Underexposure with incident metering

[Ranchu]

Because colour negative film can take a whole lot of overexposure, so exposing for shadow detail works brilliantly, still allowing retention of detail in the highlights.

If I'm reading your note right, you used a center weighted averaging meter to meter off the middle of your picture.

 

[Richard G]

If I'm reading your note right, you used a center weighted averaging meter to meter off the middle of your picture.

Well, I'm not sure. And I wouldn't want to incriminate myself. I took that photo five years ago. It was a 35mm lens. I can't actually remember what I did precisely for the metering, other than that I wanted to set the exposure for the depths of the room which was pretty dark. I don't know whether I just added a stop to the metered reading or pointed the camera down or right for the darkest part of the room or what. The M6 is completely manual I'm sure you know. I was using the M2 on that trip too, with a Gossen meter, and so manual settings were routine for me at that time. I had just used the M6 in the light at the front of the cafe and might well have had a hand held incident reading there. If that was the case I might have just added a stop or even two to the previous setting. I just can't be sure. None of this changes my essential point: choosing to expose for the shadows gives tolerable highlight detail in the overexposed parts of the frame with colour negative film. I was impressed by the detail retained in the nearest overhead light on the left.

 

[Ranchu]

It's a fine picture, no doubt. For some reason I'm really liking the wall paper..

 

[Richard G]

Thanks Ranchu. It was all class that place. I suspect it was destroyed in the big Christchurch earthquake.

 

[Sparrow]

Because colour negative film can take a whole lot of overexposure, so exposing for shadow detail works brilliantly, still allowing retention of detail in the highlights.


Fuji Superior 400(?): Exposed with an M6 reflective reading from the back of a very dark cafe late on an overcast afternoon.


Le Petit Croix by Richard GM2, on Flickr

... well again I have my own, lets call it empiric method for places like that, that reasons 'there isn't much light in here so I'm not going to overexpose anything .. so that's f2.8 at 1/30 then' 🙂

... btw, nice photo Richard

 

[Roger Hicks]

Because colour negative film can take a whole lot of overexposure, so exposing for shadow detail works brilliantly, still allowing retention of detail in the highlights. . . .

Exactly. And that's even before you start dodging and burning.

Cheers,

R.

 

[Roger Hicks]

OOps! The meter will respond to the white paper as if it were an 18% gray card, thus giving a reading 2 and 1/2 stops more stopped down than it should be. Thus we must open up 2 and 1/2 stops, not stop down that much!

OK, with that said, I want to say I have never really been able to mathematically show that 18% is a true midpoint. My reasoning goes, one stop brighter than 18% would be 36%. So two stops would be 72%. We can only get to 100%, so there's no room for another doubling. But intuitively, the jump from 72% to 100% seems like another half-stop or so. OK, that accounts for our belief that there's 2 and 1/2 stops from 18% gray to full white. And that seems fine. Now let's go in the other direction: one stop darker, 9%. Two stops darker, 4 and 1/2%. Three stops darker, 2 and 1/4%. Four stops darker, 1 and 1/8%.

Question: How far should this sequence be carried on? We are already four stops below our chosen 18% midpoint, whereas there are only 2 and 1/2 stops the other way. What kind of a midpoint is that? I don't know; I am raising a question here, I don't know why 18% is a magic middle tone.

One other thing (Roger): (Psychologist here) "Psychophysics" refers to a method of quantifying the response of the nervous system to various stimuli. For instance, if there are 100 candles lit, how many candles must we blow out before an observer can notice the change in light level? A: one. The Just Noticeable Difference (JND) for illumination level is one part in 100. Or: the ear's response to sound levels is logarithmic. Three deciBels equates to a perceived doubling or halving of sound intensity. And the JND is one deciBel. Another example: ever notice how the eye doctor finds your prescription? She starts with a power weaker than you need, and ascends through the correct one, noticing when you say, "no, the last one was better." Then she starts higher than your true prescription, and descends, until you say, "no, the one before this one." That's called the 'Method of Ascending and Descending Series," developed by experimental psychologists. That's another application of psychophysics. It's not about "what looks right!" It's about quantifying subjective inner experience.

Para 1: Oops indeed. A common mistake. Frances calls it "photo dyslexia." Sorry! It was late...

Para 2: The response of the eye is not linear. Give people 50 cards from 2% reflectance (hard to get much lower) to 95% reflectance (hard to get much higher) and they will pick something around 18% reflectance as a mid tone. Otherwise it is, as you rather imply, an almost completely useless number. It doesn't even correspond to an average landscape reflectance near Rochester NY (12-14%).

Last para: We must be using the terms slightly differently here, because I'd suggest that the example you give of the optometrist is, absolutely and precisely, "what looks right". As is the Munsell mid-tone. As are the original "first excellent print" tests that gave rise to Kodak (later ASA) film speeds. The quantification of the latter was especially devious, involving the rate of change of the d/log E curve.

You are, as you say, quantifying subjective inner experience. But that "inner experience" is, in all the cases given above, "what looks right".

Cheers,

R.

 

[lrochfort]

Hi,

Nice to see you two back together...

Ever since the opening post I've been wondering why incident light readings for landscapes but no one else has said a thing.

Regards, David

Money mostly! I only have an incident meter and haven't been able to afford another meter until now.

If possible, I tend to take an incident reading from the subject's position back towards the camera, and then slide the cone back to meter the shadows up close. I then try to make as much of an informed decision as I can about how to proceed based on the difference between the readings and the shadow areas in the subject.

I think I was just having an off period, or my 4 year old was taking more of my attention during that time. He's completely fascinated by the light meter, so it often comes back to me with the dials all over the place 🙂

However, I went out this past weekend and was able to concentrate and be more methodical. Glad to say I was able to get several rolls that I'm pleased with that had good exposure in varying light levels and conditions.

Last week I found an absolute bargain of a Pentax Spotmeter V on eBay, so I'm very much looking forward to using that if only because I think using the Zone System will provide a framework within which I can build some structure and learn in a somewhat more deliberate manner.

Cheers,
Laurence.

 

[Roger Hicks]

A little thought experiment to demonstrate the exact equivaence of spot metering a grey card, and incident light metering.

In the middle of a landscape, or indeed any other large-ish subject, we have a sheet of cardboard with two holes in it, close together. One hole is blocked with a card that reflects exactly 18% of the light falling on it in a perfectly diffuse manner, e.g. an ideal grey card. The other is blocked by a sheet of translucent material that transmits exactly 18% of the light falling on it: again, perfectly diffusely, e.g. an ideal incident light receptor.

With the same meter, we read the front of the reflective card and the back of the translucent material. How are the readings going to differ?

There may be a flaw in the argument, but if there is, I can't see it.

Cheers,

R.

 

[Ranchu]

How are the readings going to differ?

The translucent one will be 2 1/2 stops more exposure because it's an artificial highlight.

 

[Roger Hicks]

The translucent one will be 2 1/2 stops more exposure because it's an artificial highlight.

Eh?

Are you making a joke? Or did you not read what I wrote? Or did you not understand what I wrote?

If the first, I apologize, but otherwise, I can't make any sense at all of what you wrote.

Cheers,

R.

 

[Ranchu]

That was a joke. Juan seems to me to be a person who would know what he's looking at, Roger.

 

[icebear]

... well again I have my own, lets call it empiric method for places like that, that reasons 'there isn't much light in here so I'm not going to overexpose anything .. so that's f2.8 at 1/30 then' 🙂

....

ROFL , that made me almost spill my coffee, great, love it ! 😀

 

[Rob-F]

Para 1: Oops indeed. A common mistake. Frances calls it "photo dyslexia." Sorry! It was late...

Para 2: The response of the eye is not linear.

Last para: We must be using the terms slightly differently here, because I'd suggest that the example you give of the optometrist is, absolutely and precisely, "what looks right".

You are, as you say, quantifying subjective inner experience. But that "inner experience" is, in all the cases given above, "what looks right".

Cheers,

R.

1. Yep, I knew you knew better than that, Roger!

2. We are using the terms differently, I think. In one sense, it is, "I like what I see." That one is subjective. Like for art critics, maybe. The other sense has to do with verifiable, repeatable, quantitative measurement. Like when the split images line up in the rangefinder. It not only looks right, it is right. In the case of the optometrist, it has to be right, or else many of us would have such bad eyeglasses, we wouldn't be able to find our cameras! (Speaking for myself, here.)

@Juan, et. al.: I don't see any way to argue with Roger's position that the only way to ensure a given shadow registers on film, is to measure that shadow. Not that allowing a safety margin won't work most of the time; but "when it absolutely positively has to be there . . ." then using judgment, experience, and doing "what usually works" (like rating the film slower than box speed, for example) might be too risky. And "Who cares what's in the shadows?" merely avoids the issue.

I think the various methods/approaches described here are probably all valid, as long as one get acceptable results. But some approaches are more scientifically defensible, while others are more casual! One way might be good enough for a street photographer; the other, good enough for NASA.

 

[Roger Hicks]

1. Yep, I knew you knew better than that, Roger!

2. We are using the terms differently, I think. In one sense, it is, "I like what I see." That one is subjective. Like for art critics, maybe. The other sense has to do with verifiable, repeatable, quantitative measurement. Like when the split images line up in the rangefinder. It not only looks right, it is right. In the case of the optometrist, it has to be right, or else many of us would have such bad eyeglasses, we wouldn't be able to find our cameras! (Speaking for myself, here.)

@Juan, et. al.: I don't see any way to argue with Roger's position that the only way to ensure a given shadow registers on film, is to measure that shadow. Not that allowing a safety margin won't work most of the time; but "when it absolutely positively has to be there . . ." then using judgment, experience, and doing "what usually works" (like rating the film slower than box speed, for example) might be too risky. And "Who cares what's in the shadows?" merely avoids the issue.

I think the various methods/approaches described here are probably all valid, as long as one get acceptable results. But some approaches are more scientifically defensible, while others are more casual! One way might be good enough for a street photographer; the other, good enough for NASA.

If you can use a quantitative measure -- be it an optical prescription, ISO speeds, whatever -- to make things "look right" (image on focus, adequate exposure), then I'd say you're making things "look right" via psychophysics. And how do you get the quantitative measure? By asking enough people what looks right and measuring that against the relevant variables you can control...

This is not quite the same as "I like what I see", though of course it shades into it. WHY did so many people see the "first excellent print" as excellent? WHY do most people pick 18% as a visual mid tone? Although you can construct all kinds of theories -- some of them testable -- it's like the punchline to the old joke, "Why does it take three women with PMS to change a light bulb?" The answer (fully approved even by the vast majority of feminists) is "It just DOES, OK?"

This is to me the nature of psychophysics: finding common ground, and quantifying it. Which is a lot harder with art criticism!

Finally, I do not want anyone to think that I believe that the only valid means of metering negatives is spot metering the darkest shadows in which you want texture and detail. Clearly it isn't. As you point out, just about any method can be made to work. All I'm saying (again as you point out) is that the only way to be sure of adequate shadow detail, without giving more exposure than you need, is to meter the shadows directly.

The earliest commercial spot meter, the SEI Photometer, didn't even bother with a mid-tone index, but instead used a shadow index (to avoid underexposing shadows in neg films) and a highlight index (to avoid overexposing highlights in slide and movie films). But distressingly many people seem unable even to grasp the concepts of film speeds being keyed to the shadows for negatives and to the highlights for transparencies. In the teeth of the evidence, the published ISO standards and common sense, they believe that a metering off a grey card of arbitrarily chosen reflectance will invariably suffice. Very often it will. Sometimes it won't (unless you fudge).

Cheers,

R.

 

[Juan Valdenebro]

Dear Juan,

Highlight 1: No. That's the point. You never know how much darker the shadows will be. One stop? Five stops? If it's five, well, bad luck.

Highlight 2: Studio advertising; travel; products; food; step by step; all professionally. Across maybe 30 years. In other words, enough to have a fair idea of what I am talking about.

Highlight 3: It's not an opinion. It's a statement of fact. You have yet to explain the difference between reading the light falling on one defined surface (a grey card) and another (a translucent "artificial highlight"). Please do not confuse facts and opinions.

Incident readings -- including your willfully complicated version with grey cards -- work fine for slides. But not for negatives.

Reflect, perhaps, that someone who has written a well-regarded book on exposure theory and practice has probably studied the subject quite deeply and carefully. More deeply and carefully, on present evidence, than you have.

Cheers,

R.

Hi Roger,
I see no reason for a never ending discussion... I mean both of us have the right to an opinion, and instead of what you said, "explain, convince me, or think how I think!" I consider communication is a bit an uthopy, and when we both say "A" at the start of our statements, "A" means different things to both of us even if we try to imagine both "A's" are exactly the same and a good basement to our discussion... But I'll try to answer one last time for courtesy...

You say: "Highlight 1: No. That's the point. You never know how much darker the shadows will be. One stop? Five stops? If it's five, well, bad luck." OK, I understand you mean if you have a scene correctly exposed for medium and high values, you consider it wrongly exposed if you don't give enough texture to the shadows... As I said, me too, but mainly on direct sun scenes where, as I stated previously, I judge the shades are visually interesting. In my opinion, and to me it's OK yours might be different, in most photography shadows don't matter at all in what an image gives: they don't affect what a photograph produces in our soul.
You say: "Highlight 2: Studio advertising; travel; products; food; step by step; all professionally. Across maybe 30 years. In other words, enough to have a fair idea of what I am talking about." OK, it seems you've used slide film and metered for it... Haven't you found situations where incident metering gives a reading that's below optimal? I have, as many others. I was even taught about it... If you haven't seen that happen, sincerely it's a miracle. Why would other metering mediums exist if incident metering is a perfect absolute for every light situation?
You say: "Highlight 3: It's not an opinion. It's a statement of fact. You have yet to explain the difference between reading the light falling on one defined surface (a grey card) and another (a translucent "artificial highlight"). Please do not confuse facts and opinions." OK, this is about considering or not all metering mediums are the just same thing... Couldn't it be you the one that's confusing facts and opinions? I mean maybe you're not as absolute and perfect as an incident meter...😉 Roger, seriously I see big differences between them... To me it's not the same metering the light that hits the scene, than metering the light reflected by the scene... One of the reasons is that an incident meter meters, inside its hemisphere, several different lights coming from different angles, and promedy that, in a way that is not the same as when a spot meter considers the lights that are being indeed actually sent to hit film...
Why don't we leave this here like this: I accept you're a great photographer who cares more about detailed shadows than I do, and who knows better than I do how to achieve it, while I'm an inferior photographer who cares less about shadow detail, being it possible that I don't know how to get it.
Cheers,
Juan

 

[Sparrow]

... I'm just thankful no one has mentioned the metre's K-factor yet, would't it be tricky if they didn't all read the same value for the same scene-luminosity?

 

[David Hughes]

Money mostly! I only have an incident meter and haven't been able to afford another meter until now.

If possible, I tend to take an incident reading from the subject's position back towards the camera, and then slide the cone back to meter the shadows up close. I then try to make as much of an informed decision as I can about how to proceed based on the difference between the readings and the shadow areas in the subject.

I think I was just having an off period, or my 4 year old was taking more of my attention during that time. He's completely fascinated by the light meter, so it often comes back to me with the dials all over the place 🙂

However, I went out this past weekend and was able to concentrate and be more methodical. Glad to say I was able to get several rolls that I'm pleased with that had good exposure in varying light levels and conditions.

Last week I found an absolute bargain of a Pentax Spotmeter V on eBay, so I'm very much looking forward to using that if only because I think using the Zone System will provide a framework within which I can build some structure and learn in a somewhat more deliberate manner.

Cheers,
Laurence.

Thanks, like I said I was just curious.

EDIT, much later: if you slide the cone back then you are presumably using a reflective reading then? FWIW, in landscape photography I point the meter/camera at a bit of grass in the mid to foreground and use that reading. But again I'd take that as a start and adjust slightly if I think it's needed. I'd only consider a spot reading if a minor part of the view was important, the classic example being bluebells or other flowers in woodland.

Regards, David

 

[Roger Hicks]

Hi Roger,
I see no reason for a never ending discussion... I mean both of us have the right to an opinion, and instead of what you said, "explain, convince me, or think how I think!" I consider communication is a bit an uthopy, and when we both say "A" at the start of our statements, "A" means different things to both of us even if we try to imagine both "A's" are exactly the same and a good basement to our discussion... But I'll try to answer one last time for courtesy...

You say: "Highlight 1: No. That's the point. You never know how much darker the shadows will be. One stop? Five stops? If it's five, well, bad luck." OK, I understand you mean if you have a scene correctly exposed for medium and high values, you consider it wrongly exposed if you don't give enough texture to the shadows... As I said, me too, but mainly on direct sun scenes where, as I stated previously, I judge the shades are visually interesting. In my opinion, and to me it's OK yours might be different, in most photography shadows don't matter at all in what an image gives: they don't affect what a photograph produces in our soul.
You say: "Highlight 2: Studio advertising; travel; products; food; step by step; all professionally. Across maybe 30 years. In other words, enough to have a fair idea of what I am talking about." OK, it seems you've used slide film and metered for it... Haven't you found situations where incident metering gives a reading that's below optimal? I have, as many others. I was even taught about it... If you haven't seen that happen, sincerely it's a miracle. Why would other metering mediums exist if incident metering is a perfect absolute for every light situation?
You say: "Highlight 3: It's not an opinion. It's a statement of fact. You have yet to explain the difference between reading the light falling on one defined surface (a grey card) and another (a translucent "artificial highlight"). Please do not confuse facts and opinions." OK, this is about considering or not all metering mediums are the just same thing... Couldn't it be you the one that's confusing facts and opinions? I mean maybe you're not as absolute and perfect as an incident meter...😉 Roger, seriously I see big differences between them... To me it's not the same metering the light that hits the scene, than metering the light reflected by the scene... One of the reasons is that an incident meter meters, inside its hemisphere, several different lights coming from different angles, and promedy that, in a way that is not the same as when a spot meter considers the lights that are being indeed actually sent to hit film...
Why don't we leave this here like this: I accept you're a great photographer who cares more about detailed shadows than I do, and who knows better than I do how to achieve it, while I'm an inferior photographer who cares less about shadow detail, being it possible that I don't know how to get it.
Cheers,
Juan

Dear Juan,

Sorry, not in this case.

I am perfectly happy to accept that you may well be a better photographer than I. And I sincerely doubt I normally care more about shadow detail than you. As you say, often, no-one cares. But "don't care" is not always the same the same as "doesn't matter".

In other words, much as I appreciate your kind and generous offer not to prolong the discussion, and although I think you are being far too kind and polite in your assessment of our relative abilities as photographers, this is not a question of opinion. It's pure exposure theory (and practice).

Yes, the shape of the incident light receptor will greatly influence the reading: an Invercone reads a lot more than a plain dome, with very different weighting; ; a rolling shutter (as on the Gossen Sixtomat) is different again; and there are times when only a flat receptor will give the same results as a grey card, which is why a few (a very few) meters have interchangeable receptors, or receptors that can be retracted or hooded.

But you're the one relying on near enough/doesn't matter when it comes to shadow detail, and if we're playing that game, I'd suggest that the same is true of incident light receptor shapes. It seldom matters. Except when it does.

Cheers,

R.

 

[lrochfort]

This thread has certainly been interesting, but could I perhaps request a closing summary for myself, and future readers?

I understand that for negatives outside of the studio, metering the shadows with a spot meter is most desirable. However, what if one only has an incident meter with a sliding dome/cone/white bit of plastic? What is the best compromise in that situation?

Thanks all!