That old 18% - a blast from the past

In connection with my recent work in the area of incident light exposure metering, I have looked at the manuals for many exposure meters, and this is a source of never-ending amazement.

For example, the manual for the Minolta Auto Meter IV F (ca 1991) contains this passage:


Wow! All reflected-light meters! I'm already a bit itchy.

Now first, what exactly might that be attempting to say? This statement was made in the film era, and it might be reasonable to think that it means, for positive (reversal) film, that for the metered object, the exposure result would be a density that corresponded to 18% transmission of light. Hold that thought.

Now if we consider the calibration of an exposure meter as defined by the ISO standard, the parameter K describes that calibration. But there is no value of K stated for the Minolta Auto Meter IV F in its manual.

However, in the manual for its predecessor, the Auto Meter III F, a value of K is stated in the manual: 14. (And we have often heard that the Minolta reflected light exposure meters do usually have a K of 14.) So we will assume that value of K for the Auto Meter IV F.

Now, if we assume metering with K=14, and make reasonable assumptions about other factors (such as the lens transmission fraction), and assume that the exposure index of the meter is set to the ISO speed of the film, and go through all the algebra, we find that the object upon which metering was done will end up with a density on the film that corresponds to about 12.5% transmission of light. Not 18%.

Many of you may recall an almost identical situation with regard to a digital camera.

So where did the author of that passage get that idea? Probably because he (like so many others) had heard at lot that, "exposure meters are calibrated to 18% reflectance". That statement says almost nothing. In the Minolta manual, it has been recast so it almost says something. But sadly that something is not correct.

There is a lot of that going around. And has been for a while.

Best regards,

Doug

Hi Doug,

I have always been led to believe that the 18% gray was established because of print industry but that actual light-meters can range from 10-18% and most between 12%-13%. The 18% gray card should be held at a 45% angle to the camera which would make its reflectance in the 12-13% range.

I have a gray card but do not use it for metering. I only use it for white balance where the % is not an issue.

Hi Doug,

Here is some more info about the origins of that persistent 18% value.

Funnily, if one uses a Lab or Luv value (in Bruce Lindbloom's Color Calculator) of [50,0,0], supposedly somewhat halfway on the L scale, that produces an XYZ coordinate with Y (luminance) at 18.4187% ...

So we can all be somewhat right, right?

I guess I'll just follow what my camera actually records in the Raw file, as can be revealed with RawDigger, or it's little brother FastRawViewer (currently still a free Beta version). One can the calibrate (if need be by selecting a different ISO) one's handheld meter to give e.g. a similar incident light exposure suggestion as would be required for optimal Raw exposure.

Cheers,
Bart

Hi, Bart,

Bart_van_der_Wolf said:

Here is some more info about the origins of that persistent 18% value.

Funnily, if one uses a Lab or Luv value (in Bruce Lindbloom's Color Calculator) of [50,0,0], supposedly somewhat halfway on the L scale, that produces an XYZ coordinate with Y (luminance) at 18.4187% ...

So we can all be somewhat right, right?

Click to expand...

My problem is not with the notion that a relative luminance of about 18% can reasonably be said to be "mid gray". Your example from the L*ab color space is a good demonstration.

A similar position can be derived from the Gray Gamma 2.2 color space, in which a value of the coordinate of "50%" corresponds very closely to a relative luminance of about 18%.

So I'm certainly game for considering a relative luminance of about 18% as "mid gray".

My objection to the statement in the Minolta manual that a metered exposure based on a meter with K=14 is expected to produce on, for example, positive (reversal) film a density whose "relative luminance" is about 18%. It is not anywhere close to that (and is not intended to be under the tortured underlying "philosophy" of the ISO standard for exposure metering).

That discrepancy does not at all involve the notion that a relative luminance of 18% constitutes "mid gray".

I guess I'll just follow what my camera actually records in the Raw file, as can be revealed with RawDigger, or it's little brother [URL="http://www.fastrawviewer.com/"]FastRawViewer (currently still a free Beta version).

Click to expand...

Thanks for those nice references. I will try and lay in the free one!

Best regards,

Doug

Hi, Bart,

I said:

Doug Kerr said:

A similar position can be derived from the Gray Gamma 2.2 color space, in which a value of the coordinate of "50%" corresponds very closely to a relative luminance of about 18%.

Click to expand...

In fact, in the Gray Gamma 2.2 color space, a coordinate value (the single coordinate there is "K", presumably mnemonic for "black", borrowed from CMYK, since it is in effect the "blackness" of the color) of 50% (on a scale of 0-100%, of course, with 100% being full black) corresponds to a relative luminance of 21.6%.

A relative luminance of 18% comes from a K value of about 54.

This still all "justifies" considering a relative luminance of about 18% to be "mid gray".

Bet regards,

Doug

I am reminded of the "Chuck Westfall kitchen-table Canon dSLR exposure meter calibration test" of many years ago.

Here we take a metered exposure of a uniform-luminance target and examine the exposure result.

It is expected to be K=55 in the Gray Gamma 2.2 color space (a peculiar and confusing choice).

That corresponds to a relative luminance of about 17.3%.

If we consider:

• For the inbuilt exposure metering system, a K of 12.5 per the ISO standard.

• The exposure index of the metering system set to the ISO SOS of the sensor (not understood well at the time - actually I think not yet officially defined by ISO, just followed by Canon)

then we might expect an exposure result of about 17.7%.

How about that!

Recall that the Minolta manual did not contemplate the use of the ISO SOS as the exposure index for the meter but rather the ISO speed, 1/2 stop larger.

Which is why, as I said at the outset, we might expect an exposure metered with the Minolta meter to have what, on reversal film, we could reasonably consider a relative luminance of about 12.5%.

The difference comes from:

• The Minolta meter has K=14, whereas the Canon exposure metering systems are presumed to have K=12.5.

• The Minolta meter is given an exposure index that is the ISO speed of the film, whereas the Canon meter is given an exposure index that is the ISO SOS of the sensor.

• The Canon meter is through the lens, and thus automatically compensates for the expected transmission fraction of the lens. The Minolta meter cannot do that.

Best regards,

Doug

Hi, Bart,

Fast Raw Viewer is neat!

Thanks.

Best regards,

Doug

It is a warning sign in a philosophical argument or a technical one that when inconsistencies, contradictions, or ad hoc fixes arise the original premises may be flawed. The Zone System was devised to describe and control the behavior of black and white photography and its extension to digital picture-making is where the flaw lies.

Without graphs or numbers this is how the original concept travels. A piece of film receiving a Zone V Exposure acquires, after "standard" development, a Negative Density V. This Negative Density V when projected upon photographic paper in a "standard" way causes the paper to yield Print Value V. The constraint that guides "standard" in this work flow is that Zone V exposure ends up as Print Value V and simultaneously Zone I exposure ends up as Print Value I and Zone IX exposure ends up as Print Value IX.

The heart of the Zone System is that it provides for non-standard development regimes where Zone exposures can be unlinked from Print Values. This enables a variety of subject luminances to be assigned an arbitrary range of grey values on photographic paper according to the artistic proclivities of the photographer.

Digital picture-making and photography are very different animals even at the very front end of their respective work-flows. The digital sensor is a transducer that provides output information that is ultimately processed by various "engines" into a picture. Film is not a transducer and provides no output. Rather it becomes the picture itself.

And what of that pesky 18% grey? After a few thousand journeys through Zone V exposure, Negative Density V, and Print Value V I can say that under a reflection densitometer Print Value V, if not exactly 18% grey, sure looks reminiscent of it.

Just leave consideration for the use of the Pentax or similar Spotmeter to allow one to practice placing a particular element in a favorite zone for LF photography!

Asher

Hi, Asher,

Asher Kelman said:

Just leave consideration for the use of the Pentax or similar Spotmeter to allow one to practice placing a particular element in a favorite zone for LF photography!

Click to expand...

So suppose that in the scene the central feature is a gray door that you just know needs to go into zone V.

You take a spot reading on it.

From the meter reading, how do you decide what exposure to use?

Let's assume you are using B&W negative film.

Do you set the ISO speed of the film into the meter for the exposure index and use the photographic exposure (shutter sped and aperture) recommended by the meter for the shot? Or something else.

What density do you expect the image of that door to have on the film (if given standard development)?

Thanks.

Best regards,

Doug

Hi, Asher,

Is your typical practice when "placing a particular element in a favorite zone" like this:

• You choose a particular element as the "marker", and decide in what zone you want it to be rendered.

• [Optional] If we are speaking of B&W negative film, perhaps from a number of spot readings on the scene, you conclude as to what development to use, and what exposure index that implies.

• You set the exposure index of the meter to that value, or if you are not contemplating a custom development (perhaps such is not applicable to the type of film being used) set the exposure index to the rated ISO speed of the film.

• You take a spot reading on the "marker" element.

• You take the photographic exposure recommendation of the meter and adjust it up or down by the number of stops that the target zone for the element differs from Zone V. For example, if your target "placement" for this element is Zone III, you will shoot at an exposure two stops less than recommended by the meter.

Thus, we do not actually have in mind a certain negative density for that element. Rather, we consider that the photometric exposure that would be accorded an element by shooting at the photographic exposure recommended by the meter, metering on the element, to be the desired "placement" for a zone V element, and work from there.

This of course would be the infamous "mid gray" placement. But we do not need to know exactly how that would be quantified in terms of film density. Or care. The meter and film people have taken care of that for us.

Does this make sense?

Thanks.

Best regards,

Doug

Hi, Maris,

Thanks for your valuable observations.

Maris Rusis said:

It is a warning sign in a philosophical argument or a technical one that when inconsistencies, contradictions, or ad hoc fixes arise the original premises may be flawed. The Zone System was devised to describe and control the behavior of black and white photography and its extension to digital picture-making is where the flaw lies.

Without graphs or numbers this is how the original concept travels. A piece of film receiving a Zone V Exposure acquires, after "standard" development, a Negative Density V. This Negative Density V when projected upon photographic paper in a "standard" way causes the paper to yield Print Value V. The constraint that guides "standard" in this work flow is that Zone V exposure ends up as Print Value V and simultaneously Zone I exposure ends up as Print Value I and Zone IX exposure ends up as Print Value IX.

Click to expand...

Nicely said.

The heart of the Zone System is that it provides for non-standard development regimes where Zone exposures can be unlinked from Print Values. This enables a variety of subject luminances to be assigned an arbitrary range of grey values on photographic paper according to the artistic proclivities of the photographer.

Digital picture-making and photography are very different animals even at the very front end of their respective work-flows. The digital sensor is a transducer that provides output information that is ultimately processed by various "engines" into a picture. Film is not a transducer and provides no output. Rather it becomes the picture itself.

Click to expand...

An interesting, and important, distinction.

And what of that pesky 18% grey? After a few thousand journeys through Zone V exposure, Negative Density V, and Print Value V I can say that under a reflection densitometer Print Value V, if not exactly 18% grey, sure looks reminiscent of it.

Click to expand...

Makes sense to me.

Thanks.

Best regards,

Doug