18 percent grey & exposure theory needs clarifying :-)

<p>Your best bet when photographing a difficult subject is to take that 18% gray card and place be between the camera and subject and meter off that. This is an alternate way of getting an incident reading.<br>

The other, and in my mind much more preferable, way is to use a spot meter and use the Ansel Adam's Zone System but that is outside the scope of this question.</p>

I wanted to update this thread because it's one of those evergreens that gets dredged up from time to time.

While 18 percent gray cards are still invaluable if you know how to use them, for most people it's now a non-issue because in-camera meters are much smarter than they were in the 1960s.

Here's an explanation for a typical metering system, this is Nikon's but they're all similar:

For difficult lighting conditions, the best tool is still an incident meter, but a gray card works well too as long as you follow the directions which call for you to angle the card and use the exposure compensation recommendations that come with the Kodak-brand cards.

The history of how the 18% reflectance value was arrived at is a bit hazy, but the most believable story goes thus:

Kodak had a virtual monopoly over amateur snapshooting equipment at one time. The famous phrase "You push the button, and we'll do the rest." was totally true, with Kodak producing camera, film and taking care of the processing and printing. Therefore Kodak's labs had access to thousands, if not millions of amateur negatives. The story goes that, in the interest of improving results, Kodak's scientists analysed thousands of these amateur snapshots to empirically come up with an 'average' scene reflectivity, which turned out to be 18%.

This was also not far from the findings of other companies, like the then infant photoelectric metering company Sangamo Weston. The scale of early Weston meters had 'O' (Overexposure), 'U' (Underexposure), and 'C' (Correct) markings. These set limits of 3 stops between 'O' and 'C', and 4 stops between 'C' and 'U'. So if we take the overexposure point to be pure white at around half-a-stop more than 100% Lambertian reflectance*, then the 'Correct' exposure marker aligns almost perfectly with a Lambertian reflectance of 18%.

This empirically arrived at figure has survived to be embodied in the calibration constant of most exposure meters sold today, and in the centre-weighted or average setting of TTL meters in both digital and film cameras.

Meters that use an incident dome or cone are also calibrated to give an exposure as if the incident light being measured were being reflected from a Lambertian surface of 18% reflectance.

*A Lambertian surface is a theoretical matt surface that reflects equally in all directions through a solid angle of 180 degrees. It can be almost perfectly simulated practically by a specially prepared PTFE material (commercially called Spectralon). Many natural surfaces like chalk or fresh powder snow also approximate a Lambertian surface, as do common man-made surfaces like full-matt paint or common copier paper.

The above examples of chalk, white Spectralon, whitewash paint and copier paper also come close to having 100% reflectance, and can take the place of an 18% card simply by adding 2.5 stops to the exposure indicated by metering such a surface (100% reflectance minus 2.5 stops = 18%).

It's worth noting that the theoretical and measured reflectance values of Ansel Adams' Zones V and VIII are not a full 3 stops apart. According to Adams' written descriptions of those surfaces they respectively have reflectances of 18% (stated as such) and approximately 100% ("white with texture"); placing them 2.5 stops apart, and not 3.

1. I think in terms of "Zones" whenever i meter. FWIW, i only shoot film with Cameras/Meters circa 1970-1980.

I miss stuff once in awhile, but for the most part, the metering side of The Zone System has served me well.

2. However, i have always been confused by Zone-5 and 18% gray.

You will have to pardon my math vocabulary (lack of rather) but is it Algebraic, or Logarithmic.?

If Zone-5 is f/5.6 at 1/500. Then Zone-6 is f/5.6 at 1/250..... twice as much light. Or, if you go the other way it is half as much light.

That is how we are taught to think of stops. They either Halve or Double the light.

So, using that math

Zone-5 = 18%

Zone-6 = 36%

Zone-7 = 72%

....... we run out of math.

The LRV of the Zones must be based on something besides simple math.?

I once tried testing my meter. I found that the reading from the incident light was different about 1/2 to one stop to the reflective reading using a gray card. I once read there was a reason for it. But I don't remember. Is anyone familiar with my "problem"?

You need to make sure your dome is "Down" so it is getting reflective value (if that is even possible) from just the gray card.Kind of a tricky maneuver. At that point you would be better off to use the reflective meter.

Now, back to my math problem.........

I never liked dealing in percentages; they can be misleading.

An item sells for $10 and years later is worth $100. It is now worth ten times as much - a 1000% increase. Most people would agree with that. But no, that is a 900% increase.

An item sells for $10 and is later valued at $20 - a 100% increase not a 200% increase.

A stock sells for $10 a share and drops to $5 a share, a 50% decrease. When it goes back up to $10 a share that is a 100% increase. You have made 50% profit by doing nothing.

Maybe hanging the 18% value into a real number would clarify matters but I don't know how.

Wait 'till you find out about the ANSI "k-factor" that may or may not be used when calibrating light meters.

All modern meters are calibrated with a K factor for reflected light meters. For incident light meter they use the C factor.

Most of you will have little knowledge of semsitometry or densitometry. Semsitometry is the science of making precise exposures on film to study its characteristics and densitometry is the science of measuring the blackening of photo films. These sciences had its birth with the publications of Hurter and Driffield in1890. They exposed film in 1 stop (2X) increments and grafted the resulting blackening. If ordinary numbers were used, the graph is hopelessly too long. This is because, by the 14th exposure the light energy applied is 8,192 times the first.

The graph called the H&D curve is best displayed using logarithmic notation. Such math graphs elegantly. The use of logarithmic notation remains is use today. It was more common before the calculator or computer. This is the language of the slide rule.

Anyway, light exposes film; the film is developed and blackening results. How to label this backing? A beam of light of known value is played on the film. The amount of light that emerges is measured.

T = Transmission The amount of light that traverses the film divided by amount of light played on the film.

If 50 units makes it through and 100 units played on the film, T = 50 ÷ 100 = 0.5. We always express this value as a percent. Percent means per 100 so we multiply by 100 thus 0.5 X 100 = 50%

O = Opacity The amount of light that plays on the film divided by the amount that gets through. In this case 100 ÷ 50 = 2

D = Density is O expressed in logarithmic notation base 10. Thus 2 expressed as a log base 10 is 10^0.3. (ten elevated to the 0.3 power. In is common practice to omit the 10 base thus we say the number 2 represented in log notation is 0.3

Tough stuff!

OK – what about 18% gray? Look at step 7 of this gray scale of a photograph print. The density is 0.75. The reflectance is 0.178. This value expressed as a percent is 0.178 X 100 = 18%. In other words step 7 is the 18% gray target we are talking about.

Why is this important? If we correctly expose and process film and target an 17% object. On the photo film, the image of that object will read Density 0.75. A print made from that film will display an image of that object. The reflected Density of that object will read Density 0.75.

The value 0.75 Density is Transmission or reflection 18%. This is the center of the gray scale of pictorial photo film. This is the only tone that has the same Density on film as on print. This is the pivot point of the photographic gray scale.

[ATTACH=full]1314194[/ATTACH] Most of you will have little knowledge of semsitometry or densitometry. Semsitometry is the science of making precise exposures on film to study its characteristics and densitometry is the science of measuring the blackening of photo films. These sciences had its birth with the publications of Hurter and Driffield in1890. They exposed film in 1 stop (2X) increments and grafted the resulting blackening. If ordinary numbers were used, the graph is hopelessly too long. This is because, by the 14th exposure the light energy applied is 8,192 times the first.

 

The graph called the H&D curve is best displayed using logarithmic notation. Such math graphs elegantly. The use of logarithmic notation remains is use today. It was more common before the calculator or computer. This is the language of the slide rule.

 

Anyway, light exposes film; the film is developed and blackening results. How to label this backing? A beam of light of known value is played on the film. The amount of light that emerges is measured.

 

T = Transmission The amount of light that traverses the film divided by amount of light played on the film.

If 50 units makes it through and 100 units played on the film, T = 50 ÷ 100 = 0.5. We always express this value as a percent. Percent means per 100 so we multiply by 100 thus 0.5 X 100 = 50%

 

O = Opacity The amount of light that plays on the film divided by the amount that gets through. In this case 100 ÷ 50 = 2

 

D = Density is O expressed in logarithmic notation base 10. Thus 2 expressed as a log base 10 is 10^0.3. (ten elevated to the 0.3 power. In is common practice to omit the 10 base thus we say the number 2 represented in log notation is 0.3

 

Tough stuff!

OK – what about 18% gray? Look at step 7 of this gray scale of a photograph print. The density is 0.75. The reflectance is 0.178. This value expressed as a percent is 0.178 X 100 = 18%. In other words step 7 is the 18% gray target we are talking about.

 

Why is this important? If we correctly expose and process film and target an 17% object. On the photo film, the image of that object will read Density 0.75. A print made from that film will display an image of that object. The reflected Density of that object will read Density 0.75.

 

The value 0.75 Density is Transmission or reflection 18%. This is the center of the gray scale of pictorial photo film. This is the only tone that has the same Density on film as on print. This is the pivot point of the photographic gray scale.

The middle density of a film is not 0.75. It's higher than that. For color slide film it's about 1.0 and for color negative it's about .8 to 1.10 for the red layer. For B&W film it's more than 1.0.

....... we run out of math.

The LRV of the Zones must be based on something besides simple math.?

No. It just places zone VIII at a highly unlikely 144% reflectance.

I very much think the 'problem' was with Ansel's arithmetic. Either that or zone V needs to be relocated at 12.5% reflectance. Despite Adams' insistence to the contrary.

Then again, that places zone III at a very dark 3.125% reflectance, and zone II at a reflectance rarely achieved by natural materials and around the reflectivity of black velvet.

Unless Adam's - a trained concert pianist - was subconsciously slipping semitones into his obvious attempt to create a parallel between the (totally illogical IMO) chromatic musical scale and a visual one. Maybe Ansel had synaesthesia?

Since the zone system is just an extremely over-complicated version of the old advice to "expose for the shadows and develop for the highlights" it hardly matters much. Especially since hardly anyone develops individual sheets of film these days.

for color negative it's about .8 to 1.10 for the red layer. For B&W film it's more than 1.0.

Total nonsense!

The red curve of Portra 160 tops out at a density range of around 1.5, and the more contrasty Ektar 100 barely reaches a density range of 1.7. In addition, no correctly exposed and developed B&W film should exceed a density of about 1.8 in the printable highlights. Yes, you can develop B&W film to exceed a density of 2.0, but it won't print or scan easily. A normal contrast range subject should be developed to hold its textured highlights under 1.8 (+B+F). Any higher than that and your developing technique needs attention.

See Ilford's technical datasheet for FP4+. According to Ilford, when developed in Ilfotec HC its density flattens off well below 2.0D.

Colour reversal film is a special case. It's really a positive print on transparent material. The negative having been dissolved away in the bleach and fixing baths.

a

The middle density of a film is not 0.75. It's higher than that. For color slide film it's about 1.0 and for color negative it's about .8 to 1.10 for the red layer. For B&W film it's more than 1.0.

For B&W film the pivot point is 0.75 plus bas fog which is usually 0.1. For color negatives we only pay attention to the red emulsion layer. We set this value at about 1.05 red density. The objective is to avoid the use of cyan filters when printing on an enlarger. The 18% gray tone on the print measures 0.75 for all three emulsions however most gravitate to 0.80 to give the print a little more density. For many years I managed an operation that made test film and prints for the world-wide photofinishing industry. In other words I am no beginner.

I wanted to update this thread because it's one of those evergreens that gets dredged up from time to time. . . For difficult lighting conditions, the best tool is still an incident meter,

For difficult lighting conditions, in 2019, the "best tool" might be an incident light meter:

however the best tool might also be the TTL Meter in the Digital camera, combined with using an exposure bracket and the subsequent interrogation and analysis of the histogram and 'blinkies'.

WW

Total nonsense!

The red curve of Portra 160 tops out at a density range of around 1.5, and the more contrasty Ektar 100 barely reaches a density range of 1.7. In addition, no correctly exposed and developed B&W film should exceed a density of about 1.8 in the printable highlights. Yes, you can develop B&W film to exceed a density of 2.0, but it won't print or scan easily. A normal contrast range subject should be developed to hold its textured highlights under 1.8 (+B+F). Any higher than that and your developing technique needs attention.

 

See Ilford's technical datasheet for FP4+. According to Ilford, when developed in Ilfotec HC its density flattens off well below 2.0D.

 

Colour reversal film is a special case. It's really a positive print on transparent material. The negative having been dissolved away in the bleach and fixing baths.

1.5 minus 1 is 0.5 and that's about 3.3 stops in term of print because the paper has much higher contrast.

You need to make sure your dome is "Down" so it is getting reflective value (if that is even possible) from just the gray card.Kind of a tricky maneuver. At that point you would be better off to use the reflective meter.

Now, back to my math problem.........

Denny: I guess I wasn't clear. I'm using the same meter. It has a dome for incident readings. Then you swap in a 10 degree viewfinder to take reflective readings. So I'm looking at the gray card exclusively. Here's my question again. Maybe someone has an answer.

"I once tried testing my meter. I found that the reading from the incident light was different about 1/2 to one stop to the reflective reading using a gray card. I once read there was a reason for it. But I don't remember. Is anyone familiar with my "problem"?"

Denny: I guess I wasn't clear. I'm using the same meter. It has a dome for incident readings. Then you swap in a 10 degree viewfinder to take reflective readings. So I'm looking at the gray card exclusively. Here's my question again. Maybe someone has an answer.

"I once tried testing my meter. I found that the reading from the incident light was different about 1/2 to one stop to the reflective reading using a gray card. I once read there was a reason for it. But I don't remember. Is anyone familiar with my "problem"?"

I understand.

Most meters with a dome also have a way to either recess or shield the dome so it is not taking a 180 degree reading.

If your meter cannot do that, then your dome is seeing a lot more area than your reflective is.

My Sekonic 308 does not recess, but it has an attachment for the dome that blocks out all of that "side" light.

1.5 minus 1 is 0.5 and that's about 3.3 stops in term of print because the paper has much higher contrast.

Whaaaa?

Where did 'minus 1' come from?

And where did 1.5 come from?

The 'contrast' of a print rarely exceeds 100:1 or a reflective density range of 2.1D (at most). But nobody was talking about print contrast.

You can't just add or subtract densities at random and relate them directly to a print density. The gamma curve of Bromide paper is 'S' shaped, and less steep at the bottom and top.

Here's my question again. Maybe someone has an answer.

"I once tried testing my meter. I found that the reading from the incident light was different about 1/2 to one stop to the reflective reading using a gray card.

Grey cards can fade or get dirty over time, and a half-stop discrepancy only requires a few percent change of reflectance (18% to 13%, or 18% to 25%).

This has implications for the required manufacturing tolerance of a grey card too. I suspect that a cheaply printed grey card might easily miss the mark by a considerable margin.

FWIW, I once tried to print a 50% reflectance halftone dot pattern. It sounds easy; you just make a pattern of small squares with half of them black and the other half white paper - right? Wrong! It just didn't work out like that. So producing an exact18% reflectance surface isn't trivial.

Personally, I prefer to use a disposable white card - AKA a sheet of copier paper. This can't fade lighter and even if it isn't exactly 100% reflectance, a small percentage error at close to 100% is less drastic than the same percentage error at 18%. OK, you have to increase the white card reading by 2.5 stops, but it's hardly higher maths. The 'white card' also effectively increases the sensitivity of your reflective meter by 2.5 stops as a bonus. And you throw it away or recycle it after use, so it has little chance to get dirty.

Joe, I seem to recall there was an actual difference in the measuring that I read about that gave a one stop difference.

Ok I found some discussion on this difference in readings. CAn anyone clarify this for me?

Question about 12% vs 18% middle grey: Beginners Questions Forum: Digital Photography Review

Here's another that I also copy for simple reference.

18% vs 13% grey card: which should have a centered histogram centered?: Studio and Lighting Technique Forum: Digital Photography Review

18% gray is the reference used by printers. Light meters and the meters in cameras are calibrated to read 12% gray (not 13% gray). There is about 1/2 stop difference between the two.

http://www.bythom.com/graycards.htm

It is thought that Ansell Adams insisted on 18% gray, which is why it became the standard for film photography. That was fine in the days of film since B&W or color negative film had a large exposure latitude. It was also a simple matter to adjust the developing to make metering an 18% gray card work as a means of exposure control.

Digital camera detectors are more like slide film in that they have very little exposure latitude so the exposure has to be dead on if you want the best results.

What this means is if you take a reading of a general outdoor scene; i.e. fields, trees, mountains, sky, and clouds; you will get a reading based upon 12% gray. If you then read a gray card in the same light it should read that you need 1/2 more exposure.

It is OK to meter a gray card if you want, just remember to DECREASE the exposure by 1/3-1/2 stop from the meter reading and you will be getting the correct exposure with your digital camera.

The history of how the 18% reflectance value was arrived at is a bit hazy, but the most believable story goes thus:

Kodak had a virtual monopoly over amateur snapshooting equipment at one time. The famous phrase "You push the button, and we'll do the rest." was totally true, with Kodak producing camera, film and taking care of the processing and printing. Therefore Kodak's labs had access to thousands, if not millions of amateur negatives. The story goes that, in the interest of improving results, Kodak's scientists analysed thousands of these amateur snapshots to empirically come up with an 'average' scene reflectivity, which turned out to be 18%.

 

(snip)

Averaging meters, reading over most of a scene, often enough work just fine.

That is, more often than you might expect. Usually, it is only if a large amount

of background is very light or very dark, that you have to adjust.

As well as I know, machine printed negatives are normally adjusted based

on an average over the negative.

It is also usual for printing from color negatives to generate a color balance that averages

to gray. I suspect that also came from Kodak measuring a lot of negatives.

This can cause problems when the background has much of one color, but even

then it works better than you might expect.

"As well as I know, machine printed negatives are normally adjusted based on an average over the negative."

Which means you shouldn't worry about the correct exposure when taking the shot for if you send the film out to be printed, the printer analyzer will decide what exposure it will print at no matter how careful you were with the taking exposure.

I once took some photos of some surfers. I had the camera set on manual and based the exposure on an incident light meter and checked with the "sunny 16" rule. It was a sunny day and the light condition didn't change during the time I took the shots. I sent the film out and when it came back I checked the negatives. (The first thing that I do) to see how well I did. All the negative exposures looked good with equal density for them all. The prints were much different with some light, some dark and some okay. If there was a lot of white water in the scene the printer analyzer would think, "Way to light I had better darken down," and I would get a dark print. If there was too much dark water, the printer analyzer would think it was too dark and lighten up the print too much. Being super critical at the time of the film exposure would be a waste of time it the film was to be sent out for prints. I sometimes would take a photo of a gray card for the first shot and tell the tech to set the printer for that for all the prints.

Joe, I seem to recall there was an actual difference in the measuring that I read about that gave a one stop difference.

Never heard of such a discrepency.

However, WRT 18% versus 12.5%:

My experiments with Nikon DSLRs, plus a limited number of other digital cameras, leads me to believe that they are indeed geared toward an 18% average reflectance exposure.

Try this. Put the camera on CW average or spot metering and point it at an evenly lit blank surface (in focus, because focus affects exposure). Take a shot and note the histogram line - it should be a quite narrow pile of pixels somewhere toward the centre of the graph.

Now add 2.5 stops exposure compensation, or manually expose 2.5 stops more than before. You should find that the pile of pixels has moved to nudge the RH edge of the graph - just short of inducing 'blinkies'.

Add another half stop and it should push the pixels over the edge and into overexposure territory.

QED. The margin between just 'pure white' - i.e. 100% Lambertian reflectance and the camera's exposure norm is 2.5 stops, and 18% reflectance is exactly 2.5 stops lower than 100%.

If the camera tolerated a 3 stop increase without overexposure, then that would indicate an averaged exposure of 12.5%.

The 13% figure (actually 12.5% rounded. 12.5% being exactly 3 stops less than 100%) was used on early Weston meters, with the 'O' marking on the meter scale being 3 stops higher than the 'C' mark.

I suspect that some people with an addiction to neatness and nice round numbers would prefer it if the universe conformed and handed us exactly 3 nice whole stops below the 100% Lambertian reflectance reference as 'average'. But it doesn't seem to work that way.