Kodak grey card usage

[Bill Burk]

The "calibration light" is a red herring IMHO.

Well, the calibration light being a backlit screen explains why it is not a reflectance percentage that you are calibrating to. So it's a mistake of using the wrong words if you say "a light meter is calibrated to 12% reflectance". That's all I'm saying. It helps answer the question "What do you mean when you say a meter is not calibrated to a certain percent of gray?".

 

[Diapositivo]

I think I have got a resumé of the conclusions so far.

Hm for black and white film is an empirical, observed data. If the film is ISO rated, let's say ISO 100, then we know that Hm is 8 / ISO and so for each ISO Hm is a quantity which is known a priori, in this case 0.08 lux.seconds, or LogHm = -2.1.

Now if Hg is defined as Hm * 10, and this definition must occur somewhere in the ISO standard, although I am not sure light meter makers honour it, then in this ISO 100 case Hg = 0.8 lux.seconds, or LogHg = -1.1.

So there is, in black & white negative, a fixed relation between ISO and Hm, a fixed relation between Hm and Hg (ten times, LogH 1, 3.3 EV) and therefore a fixed relation between ISO and Hm.

Re-reading the thread I understood that RobC simply and automatically wants to apply this simple relation to slide film. So he expects, or should I say he demands, an HmR (R for Reversal) which is, simply, LogH 1.0 less exposure than metered. Now, either I am not in good command of this language, or RobC is not very good at making his thought clear, or both.

Rob, forget the B&W relation between Hm and Hg when you deal with slides. They are not 3.3 EV apart. They are very close, because the speed point, in slides, is defined in a totally different way. What the relation is, for slides, between HmR and Hg, varies depending whom you ask to, but it is entirely different

A light meter manufacturer should, in theory, aim at creating an exposure of LogH -1.1 when the light meter is set at ISO 100.

The fact that Minolta states (but it would be better to read the original terms of the letter) that they aim to reach a density of LogH -1.0 seems to show that they, from the start, choose not to calibrate according to ISO standard. The speed point being fixed, it means they chose to place the Hg point at LogH 1.1 higher, not LogH 1.0 higher, than Hm.

What role play K and b in this? They enter into the calculation of the distance between Hm and Hg.
The fixed relation of "ten times" stems from the formula

Hg = [(b * K) / 0.8] * Hm

if we populate the equation with b = 0.65, the number which is used generally in ISO standards according to Conrad, and K = 12.5, we obtain 10.16, in logarithmic terms 1,007. So the "ten times rule" depends from these assumptions: b = 0.65 and K = 12.5.

If we now substitute K = 14.0, which is the declared value adopted by Minolta, and keep b = 0.65, the "distance factor" becomes 11,375 and LoghH = 1,056. That is exactly 1/6 EV (because it is 0.05 LogH) different than the normal distance factor of 10.16.

That means, in turn, that Minolta must have chosen a b very similar to 0.72 because that is the value which, coupled with K = 14 which Minolta declares, and with a target density of LogH = -1.0 (instead of -1.1) which Minolta declares, creates a "distance" between Hm and Hg of 3.6 EV (LogH 1.1), or 1/3 EV more than the normal 3.3 EV.

It's an overall exposure 1/3 EV different from what one might expect, half due to EV = 14 (instead of 12.5) and half due to b = 0.72 (instead of 0.65).

Maybe the original letter from Minolta to RobC also mentions the b factor (?)

 

[Chan Tran]

Well, the calibration light being a backlit screen explains why it is not a reflectance percentage that you are calibrating to. So it's a mistake of using the wrong words if you say "a light meter is calibrated to 12% reflectance". That's all I'm saying. It helps answer the question "What do you mean when you say a meter is not calibrated to a certain percent of gray?".

The reflected light is calibrated to a luminance and really has nothing to do with subject reflectance. The incident light meter is calibrated to an illuminance. But if the the incident meter and the reflected meter were to read the same there is a reflectance between the 2 of them.

 

[Bill Burk]

The reflected light is calibrated to a luminance and really has nothing to do with subject reflectance. The incident light meter is calibrated to an illuminance. But if the the incident meter and the reflected meter were to read the same there is a reflectance between the 2 of them.

Sure you can infer a relationship.

Many have concluded that 12% is close to the reflectance that will agree with the cardioid dome. Suppose you successfully locate a gray card that is 12%. We have learned from a meter manufacturer that you have to use a point source of light in a dark room. We've seen wiltw demonstrate you have to hold the card at a specific angle to the light because the light reflected back to the camera varies widely at different angles.

Now it wouldn't be hard to make a 12% gray card work, since it's thought by many to be the reflectance that matches the dome. To make an 18% gray card nearly agree with an incident meter reading, I think you need to use the flat disc.

 

[Diapositivo]

The reflected light is calibrated to a luminance and really has nothing to do with subject reflectance. The incident light meter is calibrated to an illuminance. But if the the incident meter and the reflected meter were to read the same there is a reflectance between the 2 of them.

Chan, what's the difference between a luminance and the light which is reflected by the subject? In photography we call "illuminance" the light hitting the subject, "luminance" the light that the lens sees - either because it is light reflected by the subject, or because it is light emitted by a light source. The light meter is calibrated to some light: moonlight or sunlight, it cannot distinguish it.

The relation is intrinsic in the idea of reflected light metering. It's in the y = kx equation which I illustrated many posts ago. Y = Luminance. k = percentage of reflectivity; X = Illuminance.
Both instruments want to know illuminance.
Incident light meters measure it directly;
Reflected light meters measure luminance y of the subject, assume a reflectance k, and derive illuminance x. They work "the other way round" than a reflected densitometre, which knows illuminance, knows luminance, and derives reflectivity.
What I mean is not that there is some sort of equivalence. What I mean is that it is logically and scientifically wrong to deny that a reflected light meter must be calibrated to a certain reflectance percentage.

If you put your incident light meter (which measures the level of illuminance) in front of your subject with four cats, Lucifer (black), Grisette (10% grey), Kodacat (18% grey) and Snowflake (80% white), in uniform light the light meter will give you a value that will make all luminances coming from the cats fall right on the film.
A reflected light spot meter is calibrated to a luminance, that is: to an illuminance reflected by the subject (after reflection it is called luminance). Assuming a certain reflectance, the instrument calculates illuminance falling on the subject.
If you now use a reflected spot light meter pointing it to each of the four cats, he will read four different luminances (which are the result of the same illuminance reflected by four different reflectances). If you expose for the luminance which is coming from Lucifer, the four cats will be overexposed. Why that? Because the light meter is not calibrated for Lucifer's reflectance!

If you expose for the luminance coming from Snowflake, all four cats will be underexposed. Why that? Because the instrument is not calibrated for Snowflake's reflectance!
If you expose for the luminance coming from Grisette, things will improve, but still no cigar.
If you expose for the luminance coming from Kodacat, all four cats will come out well exposed. And the exposure will be the same as the incident light meter. And why that? Because the light meter is calibrated for Kodacat's reflectivity and, when you throw a luminance coming from it, it will measure the illuminance "x" just as the incident light meter.

That fact that a flat, potentially flaring grey card is difficult to use to simulate a grey cat (you have to care about inclination, glare etc.) doesn't mean that there is not a shade of subject for which the reflected light meter is calibrated. There must be one. There is only one shade of grey cat which matches the reading of the incident light meter. That's the calibration shade of grey. Or, if you prefer, the calibration luminance given a calibration illuminance (in laboratory, you calibrate instrument by giving them the calibration luminance straight away because it's easier).

Denying that there is a calibration shade of tone amounts to denying that a reading on Lucifer, or a reading on Snowflake, gives a "wrong" exposure, IMHO.

 

[Diapositivo]

Regarding the 18% or 12% shade of grey: Gossen, Sekonic also sell grey cards, which are produced explicitly for photographic purposes. If 12%, or 16.5% had been the calibration grey, they would sell 12%, 16.5% grey cards. They sell 18% grey cards because their instruments are calibrated for 18% reflectance.

 

[RobC]

What you need to do is to use an incident meter to measure the light and then given that measurement you should be able to calculate what your spot meter reading from a grey card should give you so that you can angle the grey card until it gives you the spot meter reading you would expect and that way you will know your grey card is actually reflecting 18% towards the camera. So off you go and work out exactly how to do that and then you'll have a reliable reflectance percentage you can put in your images to test if they actually turn out that colour in your slides. Any problems, just ask benskin. Between the two of you, with a bit of luck and a fair wind, you might actually be able to take a meter reading this side of Christmas and get it right.

 

[Diapositivo]

Now it wouldn't be hard to make a 12% gray card work, since it's thought by many to be the reflectance that matches the dome. To make an 18% gray card nearly agree with an incident meter reading, I think you need to use the flat disc.

I agree on both counts.
It wouldn't be tough. And a grey card over a flat surface will give a coinciding reading.

I would like to spend a word or two in favour of this poor and mistreated grey card.
Two cases: flat subject, and tridimensional subject.

Flat subject.
You have an exchequer, with black and white boxes. Let's say 4% and 90% reflectance.
The average reflected light meter will give a reading that is around 1 EV below what it would give if the boxes were all white (it will see a 47% reflectance). Being calibrated for 18%, it will place the exposure at 1.2 EV or so above middle grey. That would make the white boxes light grey, and the subject would be underxposed. We want to place white on white (and black on black).
We could use a spot light meter, take the two readings, and make an average of those. (with slides, I would meter white, and place them 2.2 EV above metering).

We can use an incident light meter (in this case a flat disc would probably work better because the subject has no side in shade).
Or we can place an 18% grey card on the exchequer and meausure that. Our lightmeter, being calibrated for 18%, will calculate exactly the same exposure than the incident light meter would do. That will make the white fall on white (like all incident light metering).

Tridimensional subject
Even when a subject is lighted with uniform light, tridimensionality causes - generally speaking - a part of the subject receives less light that some other parts. (yes we can throw light at the subject so that it is all uniformly lit, but with most subjects we wouldn't do that, because it's actually the gradient of light along the surface that transmits the idea of volume).
If we imagine a sphere as being our subject, the light will make a gradient of shade along the sphere. If we imagine a cube, of which we see three faces, some faces will be brighter than some other.
If using an incident light metering, we would use a lumisphere. A lumisphere collects light from a wide angle, and its part in the shade will "average" with its part in full light and it will nicely give us a reading which somehow allows some more exposure for the part of the subject which is reeiving less light. The cheek in light will be "averaged" with the cheek which is in the shadow of the nose, for instance, if the subject is in front and the light is at one side. That works very well unless the subject is lit in a harsh way and the shaded part really is way darker than the lighted one (in that case, with slides, you have to do some thinking).

Now let's say we want to obtain the same reading with a flat 18% grey card. Not easy! It's not the 18% quality which is the problem, is the "flat" part which is the problem. So how can we try to take into account the lesser reflectance coming from the cheek which is in the shade of the nose, or of the temple which receives less light, with a flat surface?

The instructions tell us to incline the flat card 1/3 toward light source along both axes, vertical and horizontal. Easier said than done.
But why this? Only in order to try to "average" the light for the two different temples, IMHO. (Proponents of the 12% calibration theory say this is in order to make the card reflect ~12%, but then, why not to buy a 12% card and using it much more simply?).

Considering this kind of use, the two measures (incident and reflected on 18% grey card) will never really match well.

If using, insted, the grey card for what it does well, i.e. flat subjects, the two measures coincide well, as shown by wiltw in this thread. And a 12% card would not make the job.

 

[Diapositivo]

RobC I asked you to paste here the question you asked to Minolta and their answer. You gave me a very impolite answer (even allowing for your normal politeness standard, which is wanting) and you now go on with some more insignificant remarks.
You could try to participate in this conversation instead of disrupting it.

 

[RobC]

Oh, I do apologise for making a post in the topic I started, telling you how to verify if your grey card is reflecting 18% or not. I knew i shouldn't have bothered. Wake me up after Christmas.