variance in camera light meters

David Lyga · Jan 15, 2014

I have a Canon AE-1 and an Olympus XA. Both have working meters. Both deliver identical and accurate readings for daylight situations with, of course, the same film speed. But, oddly, the XA shows two to three stops less exposure needed under incandescent lighting than the Canon requires.

Now, most of us know that color film requires about two stops less exposure under incandescent lighting (without the blue filter) than the same speed of traditional B&W film does under that same incandescent situation. This leads me to come to a possible (but erroneous?) conclusion that the Canon metering is optimized for traditional B&W negative exposure and that the XA is optimized for color (slide?) film. Comments? The XA metering is from a CDS cell and, I believe that the AE-1's cell is silicon blue photodiode. - David Lyga

David Lyga · Jan 15, 2014

No, Michael, I closed all options for variance other than the actual meter responses in question. I metered 'normal' scenes in both cases: outdoors, a park scene and, indoors, a room lit by a ceiling tungsten light. In all cases the field of view was virtually identical, obviating the possibilities of different reflectance values coming into the equation. I could not come to another conclusion than the one I posited. But, your assessment that different types of cells deliver different spectral responses is something that should be discussed more fully, as the manufacturers have not done that, ever, as far as I can see. - David Lyga

David Lyga · Jan 15, 2014

Yeah, but don't you think that the manufacturers should have offered this information freely? Amazingly, the 'public' seems to have never voiced concern with these spectral differences. It just seems that it is sufficiently important an issue to have been parsed and discussed at length, and, herefofore, seems to have been not even 'noticed'! Thank you Michael. - David Lyga

BrianShaw · Jan 15, 2014

David Lyga said:
Yeah, but don't you think that the manufacturers should have offered this information freely? Amazingly, the 'public' seems to have never voiced concern with these spectral differences. It just seems that it is sufficiently important an issue to have been parsed and discussed at length, and, herefofore, seems to have been not even 'noticed'! Thank you Michael. - David Lyga

These data (spectral response characteristics of metering cell technologies) are freely available. The "problem" is that they exist in supplier specifications and engineering journals. One must not only know how to access them, but know how to read and interpret that level of knowledge. I have seen such discussion in books on photographic metering too. As for the manufacturers, as long as their product meets their customer requirements (a decent exposure most of the time) then they have no need in educating on the underlaying engineering... escpeciall with regard to how it compares to a competetors. This is true of all consumer products. It is even mroe true when we are talking historical products.

Chan Tran · Jan 15, 2014

If anything I have to increase exposure for incandescent lighting with color film unless it's balanced for tungsten. Because the blue layer is underexposed color correction is not possible. Overexposure ensure adequate density in the blue layer. Extra density in the red layer doesn't matter much it can be filtered out during printing.
I seriously think the XA is wrong and yet when the XA meter is wrong it could still deliver correct exposure because (like the OM-2) it has a meter circuit just for displaying and another totally independent circuit to do auto exposure control.

bernard_L · Jan 15, 2014

Is it the spectral response (against different spectral contents of incandescent vs daylight) or is it that one of the two meters is deviating from true response at low light levels (artificial light is typically much weaker than sunlight)??

To sort this out you can try
(1) with both cameras, meter in sunlight a piece of red paper, fabric, etc... Do they still agree?
(2) with both cameras, meter incandescent lighht at high intensity, e.g., staring into bare bulb at same close distance. Do they still disagree (as much)??

Would be interesting if you report the results here...

E. von Hoegh · Jan 15, 2014

bernard_L said:
Is it the spectral response (against different spectral contents of incandescent vs daylight) or is it that one of the two meters is deviating from true response at low light levels (artificial light is typically much weaker than sunlight)??

To sort this out you can try
(1) with both cameras, meter in sunlight a piece of red paper, fabric, etc... Do they still agree?
(2) with both cameras, meter incandescent lighht at high intensity, e.g., staring into bare bulb at same close distance. Do they still disagree (as much)??

Would be interesting if you report the results here...

It might be more telling to meter red, blue, and green, sunlight and tungsten - be very certain you actually have lightbulbs with filaments, not lookalike CFLs. Another informative experiment would be to meter in sunlight, both direct and in open shade, with and without a strong UV filter on the cameras.
You might be surprised at what you see.

snapguy · Jan 15, 2014

get us

Mechanical and electronic gizmos hate us and are out to get us. Their plot is working.

E. von Hoegh · Jan 15, 2014

snapguy said:
Mechanical and electronic gizmos hate us and are out to get us. Their plot is working.

Not all of us. I get along just fine with my widgets. But then, I know how they work - that's the key.

BMbikerider · Jan 15, 2014

The variation in light levels may not have anything to do with the differences. How exactly is the light getting to the metering cell? If they do NOT strike the cell at exactly the same angle, and intensity there is almost certainly going to be a difference.

dmb · Jan 15, 2014

The shutter speed shown in the XA viewfinder is not directly linked to the actual shutter speed that is used. It is operated by a separate circuit from that controlling the electromagnetic shutter. My old XA always shows about 2 stops slower speeds than a separate meter would indicate however the actual exposures are fine. It is possible to adjust the viewfinder meter but it is NOT for the faint hearted.

E. von Hoegh · Jan 15, 2014

dmb said:
The shutter speed shown in the XA viewfinder is not directly linked to the actual shutter speed that is used. It is operated by a separate circuit from that controlling the electromagnetic shutter. My old XA always shows about 2 stops slower speeds than a separate meter would indicate however the actual exposures are fine. It is possible to adjust the viewfinder meter but it is NOT for the faint hearted.

I was beginning to wonder about how the cell was mounted on the XA - it's not TTL metering is it?
A more meaningful comparison could probably be made between say an original LunaPro (cds cell) and a LunaPro SBC. Calibration, metering angle, etc should all be the same, the only significant difference would be the spectra response of the cells themselves.

gone · Jan 15, 2014

I wouldn't think the Canon is optimized for B&W from my experience (that's all I shoot). I do not like/trust the Canon metering because I had exposures all over the place on every AE-1, A1, AE1P camera I ever owned, and I tested the shutter speeds on all the cameras before shooting them. I finally gave up and bought an FTb and it's been spot on w/ the exposures. I think the Canon system is good for color, but not necessarily B&W film. Of course these are old cameras and who knows how they behaved back when they were new. For what it's worth, every Nikon or Nikkormat I've owned didn't have any exposure issues. I trust the Nikon meters, not so the Canons.

mopar_guy · Jan 15, 2014

David Lyga said:
I have a Canon AE-1 and an Olympus XA. Both have working meters. Both deliver identical and accurate readings for daylight situations with, of course, the same film speed. But, oddly, the XA shows two to three stops less exposure needed under incandescent lighting than the Canon requires.

Now, most of us know that color film requires about two stops less exposure under incandescent lighting (without the blue filter) than the same speed of traditional B&W film does under that same incandescent situation. This leads me to come to a possible (but erroneous?) conclusion that the Canon metering is optimized for traditional B&W negative exposure and that the XA is optimized for color (slide?) film. Comments? The XA metering is from a CDS cell and, I believe that the AE-1's cell is silicon blue photodiode. - David Lyga

What kind of incandescent light?
Regular household tungsten light bulbs will be more red than Photofloods. Must film makers that reference a tungsten speed are refering to Photoflood bulbs and not household type bulbs.

David Lyga · Jan 15, 2014

Chan Tran said:
yet when the XA meter is wrong it could still deliver correct exposure because (like the OM-2) it has a meter circuit just for displaying and another totally independent circuit to do auto exposure control.

This is facinating to read. Thank you, Chan Tran. Also, thank you dmb for, essentially the same input. - David Lyga

mopar-guy: a couple of years ago I hoarded three hundred 100W bulbs and placed them in storage because I was terrified of the flourescent assault. I never regretted that treachery. 100W bulbs did the lighting. Towards the end of the year that they were last going to be legally saleable, they were being almost given away by stores having amazing sales. - David Lyga

momus: apparently that AE-1 IS optimized for B&W because it reads like B&W film reads light. With color, it would overexpose under the unfiltered tungsten lighting. With the proper blue filter I am not so sure. - David Lyga

both bernard_L and E. von Hoegh: I will try the added intensities of tungsten light and, to the best of my ability, the different colors. Will report here possibly tomorrow. - David Lyga

benjiboy · Jan 15, 2014

snapguy said:
Mechanical and electronic gizmos hate us and are out to get us. Their plot is working.

snapguy, your real name isn't Woody Allen is it ?

David Lyga · Jan 16, 2014

OK, I did the tests. These is a congruity with higher levels of tungsten lighting: both match. The different colors read do not seem to matter, only the light intensities. Apparently that CDS cell on the XA did not respond well to low tungsten (probably low daylight also) levels.

I don't want to hype this too much as there are profound variances to be found with meters on cameras that are decades old. But it would be interesting for all to compare their various light-meters and see what happens. One does not have to even use up film, as the readings are revealing. - David Lyga

David Lyga · Jan 16, 2014

I guess we'll have to leave it at the 'pragmatic shortcoming', Michael, as I am hardly that sophisticated. But, you are most likely correct and maybe some will delve further into this. - David Lyga

Chan Tran · Jan 16, 2014

I made comparison most of the time. I found many of the old timers camera meter response are non linear.
I use the dichroic color head as light source. Dialing in equal amount of filters made neutral density. Monitoring the filtration tightly with a color analyzer keeps the color balance from shifting when adjusting the filters. Use a good reference spot met the to check light level.

fotch · Jan 17, 2014

There is an old saying, something like a person with two watches (clocks) will never know the exact time. Just saying......

benjiboy · Jan 17, 2014

"If a man has a clock he knows what the time is, if he has several he's not sure". The quickest ways I know of a photographer to drive himself crazy is to compare light meters, and thermometers.

RalphLambrecht · Jan 17, 2014

in photographythere are countless varibles and everything depends on everything else.it is a real miacle we get any pictures at all.it is a wonderfultechnology.

E. von Hoegh · Jan 17, 2014

fotch said:
There is an old saying, something like a person with two watches (clocks) will never know the exact time. Just saying......

That saying is wrong. No timepiece keeps a perfect rate, therefore nobody knows exactly what time it is. The closest the average person can get is to listen via shortwave to the WWV time signal, but you'd better apply a correction to allow for the time it takes for the radio signal to propogate from the transmitter to your reciever, and then another correction for the distance between the radio's speaker and your ear (unless you're watching the signal on a 'scope).
A watch or other timepiece that has stopped is at least correct twice per day (assuming it has a 12hr dial, once per day if a 24hr dial) but you'll never know precisely when.

DREW WILEY · Jan 17, 2014

I keep several matched meters. When one of them reads differently anywhere over the entire scale, I sent it in to Quality Light Metric for
recalibration (about $100, but only needs to be done about once every 10 yrs).

E. von Hoegh · Jan 17, 2014

DREW WILEY said:
I keep several matched meters. When one of them reads differently anywhere over the entire scale, I sent it in to Quality Light Metric for
recalibration (about $100, but only needs to be done about once every 10 yrs).

But, if you use multiple meters, you'll need a minimum of three identical meters. If you have two, and they disagree, you have no way of knowing which is wrong. Since the odds are strongly against two meters drifting out in the same direction the same amount, three meters means when two agree and one does not, the odd one gets recalibrated.
I presume that by "several" you mean "more than two".

variance in camera light meters

Privacy & Transparency

Privacy & Transparency