Light Metering

[greg_lisi]

<p>Brought my D3/70-200 2.8 II to my church where there was to be a catholic school graduation. Lighting was typical for a cathedral type church facility (lots of tungsten). I shot raw, AWB, manual, matrix, ISO ~1600-3200, no flash, and adjusting A and S as needed to try and keep the peaks in the middle (exposure). This time, just for the heck of it, I brought my Sekonic and walked around the Altar area before anyone arrived. I took some readings and jotted them down. Time to shoot, this time I ignored the constant minus side readings I was getting on the camera meter instead applying the settings to what my Sekonic was suggesting. At the end of the day, dump into NX3. I had a spot-on exposure keeper rate that I haven't been getting in the past using the averages of the camera's analogue meter (very inconsistant). Question is, meter accuracy. Is it time for me to send my D3 in for a check-up? Are there any other tests I could run to make a determination?</p>

[CvhKaar]

<p>It is usualy better to meter the light direct , as you did this time.<br>

When metering from the camera, you measure reflected light which then gets calculated back by the camera to expose for "Mid Gray" (reflection of avarage cocasian human skin..). That's also the reason for having to "Over expose" on mainly white / snowy subjects to bring back white to White, or Underexpose dark / black subjects to get real black's : the camera wants to make it all mid-gray...</p>

<p>Now when you meter direct light ( is that correct in english ?) with your light meter , you meter exactly the amount of light reaching the subject, which mostly gives you a much better reading, which you can dial in straight for exposure, since your light meter does not try to bring your subject back to "mid Gray" ( it did not even see your subject, just the available light, did'nt it ?) .</p>

<p>When I can I also prefer using a manual light meter , more so wth slide film which gives next to no margin for exposure error, and cannot be corrected after the shoot......</p>

<p>So nothing wrong with your camera, just a better method of metering used, i'd say......</p>

[chris_graamans]

<p>Greetings! If you want to make this work with the in camera light meter, use spot metering, rather than matrix. Half press the shutter with the very center of the frame on the subject (a subject's face in this case) to lock in the exposure (and focus, depending on how you programmed your camera). If lighting is even across the space, take those exposure settings, and use manual exposure - this way you only think about exposure once. If uneven, lock in on a face for each shot using Aperture priority.<br>

What you have at a graduation in a dark space, is a lot of black (robes, background) that get averaged in with the faces by matrix metering. This results in overexposed faces, medium gray robes. The same problem is common at concerts, theater performances. You have to tell the camera what part of the frame to expose for.</p>

<p>Hope that helps,<br>

Chris</p>

[pete_s.]

<p>I agree with C.P.M.<br>

It's just the difference between an incident meter that tells you how much light falls on a subject and a reflective meter that tells you how much light is reflected back.</p>

<p>BTW, when Nikon calibrates the exposure meter in the camera they do it using an incident meter :-)</p>

<p> </p>

[bebu_lamar]

<p>You have to do further testing before you know for sure you would need calibration. I found that although the built in meter on most cameras today are very good they are calibrated not as good as older cameras. It's a shame because the calibration is software based, they don't have to disassemble the camera at all.</p>

[rodeo_joe1]

<p>Personally I wouldn't waste my money having the camera checked over, unless it's under warranty. All that a service centre will do is to point the camera at a lightbox (or something similar) of known brightness and check that the readings are within spec. If you're lucky they'll check at more than one aperture with a calibrated lens. What they won't do is fix Nikon's flaky metering algorithms. </p>

<p>Stick with the external incident meter if it's giving you better results. I'm not sure about using the so-called "spot" metering built into the camera, since it meters too large an area to really count as a spot meter, especially if you have a wideangle lens fitted. Plus it introduces an element of guesswork in estimating what is a mid tone to point it at.</p>

<p>As an aside: I really can't see what's so difficult about getting a digital camera to give almost perfect exposures every time. Having a memory bank of 50,000 "reference pictures" to compare the metering pattern to is sheer lunacy. That's basically 49,999 chances for the camera to get it wrong! Whereas all the makers really have to do is ensure that the camera doesn't blow out any pixels - or at least not too many! How about a system whereby you could set a percentage limit on the area of "blown" pixels allowed. Say 0, 1, 2, 5% etc. Then the camera sets the exposure accordingly from its histogram and everybody's happy, since a little underexposure can nearly always be rescued in post.</p>

[blumesan]

<blockquote>

<p>How about a system whereby you could set a percentage limit on the area of "blown" pixels allowed. Say 0, 1, 2, 5% etc.</p>

</blockquote>

<p>I think that's a terrific idea. Especially if it could be programmed to compute from the raw (sensor captured) data, unlike the in- camera histogram which uses the converted jpeg data.</p>

 

<blockquote>

<p>Then the camera sets the exposure accordingly from its histogram and everybody's happy, since a little underexposure can nearly always be rescued in post.</p>

</blockquote>

<p>I have been led to believe that (if you save in raw format) a little overexposure is more easily rescued than a little underexposure.</p>

<p> </p>

[bebu_lamar]

<p>Quote: Rodeo Joe<br>

"Personally I wouldn't waste my money having the camera checked over, unless it's under warranty. All that a service centre will do is to point the camera at a lightbox (or something similar) of known brightness and check that the readings are within spec. If you're lucky they'll check at more than one aperture with a calibrated lens. What they won't do is fix Nikon's flaky metering algorithms. "<br>

When you said Nikon's flaky metering algorithms you meant the matrix metering system only? Because spot and center weighted metering systems are simple ones and should not be flaky. I found that Nikon's matrix metering system works well for slide and digital but very poor for negative film.</p>

[studio460]

<p>C.P.M. said:</p>

<blockquote>

<p>When metering from the camera, you measure reflected light which then gets calculated back by the camera to expose for "Mid Gray" (reflection of avarage cocasian human skin.).</p>

</blockquote>

<p>My apologies for the correction, but 18% gray is defined by the Zone System as Zone V, or "middle gray." Caucasian skin tone reflectance is commonly assigned to Zone VI, one-stop above middle-gray.</p>

[CvhKaar]

<blockquote>

<p>My apologies for the correction, but 18% gray is defined by the Zone System as Zone V, or "middle gray." Caucasian skin tone reflectance is commonly assigned to Zone VI, one-stop above middle-gray.</p>

</blockquote>

<p>Ralph, thx for straightenig this out,<br /> This was more about trieing to explain the effect of different metering methods in English ( which is already hard if your mothertongue is not english), but of course your're right in your comment.</p>

[rodeo_joe1]

<p>BeBu. Yes, I meant the ill-judged 3D colour matrix metering system that Nikon seem to have got so woefully wrong on some cameras.</p>

<p>"I have been led to believe that (if you save in raw format) a little overexposure is more easily rescued than a little underexposure." - RAW gives you a little over one stop of extra headroom. Any more than that and the highlight detail is gone for good, whereas if you stick to low ISO speeds and save at 14bit depth you can rescue about 4 stops of underexposure. I've never been sure where the advice to "expose to the right" came from, but IMHO it's outdated advice that never was based on sound science.</p>

[bebu_lamar]

<p>Joe. Actually for indoor events like the one the OP was photographing a hand held meter with incident reading would be much better especially if you don't want to bother yourself with multiple spot readings whether or not the in camera meter is correctly calibrated. Besides from the 3D matrix algorithm which is unpublished the spot and CW meter on most current Nikon DSLR's are not calibrated carefully at the factory. They are good that is they are very repeatable and do not change over time but if you pick up 2 cameras of the same model pointing at the same subject you will have 2 different readings may be about 1/2 to 2/3 stop often. So whatever standard they are calibrated to, they are not well calibrated.</p>