D7000 Matrix Metering

[simon_hickie1]

<p>Hi All. I have now read two D7000 reviews from UK based sources (ephotozine and Amateur Photographer) which have both observed that in contrasty scenes the metering tends to 'expose for the shadows' leading to washed out skies and potentially blown highlights. Real user experiences would be welcome, particularly in comparison with the D300 which I find to have excellent matrix metering. I'd just hate to have another D80 experience!</p>

[Jerry_]

<p>The matrix meter is good at measuring everything in the scene in front of the lens: if you want to control more of the exposure, you have to put the camera in large or small spot meter mode. It's the same on almost every Nikon body (the D3, the D700, the D300,) so just be prepared to think for the camera now and then.</p>

<p> </p>

[Matt Laur]

<p>Yup, what Jerry said. Nikon's DMR-2000 Photographer Mind Reading Upgrade Module is still on backorder after all these years. In contrasty scenes, you need to give the camera some guidance so that it knows what <em>you</em> find to be important. A lot of bright sky scenes involve back-lit people, and I can see how they might err on the side of preserving those details instead of darkly silhouetting them against a perfectly exposed sky.</p>

[lachaine]

<p>If it's anything like the recent Nikon model I have (not a D7000), in higher contrast scenes that exceed normal photographic dynamic range, matrix metering seems to bias in favour of properly exposing whatever is under the actual focus point, wherever that was at the time of taking the picture. So, if that part of the picture is darker, you end up with an overall lighter exposure, and vice versa. You can experiment with this yourself just by using Nikon's ViewNX 2 or Capture NX2 so you can see where the focus point was.</p>

<p>Once you know that, it makes using matrix a little more predictable, although personally, I find that it ends up not a heck of a lot different than using centerweighted metering and single point focusing on the subject itself with AE-Lock on when the shutter is half-pressed.</p>

[john_tran14]

<blockquote>

<p> the metering tends to 'expose for the shadows' leading to washed out skies and potentially blown highlights</p>

</blockquote>

<p>Many times, we heard comments similar (or opposite) to that. Maybe you should suggest Nikon to make "Hi Matrix", and "SDow Matrix" metering modes selectable for those who favor the Hilite, or Shadow</p>

[ShunCheung]

<p>I just tested this situation between my D300 and D7000. They happened to give me identical metering: @ ISO 200, 1/125 sec, and f8. Please keep in mind that this is merely one data point. My experience with the D7000 is very limited so far.</p>

<p>In high-contrast environments, I still by far perfer spot metering, but with histograms and blinking highlights from digital, there are now alternatives.</p><div></div>

[Wouter Willemse]

<p>Maybe it's me, but I never quite got why people got so worked up over the D80 (owned one too); it may have been unexpectedly metering for shadows, but in a quite predictive fashion, at least in my experience. So it was nearly always kind of clear when to switch to spot metering or when to apply some exposure compensation.<br>

No matter what, it's still a matter of recognising the potential problem of a scene, and knowing how to deal with it given the camera. Once you know how your matrix metering reacts, it's not that hard to deal with it.</p>

[simon_hickie1]

<p>The D80's matrix metering seemed to give undue prominence to what lay under the focusing point (be it shadow, highlight, mid tone or whatever) as I demonstrated in a thread about three years ago. The D300 does not do this and it's matrix metering is extraordinarily reliable. In over two years I've never felt the need to use spot or centre weighted metering and seldom need to use exposure compensation.</p>

<p>I'm very familiar with various exposure and metering techniques and in my pre Nikon slide film days typically relied on centre weighted metering with exposure locking in aperture priority mode. I got spoilt with the Nikon F70's matrix metering which was very reliable too (and only six segments!).</p>

<p>Interestingly, a friend borrowed my D80 for a short photography break last year and commented on how unpredictable the metering was. This year he borrowed my old D50 and found its matrix metering much more reliable and predictable (and he typically shoots slide film on a Canon A1 with prime lenses). </p>

[elliot1]

<p>Wouter, the problem I had with my D80 was that it was consistently inconsistent. The issue the OP states is easy to work around if the meter gives consistent results.</p>

[Wouter Willemse]

<p>As said, my experience with my D80 is different, but maybe I was lucky in ways. And I use spot metering quite frequently anyway, so maybe I avoided the nastier situations accidentally... can't tell anymore since I do not have the camera anymore.</p>

[photo5]

<p>Shun, that's not a good example. That image can be well exposed because of the even amount highlight and shadow.<br>

Here is a JPG straight from my Nikon D80 that I took in November, 2006, while on vacation in Japan. It is just a casual street photo I took near the home where I was staying. I saw this scene and set the camera to shoot at -1 EV. That is, I told the camera to expose the image at <strong>a full stop below the setting that the meter was telling the camera to use.</strong><br>

<img src="http://hull534.smugmug.com/photos/418925188_AVt4y-L.jpg" alt="" /><br>

As you can see, it completely blew out the highlights to smithereens. I tried to recover them later from the RAW file, and they are not recoverable. This is an extreme example, I know, but it is illustrative of the terrible meter the D80 had compared to the D200 (which tended to underexpose when in doubt, instead of overexpose as the D80 has done here), and the D300 which can correctly handle a scene like this.<br>

Hearing that the D7000 also overexposes makes me glad I didn't buy one. As the D7000 is an amateur camera unlike the D200 or D300 which are semi-pro, the D7000 will tend to make images with more contrast and pop with default settings, thus blowing out highlights as any point and shoot would do.</p>

[thomas_lozinski]

<p>Posting an example of overexposure on my d7000. I took about many shots of windsurfers overexposed like this before dialing down exposure. I'll try to post the histogram too, obviously overexposed. </p><div></div>

[thomas_lozinski]

<p>Here's the exif data for the windsurf shot. It was extremely bright conditions but continually did this until I dialed in some negative exposure compensation.</p><div></div>

[thomas_lozinski]

<p>Maybe I posted too soon. What is gain control? Usually when I have +1 EV it is indicated up top in the bridge info next to the metering picture. This has gain control of 1 but I don't think the camera had any EV dialed in. </p>

[ShunCheung]

<p><a href="../photodb/user?user_id=5214737">Thomas Lozinski</a>, the histogram you posted is exactly what the camera should do following the "expose to the right" principle while shooting RAW. It captures the maximum amount of information possible without blowing out the highlights (or maybe blowing out just a tiny bit of highlights). In post processing, it is very easy to correct it.</p>

<p>If you shoot JPEG only, that histogram will be problematic.</p>

<p>As I said, my earlier sample is merely one data point. Since now I have both the D7000 and D300, I can compare their meters under various lighting conditions.</p>

[photo5]

<p>Shun, we all look forward to your test results from the D7000.</p>

[leighb]

<blockquote>

<p>What is gain control?</p>

</blockquote>

<p>As commonly used in the electronics industry, "gain" indicates an amplification factor relative to nominal, which is 1.</p>

<p>Using this reasoning, a 1-stop over-exposure would be a gain of 2, while a 1-stop underexposure would be a gain of 0.5.</p>

<p>I can't guarantee that's the correct interpretation of the EXIF data, but considering the 'EV' notation is not attached to the parameter, I believe this is right.</p>

<p>- Leigh</p>

<p> </p>

[dan_south]

<blockquote>

<p>[The D80's matrix meter] completely blew out the highlights to smithereens. I tried to recover them later from the RAW file, and they are not recoverable. This is an extreme example, I know, but it is illustrative of the terrible meter the D80 had.</p>

</blockquote>

<p>Dave, can you tell from viewing the photo in Capture NX or View NX where the focus point was? The foreground looks properly exposed to me. The stones look just as I would imagine them looking if I were standing in the scene - not too bright, not too dark. I would expect a metering system to favor the foreground over the background.</p>

[simon_hickie1]

<p>My issue is that the focus point should be irrelevant. If, for example, one was employing hyperfocal focusing techniques in a landscape shot, it may be that the focus point happens to be in an area that is particularly light or dark. If I want to use spot metering, I use spot metering. I would expect matrix metering to evaluate the overall scene and deliver a full range of tones rather than favour one particular end of the histogram - dynamic range issues permitting. The D300 seems to do this reliably and consistently. Comparisons between the D7000 and the new Pentax K-5 will prove interesting!</p>

[elliot1]

<p>@ Thomas. It appears you are shooting with the Sigma 150-500mm. I this correct? If so, the lens could be the problem. I had the Sigma 50-500mm briefly and used it primarily to shoot surfing. I found it to overexpose by at least a stop or more at the beach and give a lot less contrast compared to my Nikon 70-300mm VR.</p>

<p>I think it is important to understand that if a meter is consistent (meaning it always tends to overexpose or underexpose), exposure compensation can easily correct the problem. When the meter is inconsistent (meaning sometimes it will overexpose and sometimes it will not under similar or the same shooting conditions), it is difficult to control.</p>

[simon_hickie1]

<p>Hi Thomas. It also looks like your surfing shot might be suffering from 'veiling flare'. If the meter was exposing correctly I'd also expect the histogram to be a little further to the left. The veiling flare, if that's what it is, also means that we are not seeing any blacks from the wetsuits - also reflected in the histogram. </p>

[thomas_lozinski]

<p>Even if it was veiling flare, the meter would see all the light before the exposure. I looked up the gain control and it has to do with the chosen ISO compared with base ISO not exposure compensation. The ISO 100 and 200 had zero gain, 400 and 800 have 1 gain 1600 to 12800 have gain 2.</p>

[zack_zoll]

<p>Thomas: Not necessarily so. If the camera could see flare/ghosting before it happened and expose accordingly, we wouldn't get much flare or ghosting! Most likely the meter is reading dynamically (like our eyes, with no time factor), and expects a very small amount of flare. Once the exposure is made time is now a factor, and over the duration of your exposure your camera may accumulate more flare.</p>

<p>Another issue is that when the lens relays exposure data to the camera body, it doesn't say, "Hey camera, I'm gonna' flare!"</p>

<p>I would advise setting the function button on the D7000 to spot meter. I've done this on every camera I've owned that allowed it, and it has helped my exposure immensely.</p>