = 340 Lumisphere

R = ( 3.1415 * 12.5 ) / 340

R = 0.116, about 12%

The Sekonic 308S gives specification for the flat diffuser

C = 250 Flat diffuser

R = 0.157, about 16%

Gossen is said to use K = 11.37 and working backwards from what seems to be a 9% R, that would make C = 397

 

[Stephen Benskin]

Thanks Bill.

From the link http://64.165.113.140/content/benskin/

Defining K
An Interpretation of Current Exposure and Meter Technology
Calibration Levels of Film and Exposure Devices

From Defining K

 

[Bill Burk]

Thanks Bill.

From the link http://64.165.113.140/content/benskin/

Defining K
An Interpretation of Current Exposure and Meter Technology
Calibration Levels of Film and Exposure Devices

From Defining K

View attachment 98391 View attachment 98392 View attachment 98393

I'm moving away from that fixed IP address but these papers may be found at this mirror that I will keep alive:

http://www.beefalobill.com/benskin

 

[benjiboy]

This is the fastest way I know to drive yourself crazy "a man with one clock always knows what time it is, a man with several ones is never sure".

 

[Bill Burk]

Or a person with two websites will never know where their files are.

I got rid of the fixed IP so you can access the information on

http://www.beefalobill.com/benskin

The page gives written instructions how to reach the contents (web crawlers and robots can't follow instructions but people can)..

 

[Alan Klein]

So how to take reflective readings vs. incident readings into consideration when shooting chromes vs. negatives?

 

[wiltw]

If we simply FORGET what reflectivity percentage a meter is CALIBRATED to, and then DO REMEMBER that
18% tonality is simply 'the middle tone in the range of tones from black thru white' on a print,
we understand what 18% gray truly represents.

BTW, if I meter 18% target and expose with Canon d*g*t*l camera, the spike for the 18% tonal area appears just to the left of center midline of the histogram.
And if I meter 12% target and expose with Canon d*g*t*l camera, the spike of the 18% tonal read appear right on the center midline of the histogram

 

[Bill Burk]

So how to take reflective readings vs. incident readings into consideration when shooting chromes vs. negatives?

Take the case of reflective readings and black and white negative material.

My understanding is the calculated f/stop and shutter speed settings from a reflective exposure meter are intended to put at the film plane, 10 times the exposure of the speed point (where speed point is the exposure at the point where the characteristic curve crosses 0.10 above base + fog when developed to meet ISO parameters).

Now to move onto color slides, the speed point for a particular color slide film is figured out a different way. You want to protect the highlights. So although we are no longer literally putting "10 times the speed point" on the film (because the "speed point" is somewhere else)... a color slide film rated with the same ISO as a black and white negative film... will calculate the exact same result of f/stop and shutter speed in the same light... even though the films have different speed points. Because there was an effort to make different kinds of film compatible with all exposure meters...

 

[Europan]

Interestingly, all compare within +/- 0.1 stop in incident mode using the sun on an absolutely clear day. They all read within +/- 0.1 stop reading an 18% gray card in the same light, but, they all read exactly 1 stop less exposure than they do in incident mode.

Ultraviolet, the invisible. When you measure incident sunlight, more UV reaches the cell than when pointing towards a card on which it is diffused. Generally, photocells are much less sensitive to UV than photographic plates and films but about one EI point difference seems plausible to me.

Repeat the experiment with incandescent light and you will see that I am right.

 

[Bill Burk]

I checked your idea with a slight variation, instead of using incandescent light, which has a whole spectrum of differences from daylight...

I took an incident meter reading from inside my car through the dash window. It read f/16.2 (where the fraction is 1/10th of an f/stop). Out of the car it read f/16.6

This is 4/10 of an f/stop.

Now... I used a UV meter to verify that 0 UV passes my car windshield. While the outside reading was a UV index of 8 or 199 mW/sq.m

The windshield is not perfectly clear though, so visual light reduction could account for the different readings.

So I got out my SEI photometer.

I metered a white object across the street from inside my car... and then from outside my car.

The difference is about 1 mark, on the SEI one mark is 1/3 f/stop.

So .4 metered difference, less .33 visual difference leaves less than 1/10 f/stop which can be accounted for by the loss of UV light in this test.

I think my test supports a hypothesis that UV does not play a significant role in the difference in exposure recommendation found between Incident reading and Reflected light metering off an 18% gray card.

 

[Europan]

At which time of the day did you do that measurement?

 

[Bill Burk]

About 2 PM local time under slightly hazy sky.

I don't have the dome handy so I couldn't check your idea that it blocks UV. (I loaned my older Sekonic to a scout.) So I had to find a different way to figure this out and I remembered my UV meter always drops to zero in the car.

Invisible light does influence meters. Incandescent lights throw additional Infrared into the mix. Meter calibration is done using an incandescent light (filtered with approximately an 80B blue filter). I guess you could call the test light source a compromise between indoor and outdoor light.

So you could get different results indoors with incandescent versus outdoors in the sun. Also your test might have problems caused by "inverse square law" because in the sun, the inverse square law isn't as significant.

 

[darkroommike]

The 18% gray card was a printing standard originally. Each maker gets to choose their own calibration standard on the assembly line. And there are no ultimate standards in the field. Your film, your camera, your meter, your shutter, your lens, camera flare, your latitude, your longitude, your altitude, and even your attitude can affect your results. Roll all that up and then add your preference for negative density and contrast, your choice of developer and your choice of enlarger type. All you can do is find what works for you either with rigorous tests or a more casual approach of try it and see.

 

[Europan]

Agree