Metering and over-exposure phobia

[e_No]

Now that I have a blog, I'm re-igniting an old conversation by re-posting an old article I did comparing D80 vs. D90 matrix and spot metering against straight sunny 16 results. I thought folks here at the Learning center might find this useful.

Sunny 16 & D80 vs. D90 metering

I will use this as a basis for an upcoming article where I'll compare D7000 metering against the D300 and D700 (and a D90 if I can borrow one).

Also,iIn light of all the recent discussions about D7000 over-exposure, I've been reposting some articles from my site that I wrote prior to my blog's start mid-July 2009. They concern mostly metering behavior with the Nikon D80 and D90 and strategies to balance exposure in the types of harsh, high contrast/DR situations that seem to challenge Matrix metering most.

Today's re-post, The cure for over-exposure phobia contrasts what can be done in ViewNX to address under vs. over-exposure.

If you want to keep track of all past and future articles that deal with metering, use this link:
Metering series

Next week I will have a D7000 for some testing and may also borrow a D90 to do some comparison testing (along with a D300 and D700). Keep following along if you're interested in this topic.

 

[Longshanks]

Some good info here,

Like you did, I always expose -0.7 in bright situations, and if someone was wearing a white dress or shirt it still could take a lot off pp to get some details back.

The next white shirt that I shoot will be spotmetered in zone 7.

Thanks for taking the effort.

 

[e_No]

The next white shirt that I shoot will be spotmetered in zone 7.

Cool trick, huh? I had only tried it for landscapes (to preserve details in white rock or snow-capped mountains), but it's good for clothing, too, isn't it? As in wedding dresses. One of the most difficult challenges wedding photographers face is to keep the detail in those white, sometimes shimery dresses from blowing. The companion challenge is to get the skin tones to feel natural (i.e., not overly dark for the person's complexion). Then you add flash into the equation, and things get really sporty. :redface:

 

[Longshanks]

Next week I will have a D7000 for some testing and may also borrow a D90 to do some comparison testing (along with a D300 and D700). Keep following along if you're interested in this topic.

Well you have at least 1 follower on this thread, the over-exposure phobia makes me grumpy. :biggrin:

 

[Palouse]

Fabulous stuff. Very timely too--I'm shooting (third string at best!) a wedding on Saturday!

 

[e_No]

Fabulous stuff. Very timely too--I'm shooting (third string at best!) a wedding on Saturday!

Congrats! I'm shooting one as a 2nd shooter this weekend as well... as soon as the 1st photographer gets back to me with information of where to show up.

 

[Pa]

I have owned and used the following Nikons: D70, D70s, D40, D60, D200, D90, and now D7000. They all overexposed to a degree, especially in bright conditions as you mention, but the D7000 is by far the best of the lot. I always had -0.7 eV compensation dialed in on the D60 and -0.3 on the D90, and frequently had to go more negative. But the D7000 nails it most of the time. This is based on matrix metering.

These comments come from subjective experience. I've done no controlled tests.

But I am definitely interested in your findings.

 

[Longshanks]

Pa,

Do you have the ADL feature turned on or off?
I have read that ADL underexposes a bit and then boost everything except for the highlights.

 

[SP77]

Nikon's lousy matrix metering on the consumer and lower level bodies was one of the most frustrating things about the system to me. I too am curious how the newer bodies are doing with this.

I remember getting so frustrated and annoyed with my D90 that I completely gave up on auto exposure modes and reverted to full manual with an external meter, and actually got better and far more consistent results. It couldn't even shoot a tree in the middle of a field in mid-day sun properly. It saw the leaves and the shadows underneath them and went crazy adding 3 or 4 stops of exposure to try to make the shadows on the leaves grey, which made the rest of the photo look like a nuclear bomb had just gone off even though that was only a small portion of the frame and this was "matrix" metering mode, not center-weighted or spot.

My D200 and D50 seemed to be pretty good. The D40, D80, and D90 were all awful. My dad's D60 doesn't seem to be too bad, but I haven't shot with it a ton. They went on an Alaskan cruise a year or so ago and left in green auto mode, most everything came back perfectly exposed.

 

[e_No]

Nikon's lousy matrix metering on the consumer and lower level bodies was one of the most frustrating things about the system to me. I too am curious how the newer bodies are doing with this.

Would you be surprised if I told you the D300's Matrix metering is also "lousy." See this link for some examples (all shots taken with the D300).

 

[SP77]

Would you be surprised if I told you the D300's Matrix metering is also "lousy." See this link for some examples (all shots taken with the D300).

I actually prefer the original matrix metered shots for those. :redface: White flowers in what looks like bright sun ought to be pretty brilliant looking, and that's what you got in the original shot. The others looked too dark. At least in my book it's ok to start clipping highlights here and there in the name of a more vibrant looking photo.

Anyways my issues with the matrix metering was with how it handled darker objects anywhere close to the center of the frame. It would no longer be matrix metering, but rather turn into center-weighted or spot metering all based on a darker area at or near the center of the frame trying to turn it grey, which I thought was ridiculous. Anybody with a darker shirt and it would go bonkers, adding sometimes in excess of 3 stops of exposure. It refused to leave "shadows" as shadows and insisted they must be turned grey, thus resulting in "nuclear" looking exposures. :tongue: The metering assumes that you're stupid and that you're looking at a face in a shadow and don't know to use fill-flash so it tries to "fix" that for you, except usually for me it ended up destroying the photo completely.

I haven't shot with a D300 enough to know how it behaves in such situations, but the D200 and D50 were pretty well-behaved. The D80 was awful, and so were the others I mentioned. I became extremely disappointed with the D90 because every review I read said that the metering was so much better than before, but I found it to be just as bad. Desmond posted up some D90 tests that were quite telling proving that.

 

[e_No]

But I am definitely interested in your findings.

Camera just got delivered. Will do some early looks early this evening.

 

[e_No]

I actually prefer the original matrix metered shots for those. :redface: White flowers in what looks like bright sun ought to be pretty brilliant looking, and that's what you got in the original shot.

Don't take this personally, but I think a better understanding of light might yield a different assessment. I say this as someone that's had to grow an eye for light as much as I developed it for composition. If, OTOH, the shot is about showing glowing flowers, with little priority on pedal textures, then I guess you're right. :wink:

 

[SP77]

Proper exposure is not just technical and a matter of keeping between 0 and 255. It's personal and subjective and a matter of artistic taste and whatever 'look' you might be going for as well.

 

[camerapapi]

Proper exposure is not just technical and a matter of keeping between 0 and 255. It's personal and subjective and a matter of artistic taste and whatever 'look' you might be going for as well.

+ 1. In total agreement.
I use very often the "sunny 16 rule" with my cameras. It is a guide and I modify the exposure based on what I see in my subject.
Matrix is a great metering modality because it can evaluate a scene and use algorithms to come pretty close to a very good exposure, but it is not infallible. Look at the histogram when in doubt and proceed as per your experience. I agree with Eduardo that it is not the best method of exposure for the bright areas.
Spot metering is very accurate in the hands of an experienced photographer and center weighted has been our standard for many years. Modern center weighted meters can vary their circle of acceptance from the usual 75% center weighted. I do not have experience using center weighted metering with less bias to the center but I figure that increasing the area of sensitivity outside the center could call for erroneous exposure if the photographer is not careful.
Using an exposure meter calls for good judgment on the part of the photographer. That judgment comes from experience. Nikon cameras tend to blow the red channel often and I do recommend in critical situations to always look at the color histogram. Do not think that because you are shooting RAW you have the cure for improper exposure using software. Try to get it right in camera.
As John Shaw uses to say in his workshops "If trash gets in, trash gets out."

William Rodriguez
Miami, Florida.

 

[ddietiker]

Proper exposure is not just technical and a matter of keeping between 0 and 255. It's personal and subjective and a matter of artistic taste and whatever 'look' you might be going for as well.

Bingo, evaluating a scene and deciding for yourself what is important and how you want to show it is the "correct exposure."

 

[e_No]

Proper exposure is not just technical and a matter of keeping between 0 and 255. It's personal and subjective and a matter of artistic taste and whatever 'look' you might be going for as well.

No disagreement from me there. Exposure is an artistic choice before it is a technical one, as Bryan Peterson tells us. An exposure you didn't take care to manage, OTOH, may well destroy or detract from the artistic choice you were trying to convey.

 

[Weebles]

I shake my head when people say something like "check the histogram", implying it's a way to check exposure as though it were a way to evaluate metering. And there is also the often mentioned, "check the histogram and expose to the right" comment.

However, neither of those statements, or what is often implied by them, are usually correct because the histogram is not a light meter.

Although exposing as much as possible without getting blown out highlights indicated by 'blinkies' is often a good idea, the histogram is not always the best indicator for that.

As an example, most people can not look at a scene and precisely describe what the histogram should look like prior to taking the shot. And in fact many can not look at a series of images along with an equal number of histograms and then accurately associate those images with their respective histograms. Therefore, they also can't know if the histogram for the shot they just took is correct or not. Obvious examples, such as the black cat in a coal mine analogy, notwithstanding.

 

[camerapapi]

We have had excellent discussions about meters and metering techniques but illustrations have been lacking. What about "sunny 16?"

Subscribe to see EXIF info for this image (if available)

I made this picture using the "sunny 16" rule. These auxiliary boats, near Ogunquit, Maine, were lighted by soft, directional late afternoon light. Anywhere on this planet, from 2 hours after sunrise to 2 hours before sunset, the levels of light remain unchanged unless some clouds cover the sun or in the presence of other atmospheric conditions that alter sunlight and that I do not need to explain here.
I have shot hundreds of pictures this way, so reliable it has been for me. Just keep in mind that although "sunny 16" is an excellent guide to exposure in sunlight there will be occasions when judgment will be necessary in order to obtain a good exposure. That comes with experience.

William.

Addendum: If you need to learn more about exposure I do recommend you look at Eduardo's series on exposure meters in his excellent blog.

 

[e_No]

I agree. The "correct" histogram in an outdoors shot under full sunlight looks very different than one under shade, and yet different than one indoors where shadows are prevalent, and only the subject is in light. Still, at the end of the day, if highlights are important to you, the histogram will tell you if they're clipping. It can also tell you if your midtones are where you want them to be, and even if the white balance is correct. No, it's not a meter, but then again, judging everything by the meter isn't always reliable either, and there too, one must know what he/she is looking at. The meter and histogram are tools that when used together can help assess whether you are getting out of the shot the exposure you want. And that's the key: knowing what you want, then matching it to a meter reading and histogram result that reflects what you want.

And yes, I can now tell what the histogram should look like for different types of scenery, roughly, though I am still sometimes surprised.