exposure meter: tech question

[NRothschild]

Just a different take on all the technical explanations. I'm not taqlking about a flash meter- that is a more complex subject. I'm talking about basic available light exposures. I bought a Sekonic meter because I wanted to learn more about metering. I learned a few things, conceptually, about how my cameras meter, but I rarely if ever use it. Here's why....

As said above, the camera reads reflected light and the light meter reads incident light. In theory, if you know the incident light value you can set you exposure.

In reality, to get it perfect with the limtied DR of DLSR sensors, after you figure out your light meter's exposure and apply it to your camera, you will need to shoot the scene and check for blinking highlights and/or the histogram. You are then going to tweak your exposure accordingly because you don't want to argue with camera about how it should have exposed the scene. You will lose that argument every time

So... whichever method is used, you end up doing the same thing anyway- shooting the scene and then tweaking to make sure you get the highlights the way you want. You can just save money and carry less gear by using the camera's meter.

I can understand why light meters were so important with film, where you often had more dynamic range AND, more importantly, did not get instant feedback on the results. With digital, I have not found a good reason to use the thing and, more importantly, to justify the cost and additional baggage, except as a learning tool.

Regards,
Neil

 

[Dennis_Josiah]

Great response, thanks.
--
Dennis
http://www.dchimages.com
'Seeing with the light of the world.' Gen 1:4 - Matt 5:14

 

[dannv]

You will
lose that argument every time

while i generally agree with your points. this one is not accurate. see below.

I have not
found a good reason to use the thing

reflectance meters are much more easily fooled than incidence -- particularly in the case of fuzzy logic metering systems such as Nikon's 3D Matrix. a better summary is it takes knowledge of basic exposure concepts to use any meter correctly. failing that you are just relying on trial and error which will work with any stable metering system (which one can argue Matrix isn't, by the way)...dav

--
don't wait for technology -- it won't wait for you

 

[NRothschild]

What I mean by "losing the argument" is that if I meter a scene with an incident meter, apply any standard corrections that are necessary for my particular camera/lens/light meter, shoot the image, and I burn a highlight I don't want to burn, I can't argue with the camera about the correct exposure.

This frequently happens with the limited DR of our sensors. The same thing can and frequently does with Matrix Metering.

At the end of the game, we are or in most cases should be exposing to our historgram (and by extension, the blinkies). Everything else is theory if the histogram is wrong. If I can almost always get my highlights the way I want them with only one or two test shots, using the camera's meter, I'm not sure how an incident meter can help me. If it would get me to the histogram I want with fewer shots, there might be some value, but I'm probabloy going to need one or two no matter what I do.

Remember that even though matrix metering is imperfect, at least it has some idea of the dynamic range of the scene verses the sensor. The problem that matrix metering has is that it would tremendously underexpose if it tried to never blow anything. That's where the skill and daring part comes in. My job is to figure out what to blow if something has to blow.

The incident meter assumes I have unlimited DR, which is simply not realistic. I've never spent much time testing the two meters, but I would think matrix would come closer most of the time if that 30K database of scenes has any value.

I'm not referring to spot metering an incident meter. Those meters are even more expensive and I've never quite figured out why that could be better than the camera's spot meter, which is far more easy to use.

--
Regards,
Neil

 

[dannv]

implicit in your comments seems to be the idea that an incidence meter gives you the exposure to use in the same way a camera's metering system does (or tries to). in most cases, it doesn't. if gives you a starting point(s) to begin your calculations.

i've seen nothing to indicate Nikon have updated their difference engine to account for the DR of DSLRs, much less individual cameras. so i'm skeptical of the contention matrix metering takes DR into account.

here's the best way to view it IMO:

• a good metering system with standard (ie: not extreme) scenes will return a reasonable exposure the majority of the time -- and will do it quickly and with little effort on the part of the user
• a knowledgeable photographer using an incident meter will get correct exposure every time if they don't make a calculation error -- this will take some effort on the part of the user. this approach has the benefit of handling extreme lighting more reliably.

when you factor in that everyone is going to make some calculation errors -- on balance most people prefer to avoid the effort. but note that i made a distinction between 'reasonable' and 'correct'. exposure is a choice. there is no correct exposure other than the one the photographer chose. using an incident meter gets you there more reliably than even using the camera's meter and then making your own calculations.

spot incidence meters have a much narrower FOV -- usually 1 or 2%. camera's internal spot meters are wider and change with focal length of the lens employed.
...dav
--
don't wait for technology -- it won't wait for you

 

[NRothschild]

I keep talking about histograms and you keep talking about meters. To me THE ONLY THING THAT MATTERS is the histogram.

I have no doubt that an incident meter will give me an image with a correct mid-tone if done correctly but often a correct mid-tone isn't going to preserve the important highlights.

You need to walk me through how you would use an incident meter to correctly compute the exposure such that I never blow important highlights AND get perfect mid-tones. Not a spot incident meter, though, because I can spot meter a highlight in camera and get it right the first time. The only reason (for me) to drag an incident meter would be if the incident meter had a smaller spot and that were critical to get the highlight right. But....

at the end game, I can always just dial in -1/3 stop, click, rinse and repeat to get to my no-blown highlights image and usually that is only a couple images, verses playing with a separate meter and doing calculations because I usually know up front when I need to go down more than 1/3 stop. All I care about is I don't blow the highlights and I get it done quickly and simply. If I don't like the mid-tone level but haven't blown any highlights, I can do whatever I need to do in post, assuming the mid-tone isn't obvious on the histogram- it usually (but not always) is.

--
Regards,
Neil

 

[italy74]

Hi there,

thanks A LOT to all, I neither noticed so many answers at first..... Just done now!
--

Photography is the most beautiful way to discover God around us in little and simple things.
My kits under my profile

http://italy74.smugmug.com
http://it.pg.photos.yahoo.com/ph/dlb74pr/my_photos

 

[BoBCita]

I am no expert at all but I think I understand your point:

what matters is if the picture turns out like I want it to be and you seem to believe an external meter does not give you any advantage in this respect because with a digital camera you can easily and quickly check the result (right?)

Whereas I do not understand the othe poster's point:

let us accept that an external meter can give you a physically more reliable reading (I am a physicists and I know there are standard units for everything), but does that really help with your camera? the external meter does not know anything about how your camera sensor works (neither do you or we... just the techs who designed those parts could really tell us). So why bother to spend more money?

BoB

I keep talking about histograms and you keep talking about meters.
To me THE ONLY THING THAT MATTERS is the histogram.

I have no doubt that an incident meter will give me an image with a
correct mid-tone if done correctly but often a correct mid-tone
isn't going to preserve the important highlights.

You need to walk me through how you would use an incident meter to
correctly compute the exposure such that I never blow important
highlights AND get perfect mid-tones. Not a spot incident meter,
though, because I can spot meter a highlight in camera and get it
right the first time. The only reason (for me) to drag an incident
meter would be if the incident meter had a smaller spot and that
were critical to get the highlight right. But....

at the end game, I can always just dial in -1/3 stop, click, rinse
and repeat to get to my no-blown highlights image and usually that
is only a couple images, verses playing with a separate meter and
doing calculations because I usually know up front when I need to
go down more than 1/3 stop. All I care about is I don't blow the
highlights and I get it done quickly and simply. If I don't like
the mid-tone level but haven't blown any highlights, I can do
whatever I need to do in post, assuming the mid-tone isn't obvious
on the histogram- it usually (but not always) is.

--
Regards,
Neil

 

[NRothschild]

Bob,

You understand my point exactly. And remember that although "incident meters" was discussed, my understanding of the application of the zone system requiries an incident meter with a very fine (say 1%) spot reflectance metering ability. That does not come cheap.

Taking a simple reading with an external incident meter, adjusted to any consistent differences relative to your camera, should give you a perfectly balanced mid point. In an evenly lit landscape with a green grassy foreground, the grass should be perfectly exposed because grass is very close to 18% gray in reflectance. That is one data point you are looking for. The other data points are the true outer zones of the scene; how to squash that into the 5 or 6 stops of DR that we have in our sensors is certainly more art than science. There is no "perfect mathematical exposure" except in theory where the film media has unlimited DR. A degree in Art rather than Physics would be more helpful there. I think that is an important philosophical point. (I'm a software guy)

My experience with a D2H and D70 is that, with the D2H in particular, matrix metering will often deliver a brilliant exposure- mid-tones are acceptable if not perfect and highlights are just to the left of the histogram. If it had added 1/3 stop exposure the highlights would be blown, 1/3 less and the mid-tones would be unnecessarily dark.

About half the time, I have blown highlights, even if the highlight fills the center of the frame, such as this image:



There are no blown pixels in that image.

In this example, I exposed (aperture priority) -2/3 stop from the matrix metered value. I always have to expsoe white Egrets (in bright light) -2/3 to -1/ 1/3 stops down, although the D2H is a little more consistent in that regard at about -2/3 stop, as here. I don't understand why Matrix Metering does not recognize that the central brightarea is a critical highlight, but it really doesn't matter because I know the meter and I know my subjects and I know before my first test shot if I go down -2/3 stop, I will either nail the exposure or be withing 1/3 or 2/3 stop. If I don't like the histogram, I adjust accordingly. I can do that faster than I can play with a hand held meter and with my favorite shooting (wildlife) I don't usually have much time to do this and I really don't want to think about it more than I have to.

I could spend $400 on a 1% reflectance spot meter and try to zone system the thing, but my final exposure will be identical because I want the bird's highlights as far to the right as I can get, without blowing more than a few tiny patches. Considering I shot this bird at 700mm, the field of view (vertically) is only about 1 degree and in this admittedly extreme situation, a 1% spot meter would be less refined then center weigting the scene in camera. Even for a typical scene, though, the in camera spot is pretty small. I've never calclulated the max focal length that a handheld 1% meter would provide a smaller spot than the camera's meter. To me, that isn't reason to go through the process of handheld metering.

Now... the image above, although a bit on the extreme side unless you shoot Egrets or weddings, is a very simple exposure because there is no deep shadow detail I care about. For other relatively simjple exposures of street scenes, landscapes, etc., I tend to find that I am oging down 1/3 to 2/3 stops from the matrix value. I almost never have to compensate to the plus side, It's easy to get there and in most cases my exposure strategy is very simple: the higlights are beyond the sensor's DR so I am going to expose as hot as I can without blowing important highlights. That keeps my mid-tones as close as possible. I shoot raw and adjust exposure in Capture to get the exact final exposure I want.

 

[NRothschild]

(sorry, this post was too long

To be very fair, there are far more difficult scenes to expose. I saw one last night related to another internet discussion where the scene was a stream going into shaded trees, with sunlight overpowering a central part of the scene, with another major part of the scene maybe a stop or two down. The shadow detail was also very important and it was deep to the left of the histogram. The photographer happened to be discussing this same issue: does he really need the handheld meter he's been using for 20 or 30 years with his film gear now that he is shooting digital with a Nikon body? I wouldn't weigh into that discussion with my strategy because my strategy would be rather crude: I would shoot a matrix metered frame and then try to back down past the highlights I know I can't keep, and see how the shadows look on the histogram and preview. I would then bracket around that value, depending on what I saw and deal with it in post. There are better ways to deal with that scene with spot metering (in camera or handheld) but it would take a huge amount of experience to get that scene recorded correctly without some serious bracketing. I think he had to blow even some of the lesser highlights (sun dappling the trees) but he did it in such a way that it looked good and his shadow areas didn't block up badly. He also did a bit of Photoshop work, equivilent to what he probably did with film by dodging and burning his prints.

So... maybe it becomes a religious/philosophical thing. The only way to do the scene I described with one shot would probably require a handheld spot meter and a very good feel for the blocking issues at the black end and the burning issues at the white end of the sensor's light curve (histogram). Presumably my 1/3 stop bracketed series would include the same exposure that he chose. There is a very finite number of exposures you can do. I'm taking advantage of the technical advance that comes with "free" digital film. He is using a technique developed in the day when film and processing was expensive and getting the shot right the first time was very important.

I admire and respect that skill but at the same time I am just trying to get good well exposed images with the skills that I have and, more importantly, to be able to apply a basic methodology to either a static landscape or a highly dynamic scene like a bird in flight that I can never spot meter and I may never be able to recapture the moment if I blow it.

What is most important to you? The end result- a well exposed image (possibly one of six or more in a bracketed series) or a single well executed image? It's a philosophical decision.

--
Regards,
Neil
--
Regards,
Neil

 

[BoBCita]

Well.... what can I say?: thanks a lot for your time and extensive explanation.

As I said, I am a physicist but also believe that the aim of photography is art (note that I say "aim" not that we all reach that point and no tool can replace the photographer's perspective in that respect just because no tool can read in his/her mind and realize what's in there. In this sense, it becomes even questionable (to some extent) to say that a picture is "perfectly exposed" or "not", ain't it?

--
BoB

 

[NRothschild]

"The perfect expoosure is the one you made"

That is the "art" of photography, in my book. I was very aware that you are a Physicist trying to be an artist and I tried to weave that into my tome. I don't mean that unkindly, we are kindred souls.

To really simplify things, there are results and then there is the process. In the film days, bracketing 10 shots may not have been viable (I'm going back to sheet film days). Any idiot can figure out an exposure within a stop or two with instant preview, histograms and the blinkies. There are finite exposure levels- the smallest we can do is 1/3 stop. With a little sensible bracketing, any idiot can come home with at least one image that will be identical (aside from DOF and motion blurr considerations).

For some people, this is about the process of working to that perfect exposure that THEY want without chimping and bracketing. That is a legitimate pursuit- who am I to tell someone how to enjoy themselves and I respect them tremendously if they can do it.

There is a difference between what you NEED and what you want to do and I think in these meter battles, it is important to make that distinction. Especially here where if someone says you NEED a $400 gadget, people go out and buy one

I guess I'm glad I bought my meter because I want to improve my understanding of not only the art of exposure, but the technical process or arriving at it. I would never recommend that someone should buy one, though, unless they expressed a similar objective.

--
Regards,
Neil

 

[dannv]

ok, wrapping up as best i can here. simply put: reflectance meters are more easily fooled than incidence. if your basic point is you can take a picture with either and adjust it then take another shot i can support that.

i've been ignoring your histogram comments because a) we're talking about use of meters and b) i'm not as big a fan of histogram based shooting as you are. the answer to both this issue and the later one you posed (walk you through how to use an incidence meter...) are basically the same: i get the impression your subjects are, for lack of a better word, of a standard nature that lends itself well to matrix metering and balanced histograms. for people who have more non-standard shooting styles involving content that is likely to fool a matrix meter and/or not be pleasingly depicted with a balanced histogram, an incident meter that is not easily fooled is very useful. there are many examples of mixed mode, mixed lighting shooting that give a reflectance-based fuzzy logic system fits.

if you get what you want from a matrix meter, go for it. for my part, i only use matrix when i'm in a hurry, need to shoot fast changing subjects, and am relatively confident of the subject matter -- for my shooting style these situations are rare.

WRT Bob's question about trusting a reflectance meter that "knows" the camera's sensor i'll repeat what i said earlier: i know of no indication Nikon has made any other than very gross adjustments to the matrix metering system for different DSLRs. the underlying scene db and fuzzy logic engine have almost certainly not been changed as sensors change. many of us suspect, however, Nikon put a 'fudge-factor' in on cameras like the D100 which caused them to underestimate exposure....dav
--
don't wait for technology -- it won't wait for you

 

[NRothschild]

Fair enough, dannv, I agree with most of what you said, and I think we are all talking about available light here; studio flash shooting is another thing entirely.

Just to elaborate on the histogram a bit; I think of it as a very sophisticated reflected light meter. I have learned to deal with strange light that doesn't make the classic mid-tone bulge. With tricky light, I will use spot metering and manual exposure. As I understand the use of the zone system, once you know your incident light you then need to meter the luminence of key portions of the scene and establish zones. That requires a reflectance spot meter and now we are down to which spot is more accurate or which spot is smaller. Your 1% spot meter may be smaller than my camera's spot sensor on a 50mm lens, but at 700mm mine beats yours There are ups and downs to both methods. It never hurts to have another tool in the box but this is a $400 tool we are talking about (for a decent incident/reflected narrow spot meter).

I find that matrix metering rarely underexposes. It almost always overexposes because it can't know what highlights we want. I had a recent situation where it did underexpose- dramatically. I was photographing a guy welding a pipe. The intence brightness of the torch caused the matrix meter to go way down in exposure. Interestingly, the matrixed image that was "underexposed" looked great to me and I used that version rather than something I tried to spot meter. The actual torching operation occurred so briefly that I couldn't get meter readings. Interesting metering problem.

My method is crude but effective. Assuming the matrix meter overexposed the highlights, I back off exposure until the highlights I want aren't blinking. In theory, I might want them even lower but in practice I can do that in Photoshop or Capture without any penalties. The penalty for underexposing relative to my final image, though, is noise, so I want my retained highlights as bright as possible unless I see something in that specific scene that tells me otherwise. If I am confused, I bracket around the range that is confusing me and those extra digital frames cost me nothing. Not as elegent as using an incident meter but I think we come out with the same image after post.

I think the key here is the "process"; some people such as yourself have the skill and daring and time and interest in going through the process and getting the exposure the "right" way, rather than my chimping and bracketing. I respect your skills, very very much.

I think it is important for these discussions to distinguish between what we NEED and what gives us more enjoyment in practicing our craft, especially in these gear-crazed fora.

--
Regards,
Neil

 

[Bob Topp]

I would like to suggest another point of view for metering. The discussions thus far seem to have dealt primarily with metering to reproduce a scene faithfully, although Mr. Rothschild did point out that his goal is to insure that the midtone lies on the histogram, thus making it within the realm of adjustment. The zone system was talked around, but not really addressed.

If you view the meter primarily as a means of artistic control (specifically the spot meter), then you look at a scene and determine what YOU want to set as the midtone or the highest or lowest tone. From there, you preset the exposure compensation, meter off the reference tone, recompose and shoot.

If the reference tone is to be a light tone in your composition, you dial in postivie compensation, meter off the reference and shoot. If your reference is to be a dark tone, you dial in negative compensation, meter off the reference and shoot. ...and so forth. The question the comes down to how much compensation to dial in. Going strictly by the zone system, 0 would correspond to an 18% middle gray, -2 to black and + 2 to white; however, my D70 and D200 allow me to dial in -5 to +5, so I have to re-think where I want my tones to lie, or more correctly, experiment with the camera until I know how much detail I will be able to see in various levels of compensation.

This approach may produce blown highlights, and it may produce shadowed areas with no detail. It is up to the photographer to make some tradeoffs in advance, as is true of shot selection, cropping and everything else. It is generally conceded that there is no detail to be lost in specular highlights, such as sun on the water; but in the case of Mr. Rothschild's heron, I suspect the first goal was to preserve feather detail, second to avoid blown highlights if possible.

Systems such as Farzads http://www.spotmetering.com/ have taught me a lot about using my camera's meter, but most are still grounded in film experience. Nothing will beat getting familiar with your camera's built-in metering and how it works best with your post-processing workflow.

Regards,

Bob

 

[Murray Bowles]

Actually, the drawing in the original D200 brochure shows a new box in the matrix metering algorithm which APPEARS to be there do deal with blown highlights. I remember when I first saw it i guessed they added something to look for small very-bright regions (matrix metering in general collapses the 1000 (or 500) metering pixels down into a much smaller number of regions before analyzing things).

But that's all guesswork -- Nikon hasn't published any meaningful information on matrix metering since the first FA white paper.

 

[NRothschild]

Bob,

There is actually little or no difference between the classic zone system (used to faithfully reproduce a scene) and using the histogram to avoid blown highlights.

Any discussion of the zone system I have seen, including your description, basically says to spot meter the white area (classically for slides - now digital) and plus up about 2 stops. Since not every scene includes a snow capped mountain or a white Egret, the photographer must then find a zone 6 area of the scene and plus it up one stop., etc.

The intent, assuming we have a zone 7 white snow capped mountain, is to get white to look white on the slide without blowing the highlight. To me, white on slide = 255 on histogram. The big difference is that there is a solid wall at 255 on digital where there is a little lattitude on slide.

The problem with the zone system, as I see it, and as applied to digital, is that most scenes have tiny areas near zone 7 that cannot be metered due to their small size, but we don't want to blow. Most scenes don't have the classic snow peaked mountain or snow field. I guess that was part of the black art of the zone system.

My histogram is just a tool to put white into zone 7, hopefully without blowing the whitest of the whites.

The only reason I use Matrix most of the time is that my subjects, as well as the light, are often dynamic. The best light of the day is very fragile.

I think we should not mix "creativity" with basic exposure concepts because they are two different things. If I want a snow capped mountain to look a little gray (zone 6) then I add a step to the basic formula. I would stop down one stop (or whatever I choose) from the zone formula, regardless of methodology,

--
Regards,
Neil

 

[dannv]

thanks for reminding me of that diagram. it does imply there may be something more sophisticated happening. ...dav
--
don't wait for technology -- it won't wait for you

 

[SandyR]

Yes, that's what I meant - over/under exposure. Technically I'm not using right words.
--
===
Sandy

 

[jcase]

Thanks for the link

Systems such as Farzads http://www.spotmetering.com/ have taught me
a lot about using my camera's meter, but most are still grounded in
film experience. Nothing will beat getting familiar with your
camera's built-in metering and how it works best with your
post-processing workflow.

Regards,

Bob

--
John Case
Case Photography



http://www.CasePhotographyOnline.com