Interesting observation regarding light meters

Arvee · Jan 11, 2016

It seems to be common knowledge, you can find it most anywhere, that an incident measurement is equal to a reflected measurement of an 18% Kodak gray card. I always believed this to be true.

Well, in an informal experiment, I found that this isn’t necessarily true. I have a penchant for collecting light meters and I set out to see how my stable performed. Interestingly, my Gossen meters, Luna Pro F, Luna Pro SBC (2), Luna Lux SBC (2), Sixtomat Digital, 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.

I performed the same test with my Minolta IVF, Sekonic 308S, Sekonic 448 (2), Sekonic 398M, Sekonic 428, and my really old Luna Pro CdS (gray and black) meters and they all report exactly the same measurement (no difference between incident and reflected/18% gray card) and all within +/- 0.1 stop in incident mode.

Finally, the Gossen Digisix (not the Digisix 2) reports exactly the same reading for incident and gray card readings while all the older Gossen Silicon blue models show one stop less exposure in reflected mode when measuring an 18 % gray card.

Note: this is an informal experiment but there is definitely a pattern of differences between the older Gossens and the Sekonic/Minoltas, Luna Pro CdS and the new Digisix. I also used the Pentax V analog meter, which is spot on, to confirm the reflected measurement of the gray card.

Not suggesting anything here, just an interesting observation that explained why I got different results in the field when switching between incident/reflected measurements when using the Silicon blue Gossens, excluding the Digisix.

I’d be interested in hearing what other folks are seeing with their meters on this issue. As mentioned, this is an informal experiment but I believe there is a pattern that is too consistent to ignore. Perhaps Gossen had a change of heart regarding calibration philosophy during the 80s/90s when they intoduced the new Si blue meters?

Anyone know of a reason that might explain this?

DWThomas · Jan 11, 2016

(Thinking out loud here ...) I have some vague recollection of reading that the 18% assumption was varied between makers or over time? Dunno if that might account for it. I would go off on a tangent about age-related degradation of the diffusers as to color and transmission, but I would think that would would push the difference the other way.

Mostly an argument for testing and figuring out the idiosyncrasies of one's own particular equipment, I guess.

Edit: And what about spectral sensitivity outside of the visible, vs the gray card? Do some tests with IR or UV filtration? Oooooh, a project for some science students ....

wombat2go · Jan 11, 2016

Hi Fred, I am refurbishing a dead LunaSix using its CDS LDR sensor and its optics, and brewing some new electronics.

I did some measurements outside in the full winter sunlight with the sensor lens axis pointing directly at the sun.

With lens unfiltered (ie low range reflected), the LDR was 100 Ohm, corresponding to about 22000 Lux and EV 16 or so.

With lens filtered (ie high range reflected) the LDR was 520 ohm.

With dome and no filter (ie low range incident) the LDR was 570 Ohm, corresponding to about 3460 Lux

With dome and filter on ( ie high range incident) the LDR was 3000 Ohm corresponding to about 590 Lux

From that, the dome ( by itself) attenuates direct on axis sunlight by about 6.35 times which is about 2.7 EV stops.

But that is not the full story, because when the light source is to the side of the sensor's lens, ( off axis)
the dome actually increases the brightness on the sensor. (I did not measure with off axis angles)
So in the incident mode, the sensor is receiving light from a hemisphere around the axis of the lens.
In the reflected mode, the sensor's sensitivity is a cone shape , so it will depend on how close the 18% card is,
ideally the card should be large enough to cover the cone's base -easier with a Pentax spot meter.

I am sure in the development of these meters,
the priority would have been to get them predicting exposures that gave good results on actual photos.
A fair amount of art as well as the science.

-regards

RobC · Jan 11, 2016

Reflection(spot):
2Ev = B*S / K

Incident:
2Ev = E*S/C

Where:

B = Luminance
S = Film Speed
K = Exposure Constant (reflected)
E = Incident light
C = Exposure Constant (incident)
Ev = Exposure Value

See https://en.wikipedia.org/wiki/Light_meter for values of K and C

And note that K and C, especially K are NOT percentages.

For a reflection meter dividing something by 12.5 is the same as multiplying by 8%. So if you consider that 8% is approx 3 2/3 stops less than the reading which puts it in the middle of a 7 1/3 stop range. But it doesn't put it in the midlle of a 10 stop range.

In short, meters are not claibrated 18% reflection and never have been as far I can see. Well incident meters maybe but you'll need to do the maths to check. 18% is 2 1/2 stops less than 100% so would only be the middle of a 5 stop range.

Note the formula are current formula, older formula used in old light meters may differ.

Patrick Robert James · Jan 11, 2016

The "common knowledge" isn't really accurate. Using a grey card's value of 18% is one of those things that seems logical so people repeat it. IIRC different manufacturers used different standards. I believe 12% is the value some used which would generally be the difference you are seeing between the incident measurement and the reflective measurement off the 18% gray card.

Arvee · Jan 11, 2016

My QA Engr. wife thinks the change back might be the result of the ANSI standard of 1994.

Pioneer · Jan 11, 2016

I have been going on the assumption that if I use the same meter all the time, all my film will be exposed in a similar fashion. That way if I do find major differences then there is a good chance that something with the camera (shutter), or lens (aperture) is working differently than I assume it would. That is if I am using the same film.

There are occasionally small differences between cameras using the same film but it is usually the differences between medium format (Rolleiflex) and 35mm (Pentax.)

I suspect that my assumptions are probably off but it has worked relatively well. I suppose it should be mentioned that there is always an exception to any rule of thumb.

Sirius Glass · Jan 11, 2016

Fred Aspen said:
It seems to be common knowledge, you can find it most anywhere, that an incident measurement is equal to a reflected measurement of an 18% Kodak gray card. I always believed this to be true.

I am not sure that this is a valid assumption. The reflectance meters are based on the reflection of a gray card. An incidence meter is based on received light and adjusted accordingly. The incidence meter does not account for how a subject reflects light, only as it receives light.

MattKing · Jan 11, 2016

What method are you using with the gray card? At what angle are you holding the card when you take the reading?

How old is the gray card (they fade/discolour/change over time)?

And the "standard" for reflection readings is, indeed, closer to 12% than 18%.

Sirius Glass · Jan 11, 2016

My Gossen Luna Pro SBC and Sekonic L-308S both agree on reflectance readings; my Gossen Luna Pro SBC and Sekonic L-308S both agree on incident readings. However on the average the incident and reflectance readings differ by one stop.

wiltw · Jan 11, 2016

RobC said:
See https://en.wikipedia.org/wiki/Light_meter for values of K and C

And note that K and C, especially K are NOT percentages.

For a reflection meter dividing something by 12.5 is the same as multiplying by 8%. So if you consider that 8% is approx 3 2/3 stops less than the reading which puts it in the middle of a 7 1/3 stop range. But it doesn't put it in the midlle of a 10 stop range.

In short, meters are not claibrated 18% reflection and never have been as far I can see. Well incident meters maybe but you'll need to do the maths to check. 18% is 2 1/2 stops less than 100% so would only be the middle of a 5 stop range.

Note the formula are current formula, older formula used in old light meters may differ.

^^^
Several decades ago I got into a discussion with Ctein, who is a very knowledgeable photographer who holds a degree from Caltech. He explained the reflected vs. incident meter calibration equation, and the fact that the manufacturer GETS TO CHOOSE the value of C or K constant in the equation, and that the target reflection of about 12.5% is what one needs to meter in order for 18% grey to end up a true 'middle' tone.

If one realizes that 'reflected light metering target' is not the same as 'middle of the tonal range between black and white', the need to meter 12% tone in order to cause the 18% one to be 'in the middle' is not contradictory.

markbarendt · Jan 11, 2016

Fred Aspen said:
It seems to be common knowledge, you can find it most anywhere, that an incident measurement is equal to a reflected measurement of an 18% Kodak gray card. I always believed this to be true.

Common knowledge = urban myth. The instructions with the gray card specify an adjustment to the reflective meter reading to find the camera setting. That is essentially Kodak saying "a direct gray card reading with a reflective meter will not match an incident meter reading".

All reflective measurements have that problem. The advantage of using a gray card is that it is a "known target" where the reflectivity and offset to the camera setting are "known". That is in contrast to "unknown targets" like the dirt under a bush.

Read the instructions for the gray card. Follow those instructions carefully (get angling right and do the math specified).

Take an incident reading at the gray and the results (after the math) should now be very close, we are human so probably not perfect.

Arvee · Jan 11, 2016

Sirius Glass said:
My Gossen Luna Pro SBC and Sekonic L-308S both agree on reflectance readings; my Gossen Luna Pro SBC and Sekonic L-308S both agree on incident readings. However on the average the incident and reflectance readings differ by one stop.

That's where I am seeing the difference. My SBCs and the 308 agree on incident measurements; when I measure the gray card the SBCs (both of them) tell me to close one stop compared to the 308 reading.

I just found it really odd that the old CdS Luna Pros agreed incident vs reflected mode. Then the Si blue came on the market and they disagree by one stop. But the latest meters are back to incident vs reflected mode agreement. When I used a 'hybrid' metering strategy with, for example, and SBC, it became confusing as to which mode to believe and my results were inconsistent. With the Sekonics, the measurements would confirm each other rather than disagree with each other. It took me a while to sort out the problem.

Arvee · Jan 11, 2016

I am aware of the 12% vs 18% argument which is about a 1/2 stop we are quibbling over.

The differences in the various K factors that mfrs. can choose only accounts for 1/6 of a stop at most.

The difference I am seeing is consistently a full stop in the other direction; the Gossens are telling me that the gray card is 36%, essentially the same as a caucasian skin tone, Zone VI.

All the Japanese meters tell me a gray card is 18% (and my exposures tell me that the Japanese meters are correct and the German meters are telling me to underexpose by one stop and the negs are showing me they are under exposed.

RalphLambrecht · Jan 11, 2016

all I can confirm isthat my arsenal of Gossen meters are all within 1/3 stop iat any lighting condition. a few years back one was an outsider but came back pergfect from a Gossen adjustment with a changed serial number!:confused:

Arvee · Jan 11, 2016

RalphLambrecht said:
all I can confirm isthat my arsenal of Gossen meters are all within 1/3 stop iat any lighting condition. a few years back one was an outsider but came back pergfect from a Gossen adjustment with a changed serial number!:confused:

Hi Ralph,

Here's the rub, perhaps you have an answer. All my Gossens agree with each other (both incident and reflected); all my Japanese meters agree with all my German meters in incident mode. When it comes to reflected measurements, the Gossens all say that the gray card is one stop brighter than Japanese meters tell me. The Japanese meters all agree with each other within the family.

What I want to know is why the difference. Differences in calibration protocols don't explain the discrepancy.

RobC · Jan 11, 2016

wiltw said:
^^^
Several decades ago I got into a discussion with Ctein, who is a very knowledgeable photographer who holds a degree from Caltech. He explained the reflected vs. incident meter calibration equation, and the fact that the manufacturer GETS TO CHOOSE the value of C or K constant in the equation, and that the target reflection of about 12.5% is what one needs to meter in order for 18% grey to end up a true 'middle' tone.

If one realizes that 'reflected light metering target' is not the same as 'middle of the tonal range between black and white', the need to meter 12% tone in order to cause the 18% one to be 'in the middle' is not contradictory.

But the point is that the K factor in a reflection meter of 12.5 which is the standard today is not a percentage of 12.5% it is factor which if you convert it to a percentage it becomes equivalent to 8% and NOT 12.5% or 18%.

Then 8% of the reading is 3 2/3 stops less than the actual reading. i.e. 3 2/3 stops down from how much light there is. So the K factor in a reflection meter is placing the exposure in the middle of 7 2/3 stop range wen the K factor is 12.5

For an incident meter the the calcs are different becasue E is measured differently and i haven't looked at that in detail. But the C factor is doing essentially the same but may not match a reflection meter.

Arvee · Jan 11, 2016

According to Jeff Conrad's White Paper, specified K factors (from 10.76 to 14) only account for 0.16 stop variability maximum.

I am not entirely clear on your explanation, Rob, but I think you're saying the K factor has a great deal more influence than 0.16 stop.

Check this out: http://www.largeformatphotography.info/articles/conrad-meter-cal.pdf

and this: http://dpanswers.com/content/tech_kfactor.php

Bill Burk · Jan 11, 2016

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.

RobC · Jan 11, 2016

Fred Aspen said:
According to Jeff Conrad's White Paper, specified K factors (from 10.76 to 14) only account for 0.16 stop variability maximum.

I am not entirely clear on your explanation, Rob, but I think you're saying the K factor has a great deal more influence than 0.16 stop.

Check this out: http://www.largeformatphotography.info/articles/conrad-meter-cal.pdf

and this: http://dpanswers.com/content/tech_kfactor.php

I'm saying a modern reflection meter is calbrated to an 8% reflection and NOT 12.5% or 18%.

But this highly controversial for most people who are into the numbers game.

Alan Klein · Jan 11, 2016

I noticed the same thing on my Minolta IIIf; a difference of one stop. I noticed similar difference on an APP for my Samsung Galaxy cellphone. I was curious at the time and seem to recall the Ansel Adams played some hand in getting the industry to use the 18% gray card calibration model instead of the 12% because of the way he exposed for BW film. I might not have all the particulars right, but others could do the research.

Sirius Glass · Jan 11, 2016

RobC said:
I'm saying a modern reflection meter is calbrated to an 8% reflection and NOT 12.5% or 18%.

But this highly controversial for most people who are into the numbers game.

Given that latitude of today's film as long as it is within one stop it does not matter. Zonistas need not apply.

David Allen · Jan 12, 2016

Sorry but the ‘truism’ (meters are calibrated for 18% grey) - that has been repeated down the ages - is plain wrong. Light meters are calibrated to an ANSI standard. This standard is related to a specific luminance value that can be roughly described as being the equivalent to the reflectance of 12% grey. The reason that it is linked to luminance is because it can be correlated to a measurable amount of light energy whereas reflectance will vary according to metering position, position of light source and angle - all of which vary the amount of light energy bouncing off any particular object. Because the meters are calibrated to 12% ANSI standard some people have taken to calling this the ‘K’ factor.

The reality is that meters are calibrated to luminance that is equivalent to 12% grey and Grey Cards are calibrated to a reflectance of 18% grey. The purpose of Kodak’s grey card was to deliver a constant that then needs interpretation. It was never intended as a substitute for a given scene, nor as a matched equivalent to a meter’s calibration and definitely not as an equivalent of Zone V.

If you meter a grey card with an ANSI calibrated meter you will be off by 1/2 stop plus any other errors that you might have such as incorrect EI, poor metering technique, poor assessment of the scene’s lighting, etc.

Far better is to do your own tests (thereby calibrating everything to your meter, metering technique, your equipment, your choice of film/developer, your way of agitation and your preferred way of printing).

None of this is rocket science! - you just need to expose the film sufficiently to achieve detail in the shadow areas that are important to you and develop so that you retain sufficient detail in the bright highlights that are important to you.

The variations that some have observed between differing light meters can be affected by a number of things including the meter's angle of view, colour sensitivity, accuracy of initial calibration, whether it is a 'normal' field meter or a spot meter (which can introduce variations such as flare that arise from the optics being used in the meter - just try metering a scene with a spot meter and then fit the inner tube from a toilet roll as a lens hood and observe how much a difference that can make).

Bests,

David.
www.dsallen.de

AgX · Jan 12, 2016

David Allen said:
Light meters are calibrated to an ANSI standard.

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

Arvee · Jan 12, 2016

David Allen said:
Far better is to do your own tests (thereby calibrating everything to your meter, metering technique, your equipment, your choice of film/developer, your way of agitation and your preferred way of printing).

Bests,

David.
www.dsallen.de

Agree wholeheartedly!

Interesting observation regarding light meters

Moderator

Privacy & Transparency

Privacy & Transparency