Feedback Please: The Kodak 18% Grey Card and Metering, a new look.

[Bill Burk]

Another fun calculation you can work off that Kodak chart is the hypothetical... What would K have to be if the meters were calibrated to 18%.

Taking the formula N squared / t = ( L times S ) / K and substituting...

32 squared / 1/30 second = ((16,000 lux times 0.18 ) times 400) / K

30,720 = 1,152,000 / K

30,720 K = 1,152,000

K = 37.5

So K would have to be 37.5 if the meter were to be calibrated to 18%

This must be wrong because Wikipedia article says K = 14 would be 18% (with C = 250).

Anyone know how to compute reflected light given incident light and reflectance? I think my calculation is wrong there. Maybe need to use PI at that step of the equation?

 

[RobC]

everything is based on ISO 100.

reflectance in what units

 

[RobC]

incident light * reflected %age

L should NOT be multiplied by 0.18 as you have done your conversion. Dividing by K does that so you are doing your conversion wrong.

The value you have for L is the reflectance. You already have the number but seems you don't know what it is. i.e. 16000 is reflected light (the formula you show relates to reflected light meter and not an incident meter).

AND

the standard reflection light meter formula value for B is in candelas and not Lux so you must convert Lux (if that is what you are actually using) into candelas in your formula. That means you divide it by 10 so your reflection value is actually 16000/10.

Are you really sure your meter is measuring reflectance in Lux. I don't think so becasue Lux is a measure of illuminance(incident) and not Luminance(reflectance).

So get your units right otherwise what you you write in your formulas is meaningless to me. You must repost your formula when you have actually understood what the constutuent components of the standard formulas actually represent using the correct units so we know what your starting point really is.
And don't invent spurious scenarios which don't represent anything meaningful otherwise they are non understandable.

 

[Bill Burk]

16000 lux is given as the incident light for an example row in the Kodak chart.

I tried to substitute (incident lux from the chart times 0.18 gray card reflectance) as a replacement for L (or B), but as you say the units are wrong so that has to be reworked.

Meanwhile you should take a look at that Kodak chart because it illustrates the relationship between meter readings and incident lux.

 

[RobC]

No I don't think I will look at the Kodak Chart thank you.

The formula for a light meter is in the the light meter and that is the formula which calculates the 8% midpoint based on K = 12.5 or 14 or whatever it is for your particular meter.

You must understand that old meters used non SI units and K values are different by a conversion ratio from Foot Lamberts (If thats what your meter uses) to Candelas for modern Reflection meters using Candelas. So K = 12.5 on a modern meter using Cd/M^2 would be K = 3.6 on an old reflection meter using Foot Lamberts. Getting units correct is critical to making any sense of it.

If you are using some knackered old meter using Foot lamberts and you know it used K = 3.3 then just multiply 3.3 by 3.4 (conversion to candelas) which gives you 11.22 K equivalent in modern SI units meter and then divide 100 by 11.22 to tell you what %age reflection its calibrated to place on midpoint of curve. So 100/11.22 = 8.9%.

So if you know what the K factor of your knackered old meter is you can work out what it equates to in the far more intuitively meaningful SI units.

 

[Stephen Benskin]

This must be wrong because Wikipedia article says K = 14 would be 18% (with C = 250).

Anyone know how to compute reflected light given incident light and reflectance? I think my calculation is wrong there. Maybe need to use PI at that step of the equation?

 

[Bill Burk]

Here's the example row I took from the chart.

F/stop - Luminance cd/ft squared - Luminance footlamberts - Illuminance footcandles - Illuminance Lux

Formula:

A squared / t = ( L times S ) / K

Values:

A = 32 taken from that row of the chart

t = 1/30 second given in the instructions for the chart

L = 270 footlamberts converts to 925 candela per square meter

Or you can do the conversion from Lux as (16000 Lux times 0.18 percent reflectance of the Kodak 18% gray card from which reading was taken) / PI = 917 candela per square meter

S = ISO 400 the speed given in the instructions for the chart

K = unknown.

32 squared = 1024

1024 divided by 1/30 = 30720

925 candela per square meter times ISO 400 = 370,000

Leaves us with values in our formula

30,720 = 370,000 / K

multiply both sides by K

30,720 K = 370,000

K = 12.04 (or K = 11.94 when using the calculated conversion from Lux)

Now, where Wikipedia article says I should get K = 14, this is closer now and is a more reasonable solution.

 

[RobC]

I'm not interested in that formula. Its not a light meter formula that I use.

And don't quote wikipedia and expect me to know what its talking about. I'm not bloody telepathic. If you don't post references and get it right no one has a clue what you're talking about.

 

[Bill Burk]

https://en.wikipedia.org/wiki/Light_meter

This is the formula I am using

The Kodak guide to Estimating Luminance and Illuminance is an external link from that article.

 

[RobC]

straight away that page syas L is in Cd/M^2 so why are you using Lux values. That changes what the K factor will result in being. Get it right and be consistent otherwise everyone needs to check what units you happen to be using on this occasion and do the conversions whichh you should already have done and made clear so that people know what you're talking about. Flipping between one article and another in different units makes it worse and if you can't be bothered to be consistent with your units then neither can I be bothered to do it for you.

AND

just so you are sure which units are best to use you should note that the standard formula of a reflection light meter today is:

2^Ev = (B*S)/K

2^Ev gives its results in CD/m^2 so it follows that B must always be in Cd/m^2 otherwise the formula is invalidated unless K is also converted to relate to units B is in. Its common sense to keep it all in same units as 2^Ev. i.e. Cd/m^2 then less confusion happens and it is far more intuitive.

 

[Bill Burk]

straight away that page syas L is in Cd/M^2 so why are you using Lux values. .

L = 270 footlamberts converts to 925 candela per square meter

I'm converting to candela per square meter

 

[Stephen Benskin]

K is a constant that takes into consideration a number of variables and is an expression "as a function of the lens transmission, spectral characteristics of the detector and the discrimination of the field luminance measurement." (Re-evaluation of Factors Affecting Manual or Automatic Control of Camera Exposure) In order to understand K, it's important to understand how it relates not just to metering, but to camera exposure. There is an important relationship between K and two other constants, q and P. In fact, K can be defined as K = P / q.

To make the information more manageable, it will be presented in a number of installments. BTW, what has happened to the sub and super script formatting?

The camera’s optical system is not a lossless one. The amount of light exiting the lens is less than what can be attributed to the f/Stop and shutter. This light loss is caused by a number of variables. In order to produce the desired amount of exposure at the film plane, the luminance value entering the lens needs to compensate for the loss. In order to do this, a constant is added to the camera exposure equation. The constant is known as “q”. D Connelly in Calibration Levels of Films and Exposure Devices defines q as a factor which takes into account the light transmission characteristics of the lens system. Its value is derived from the variables:

Distance from lens to object
Focal length of the lens
Lens transmittance
Camera flare correction factor (at the average mean)
Vignetting factor
Angle of image point from axis of lens

Through the lens metering can measure the actual value of q. A hand held meter cannot. It must incorporate an average value. An acceptable value for q has been determined to be 0.65. So the exposure equation of

(Lg * Sa* t) / A^2 = Hg

becomes

(q * Lg * Sa * t) / A^2 = Hg

Where:
Lg = scene mean Luminance value
t = shutter speed
A = f-number
Sa = film speed
Hg = film plane mean exposure

Most of the K equation is comprised of the q equation. The other variables fall into two categories.

 

[Bill Burk]

B in candela per square meter can be found taking Incident light in Lux multiplied by Reflectance percentage then divide by PI.

So a Sunny 16 light situation which is commonly considered 81920 Lux can be translated to B for a given Reflectance percentage by:

81920 Lux times 0.157 = 12,861

12,861 divided by PI = 4096

So B = 4096 candela per square meter in Sunny 16 conditions with a 15.7% reflectance.

Now solve for EV

4096 times 100 = 409600

409600 divided by K 12.5 = 32768

2 to the 15 = 32768

 

[RobC]

And K= 12.04 = 8.3% of what was metered.

 

[Bill Burk]

So you are saying 0.083 reflectance is the calibration?

81920 Lux times 0.083 = 6,799

6,799 divided by PI = 2,164

So B = 2,164 candela per square meter in Sunny 16 conditions with an 8.3% reflectance.

Now solve for EV

2,164 times 100 = 216,400

216,400 divided by K 12.5 = 17,312

2 to the 14 = 16,384

So now your meter reads EV 14 ?

 

[RobC]

the film speed in the Kodak chart is 125. Not valid as formulas based on 100 speed film.

Please adjust.

Besides what are you trying to say, that my meter is wrong and it doesn't use the B*S/K formula? I can assure you it does.

So quoting some different old formula from elsewhere and saying it doesn't match proves absolutely bugger all.

 

[Bill Burk]

The film speed in the Kodak chart instructions is 400

The chart is used with a light meter shutter speed 1/30 and you are instructed to take a light meter reading off an 18% gray card, the indicated f/stop refers to a row in the chart.

From that you will lookup estimated luminance and illuminance.

You can use these values of luminance and illuminance in your formula. Luminance and illuminance values are not tied to ISO speeds.

 

[RobC]

I thought we were looking at exposure diagram above.

 

[RobC]

my meter tells me what the cd/m^2 values are from the EV values it outputs if I use it 100 speed and only 100 speed.

And that kodak article says figures are approximate and your conversions are rounded so results will be approximate.

So what exactly are you trying to say because I really have no idea what you're trying prove.

F32 for 1/30 second is EV 15

 

[RobC]

All I have said is that for K = 12.5

(B*S)/K = (B*S)*0.08 = (B*S) * 8%

OR

If your reflection meter uses K = 14 (which mine does) then

(B*S)/K = (B*S)*0.0714 = (B*S) * 7.14%

 

[Bill Burk]

I believe that the Kodak 18% gray card readings need to be opened by 2/3 stop to indicate the same f/stop and shutter speed for the same film speed as you might get directly from your meter in incident mode. I believe the referenced Kodak article gives some concrete examples that are better than other web explanations about the myth of the 18% gray card have offered.

Most of the other web explanations about the myth of the 18% gray card say that you need to use a 12% gray card to make incident meter agree with gray card reading. I believe them and my calculations offered here support the 12% gray card generally. 8% sounds wrong to me as being the reflectance value of a gray card that would help people get the same reading with incident meter reading and reflected meter gray card reading.

It may help to demonstrate the different gray cards you would need with different values of K, for example it's been said that K = 14 would equate to 18% gray card and K = 12.5 equates to 12% gray card. It would be helpful to compile a list of these.

 

[RobC]

I think you forget that all a reflection meter does is put the reading in the middle of the curve. But that is only true for what?
Well it only true for a curve which records 7 1/3 stop range when K is set to 12.5 becasue K of 12.5 is 8% of 100% and 8% is 3 2/3 stops less 100% which is in the middle of 7 1/3 stop range.
So then we can argue about whether 7 1/3 stop range is really correct or is it actually 7 2/3 stops or 8 stops or 7 stops and if we don't get our dev to match the ISO test that the manufacturer did exactly then it really isn't accurate enough to be arguing over it anyway. We are differing by about a 1/3 of a stop. Are we talking about curve above fb+fog? That would be about a 1/3 of a stop. No I don't think so becasue the K adjustment is down from metered and not up from zero.
And how many people have aligned a kodak card accurately using a sextant or clinometer before taking a reading from it. My guess to within 0.1% would be zero. So what have you proved? well bugger all as far as I can tell.
You can not deny that a K =12.5 using the standard meter formula is 8% of the reading.

 

[Bill Burk]

It's true that giving an exposure of EV14 for Sunny 16 conditions will make good negs.

But the issue to me is when people doubt themselves because they can't get their meters to agree, when really, they do agree.

 

[RobC]

lets go at this from a different angle.

We'll asume that film will capture 7 1/3 stops of subject range and print exactly onto B+W paper full a full range of tones from pure black to white.

Now if we meter from an 8% card, at what density would you expect if to appear on the film curve? And at what density when metering from a 12% card and at what density when metering from an 18% card.

I just want densities and NO FORMULAS. A very simple question which does not require a contrived complex answer.

 

[Bill Burk]

I'd get a density of 0.8 for all three if I used the recommended meter reading.

But if I'd placed the 8% reading on Zone IV it would be 0.6 density on the negative

And if I adjust the 18% gray card reading by opening 2/3 stop it would be 0.9 density on the negative.