My Pepsi Challenge is all screwed up

Rich Ullsmith · Apr 24, 2010

I recently was given a Rollei 35, it's in fantastic shape and it even has a light meter built in. Me thinks hmm, maybe I should check it against my Sekonic spot meter.

I put out the gray card, set ISO at 400, f5.6, fill the frame, measure: 1/30th second.

Same settings and same angle with the spot meter: 1/15th second, exactly.

Okay I think, you can't expect a Sekonic digital and a P&S meter that is 40 years old to be identical. But to further muddy up the water, I pulled out the Contax G2. Set it up exactly the same. 1/8th second.

All batteries are fresh. Two stops difference. What gives here?

fotch · Apr 24, 2010

Daylight or electric lights? Daylight changes, the electric lights will be more consistent.

John Koehrer · Apr 24, 2010

How do you know you filled the frame with the gray card and a Rollei 35?

Chan Tran · Apr 24, 2010

It's hard to ensure that the gray card fills the frame with the Rollei 35. The G2 has TTL metering so it isn't difficult but it reads 2 stops different than you Sekonic then either your meter or the G2 or both are wrong. Sad but it must be true right?

phaedrus · Apr 24, 2010

It's time for a test roll of your favourite film through the Rollei 35 and the Contax G2. It's not unheard of to have a camera meter not calibrated to 18 % reflectance.

Trask · Apr 24, 2010

The exact angle that you read at can make a difference -- but a two-stop difference. All set to the same ASA setting? I'd suggest finding a large, uniformly-colored wall, preferably in shadow, and try with that. Eliminates the possibilities of shadows falling on the card, of your not knowing what the angle of reading is on the Rollei 35, and so forth.

Trask · Apr 24, 2010

Well, some older meters were calibrated to 12%, but none of the ones under discussion here.

Rich Ullsmith · Apr 24, 2010

It was under electric light, with maybe a little ambient through the window. Not much, like twenty minutes past sunset and it was overcast.

I took the azimuth of the Rollei meter into account and put it closer to the card, and made sure I didn't block any light. Same angle, too.

I'll pull the stuff out again and put it on the wall for uniformity. But still, two stops difference seems whacky. I'll report back later.

Rich Ullsmith · Apr 24, 2010

Okay, so I did this again with a white sheet on the dining table. Same f5.6, same ISO 400.

The Rollei and the Contax both stood at 1/250th. The Sekonic read 1/250th, plus 1/2 stop.

Obviously I was doing something wrong. And I suppose the 1/2 stop is not a big deal; they are basically P&S's.

Anyways, thanks for assisting in my photographic version of a Chinese Fire Drill.

Chan Tran · Apr 24, 2010

12% or 18% doesn't make 2 stops different. 1/2 stop is not bad.

benjiboy · Apr 25, 2010

The cameras meter inaccuracy is only theoretical until you take some pictures with it.

Stephen Benskin · Apr 25, 2010

Trask said:
Well, some older meters were calibrated to 12%, but none of the ones under discussion here.

Meters aren't calibrated to any reflectance / percentage. Let's not perpetuate myths.

Chan Tran · Apr 25, 2010

Stephen Benskin said:
Meters aren't calibrated to any reflectance / percentage. Let's not perpetuate myths.

No but they are calibrated with a factor K12 or K14 etc..
If a reflective light meter is said to be calibrated with a K14 it means that it was calibrated to read a luminance of 14 candela per square meter as EV0 for ISO 1. Or in other word, 0.14 candela per square meter as EV0 for ISO 100.

Chan Tran · Apr 25, 2010

To be fair the 14% or 18% reflectance does come in to play when one compares readings from an incident meter to that of a reflected light meter. An incident light meter with a flat receptor and calibrated to C250 would read EV0@ISO100 with an illuminance of 2.5 Lux. If one places a gray card of 18% reflectance and make a reflected light reading with a meter calibrated to K14 it would read EV0@ISO100. If one places a card of 16% reflectance a reflected light meter with a calibration factor of K12 would read EV0@ISO100.

Stephen Benskin · Apr 25, 2010

Chan Tran,

Yes, there is a lot we could get into. I just wanted to make a quick point without hijacking the tread.

As to the difference in light meters, calibration can get off, but added to that is the spectral sensitivity of the photo cell, and with older meters made before the mid-sixties, the color temperature of the calibration light source was 2700K whereas it later became 4700K. This can make a big difference depending on the color temperature under which a test is conducted. Anyone remember or remember reading that there used to be different speed ratings for incandescent and daylight? The calibration color temperature is the reason for it.

Steve

Mats_A · Apr 26, 2010

Chan Tran said:
No but they are calibrated with a factor K12 or K14 etc..
If a reflective light meter is said to be calibrated with a K14 it means that it was calibrated to read a luminance of 14 candela per square meter as EV0 for ISO 1. Or in other word, 0.14 candela per square meter as EV0 for ISO 100.

Wow! The quality and diversity of comments on this forum is amazing. And humbling. Where else could one get to know something as esoteric as this? I just love this site.

r

Mats

benjiboy · Apr 26, 2010

Stephen Benskin said:
Chan Tran,

Yes, there is a lot we could get into. I just wanted to make a quick point without hijacking the tread.

As to the difference in light meters, calibration can get off, but added to that is the spectral sensitivity of the photo cell, and with older meters made before the mid-sixties, the color temperature of the calibration light source was 2700K whereas it later became 4700K. This can make a big difference depending on the color temperature under which a test is conducted. Anyone remember or remember reading that there used to be different speed ratings for incandescent and daylight? The calibration color temperature is the reason for it.

Steve

As I wrote earlier the O.P needs to use the ******* camera and look at the results before assuming there's a problem.

Stephen Benskin · Apr 27, 2010

Chan Tran said:
To be fair the 14% or 18% reflectance does come in to play when one compares readings from an incident meter to that of a reflected light meter. An incident light meter with a flat receptor and calibrated to C250 would read EV0@ISO100 with an illuminance of 2.5 Lux.

Using a flat receptor can be a little misleading as the constant for dome needs to be used to calculate incident/reflective average reflectance approximation.. According to the ISO standard, that would be C=30 (30*10.76 = 323) as opposed to the C =250 or 320/10.76 = 23.23 and K = 1.16 (1.16*pi = 3.64.

3.64/30 = 0.12 or 12%

This is also based on the Sun being at an approximate angle of 41 degrees. This only works for the standard model for an average scene and it in no way means the meter "sees" 12% reflectance or that what the meter points at will reproduce that tone as 12% reflectance on the print.

The basic value of K is also 12.5 or 1.16*10.76 = 12.48 (rounded up to 12.5). This can be confirmed using the constants equation K = P/q as defined in Connelly,D, Calibration Levels in Films and Exposure Devices, Journal of Photographic Science, vol 16, 1968. Where P is Eg in the basic exposure equation Hg = Eg * t and q is the light loss constant in the classic exposure equation q*Lg*t/A^2 = Hg. P = 8 (rounded down from 8.11 to 8.2) and q=.65.

K = P/q or 12.48 = 8.11/.65

Stephen Benskin · Apr 27, 2010

Forgot to mention that through the lens meters don't have a K to speak of. They read the actual light transmitted through the lens and don't need one.

Rich Ullsmith · Apr 29, 2010

Yes, agree with Mats, thanks for letting me in on this discussion. The more I learn, the less I know.

Also agree with Ben. Real Pepsi challenge is putting some film through it, which is in the works. Thanks everybody.

My Pepsi Challenge is all screwed up

Privacy & Transparency

Privacy & Transparency