How to expose according to highlights ?

[fredscal]

<p>Hello.</p>

<p>Here's a little question: is there a way to set my D700 for always auto-exposing according to the brightest highlight it senses, no matter where in the frame, or what proportion of the frame it covers ?</p>

<p>My first guess would be: no. But then I'm asking because you never know...</p>

<p> Thanks.</p>

[kari_oinonen]

<p>Yes, but it needs some help from you.<br /> Spot metering, aim the active metering spot to the brightest part of the subject and overexpose that by 2-3 steps depending your ISO setting. Maybe only 1,5 if you are at very high ISO.<br /> The problem is that sometimes the highest highlight is sooo bright that this way you are underexposing your - maybe - main target.<br /> Often it is better to let a minor portion to go over, it depends.</p>

[acedigital]

<p>If what you are referring to is the phrase/strategy "expose to the right" to capture the maximum range of stops when shooting in RAW mode. By exposing to "the right" (using your Histogram view on the rear LCD), this enables you to get all the shadow detail you might miss if you under-exposed the frame.</p>

<p>Then your photo can be edited using Lightroom, Camera RAW, etc to bring things back to a normal range while retaining these details. <br>

If you meant something else, I defer to others.</p>

<div></div>

[bebu_lamar]

<p>No unless you can hack the firmware. The camera actually knows how bright the brightest high light is and where it is in the frame it just doesn't set the exposure for that. </p>

[purplealien]

<p>I don't think so. It would be handy wouldn't it? </p>

<p>I usually work in matrix and refer to the blinking highlights display after the first frame or two. With my D90 I often find myself dialling in -0.3 or -0.7 exposure compensation to preserve the highlights. </p>

<p>Chris</p>

[rodeo_joe1]

<p>It ought to be the easiest thing in the world for the camera to do, but Nikon's engineers have obviously never heard of "expose to the right" metering, or if they have, they think they know better than a vast number of working photographers and can't be bothered to offer it.</p>

<p>Nikon's overcomplicated and unfathomable 3D color (sic) matrix metering has been criticised by many of us on this forum. You'd think Nikon might actually listen to its customers wants, but no. So we're stuck with taking test shots, looking at the blinkies or histogram and then applying compensation - which doesn't always work because flakey 3D crock matrix metering is overly sensitive to <em>exactly </em>where in the frame you place the focusing point. As if the focusing point should get preferential exposure. What moron thought of that wizard scheme?</p>

<p>Alternatively we can use spot metering, <em>guess</em> what the brightest part of the scene is and point the near uselessly oversized "spot" at it. Sensitometrically, the compensation from metering the brightest spot should be +2.5 stops. In practice it's closer to 3 stops, but the ISO speed setting should make absolutely no difference to this.</p>

<p>Another alternative is to abandon TTL metering altogether and carry around a handheld incident meter. This I find to be a near bullet-proof metering method giving a high first-shot hit rate that needs a fraction of the compensation of matrix metering - or none at all.</p>

<p>In short we have to spend a lot of unnecessary time to do what the camera should be able to measure and calculate in milliseconds. Well, thanks a lot for listening Nikon!</p>

[dave_wilson1]

I don't know about the D700 specifically, but I do know that I don't care for Nikon's fancy multi point evaluative metering or

whatever it's called. As a rule I use the basic average metering and have the AE-AF Lock reset to AE Lock only and

average meter the part of the frame that I want and then push the lock and reposition so that the AF still works. This is

obviusly in P mode or A mode which I use frequently. Sometimes I just use manual, but I've gotten better and faster at

just aiming around the average meter and locking to do what I want. If in general I'm not getting what I want, I adjust the exp comp

[sem_svizec]

<p>No. In excessive DR, Nikon's MM generally goes for the midtones, with some bias for areas under active AF points etc. <br>

ETTR would be implementable as an additional metering mode. But there are in fact a few implementation issues. The exposure sensor doesn't have enough pixels to be really useful for this. The main sensor, used for metering in LV, does have enough pixels, but dedicated fast and accurate processing of the histogram would be needed; notice the predominant lack of a live histogram. There'd be issues with red and blue highlights if they used just the luminance, not separate RGB peaks. The right edge of the histogram is often difficult to determine precisely, there's the issue of a threshold how big dots you want to ignore. For instance one may wants to ignore the sun or other small bright outliers, or not. <br>

Shooting NEF and using ADL Auto may help you when you don't have time for exposure adjustment in quickly changing light, because this tends to expose raw data up to a stop lower than default with wide scene DR but leave default exposure with moderate DR. On the other hand, ADL's instant processing may mislead your judgement of exposure from preview image and histogram. </p>

[richard_driscoll]

<p>Actually the Nikon matrix meter is almost certainly even worse than described above. In a very thorough and very old Nikon description of the meter in the original Nikon FA it says that parts of the scene which meter above EV 16 1/3 <em>are ignored or clipped in the evaluation</em>. Now EV 16 1/3 is well over the 'sunny 16' brightness level and the assumption must be that such parts of the image are themselves the light source.<br>

Now, if you look in your instruction booklet at the curve describing the program used in 'P' mode (you'll find it near the back, page 318 for the D7000), it says that values over EV 16 1/3 are reduced to EV 16 1/3. It therefore looks as if today's matrix meter is doing something similar. This in turn means that bright areas of sky are not evaluated properly and will almost certainly be overexposed. That might not matter too much were it not for the fact that the colour balance will likely be ruined by one but not all of the channels clipping.</p>

<p> </p>

[richard_driscoll]

<p>I find the facility to blink red, geen and blue highlights independantly (on models D7000 and up) very useful in bright daylight when the separate RGB histograms can be hard to see. The luminance histogram I find missleading since any one of the RGB channels can be blown without it showing up on the luminance histogram.</p>

 

[richard_driscoll]

<blockquote>

<p>Another alternative is to abandon TTL metering altogether and carry around a handheld incident meter. This I find to be a near bullet-proof metering method giving a high first-shot hit rate that needs a fraction of the compensation of matrix metering - or none at all.</p>

</blockquote>

<p>I used to use my Weston 5 with Invercone on slide film with excellent results - much better than my built-in meter, but with digital I find the ability to look at the histogram and try again if necessary makes me reluctant. Perhaps I should get the Weston out again!</p>

<p> </p>

[Andrew Garrard]

<p>RJ - I have the feeling that this has been bothering you? :-)<br />

<br />

(For what it's worth, I've just been out shooting bluebells with the sun shining on them against a shady background, and had the same problem. ETTR wouldn't just have been useful for conventional post-processing reasons, it was actually the exposure I wanted.)<br />

<br />

The D700 meter, while pretty good at metering small areas, is a little short on pixels for this. The D800's (and D4's) should be better, though the algorithm changes have generally made me need to fiddle the exposure far more on my D800 than on my D700. Given the amount of dynamic range available on a D800 at low ISO, I agree that it's annoying to have to second-guess the camera. There's no reason that either camera shouldn't have been able to calculate an ETTR exposure in live view mode, though. That it's not offered suggests to me that someone somewhere has a patent on it, though I'll never rule out the "large company not listening" argument.</p>

[rodeo_joe1]

<blockquote>

<p>"RJ - I have the feeling that this has been bothering you? :-)"</p>

</blockquote>

<p>Gee! Was it <strong>that </strong>obvious Andrew?</p>

<p>If only it was that obvious to Nikon's system designers.</p>

[blumesan]

<p>For my D700 & D800 I have been using the method suggested by Kari.</p>

<p>To be a bit more scientific about it I tested both cameras by shooting a blank white wall at a number of exposures above the metered value. Checking the images with the histogram in RawDigger one can accurately determine the exposure increment that results in clipping. Then choose a value that gives you as much headroom as you desire. I have found that the value is not particularly dependent on ISO setting.</p>

[ilkka_nissila]

The issue is that the highlights can constitute a very small area of the frame and can be exceedingly bright compared to

the rest of the scene.

[ian_gordon_bilson]

<p>Coming from film,I have always found the Matrix mode a load of crap. Why trust an algorythm ?<br>

As others have said,a separate incident reading is best : optimum exposure 90% of the time. But centre-weighted is good too. It is easy to preset exposure compensation ahead of time,if you know the likely outcome of the metered subject.</p>

 

[robert_k1]

<blockquote>

<p>Spot metering, aim the active metering spot to the brightest part of the subject and overexpose that by 2-3 steps depending your ISO setting.</p>

</blockquote>

<p>Agreed. That's how I shoot transparency film, and on dSLR when accurate exposure is critical. The spot metering area should be at where texture must be preserved, and NOT at specular highlight where no texture need to be preserved. The digital ETTR is nothing but a repeat of the same method based on histogram.</p>

<p>This kind of exposure setting works fine for static subjects, such as landscape or studio, etc. But only those with lightning speed fingers can apply it for sports or on the streets. Camera manufacturers could have easily solved the problem for all, but they don't. See my comment below.</p>

<p>BTW, OP's home page link is broken on the profile page.</p>

<p>BTW, this is not Nikon specific, shouldn't it be moved to another forum?</p>

[robert_k1]

<blockquote>

<p>It ought to be the easiest thing in the world for the camera to do, but Nikon's engineers have obviously never heard of "expose to the right" metering ...</p>

 

</blockquote>

<p>Nikon is not the only guilty party, as far as I know no manufacturer is offering this. Name one and I will switch immediately.</p>

<p>In another thread, I said:</p>

<p>http://www.photo.net/digital-darkroom-forum/00aqHZ</p>

 

<blockquote>

<p>OTOH, digital bodies could have made the technique much easier by offering a couple of automated exposure modes. Such as "Expose to preserve highlight details" (i.e. ETTR), and "Expose to preserve shadow details" (for situations when shadow is more important than highlight). A digital body has all the information to set the exposure this way, whereas a film body does not. I don't know of a digital body that offers such automated exposure modes.</p>

</blockquote>

<p>Having such exposure modes would also shut up those who find the jpeg based histograms to be inaccurate and useless.</p>

[clive_murray_white]

<p>I take a lot of pictures of marble, and know that it'll almost always confuse automatic metering, so for crucial exposures on D800 I go RAW, all manual in live view + histogram, even then there some funny anomalies, what often amuses me is the discrepancy between how playback looks in the camera and what comes out in LR4, usually too dark but fixable. For fun I click on Auto tone to see what the designers think is right - always far too bright with marble.</p>

<p>I simply accept that no camera really knows what it is taking a picture of and automatic systems have to be averages - if I was looking for improvements I'd be concentrating on getting live view and playback much closer to the truth than they are now.</p>

<p>What drives me mad is the ever increasing trend for all sort of bits of machinery to be programmed to tell me what I want, should do and is best for me</p>

[dave_wilson1]

In general, if you don't know what you're camera is telling you, there's little chance of getting a good exposure

consistently. So, a good plan to start is to use a Kodak or similar 18% grey card and using the spot meter mode test out

the exposures you need and learn what those combinations produce in the end product. Or, use incident light readings

from a hand meter. Some camera bags used to be middle grey so you could just put your bag in the subject light and

read the bag. Reading a very reflective surface is going to fool your meter, so read the bag or the card or use a hand

meter. Edit: I forgot to mention that you also need to understand the exposure latitude that your camera is capable of for the specific finished product, web page, print, mural etc.

[shicane_wilson]

<p>I think in the end you should just go by your instinct to get the right shot, adjust the lens accordingly and sometimes natural light gives the best images.</p>