Why am I always underexposing with my D3100

[WayneF]

Did it do this before the overhaul?

But just saying, hopefully without offense, if we accept your opinion that "all pictures are underexposed", then there would seem to be a problem, it ought not do that. But that is not in my experience, and my guess is that seeing it might be helpful too. Not sure what it might show or help, but without a picture, no one can have any clue what we're discussing. And I'm not sure any one picture would be indicative of a general problem, so maybe 2 or 3 in different settings? (looking for the trend you suggest). With some details about settings and modes of course. If all pictures do this, then it should be easy to snap off about three of them, not of anything, the dog, the front door, a flower, the tree in the front yard, the lawnmower, just a picture, anything, not of anything, just some different situations showing typical exposures.

I don't have a D3100, but suspect it is like all other models. My guess is still that the issue must be more about generalities of proper exposure, which is always pretty hard to define critically. Everything always varies. Reflective meters are influenced by scene colors of course. Matrix metering going out of its way to look for unusual spots of high contrast, we never know what it does. Camera settings of Contrast and White Balance are large data shifts which affect extremes, even in Raw software...

In the raw editor, it would be instructive to pick any typical full range image, and run the White Balance slider from one end to the other while watching the histogram. At Tungsten end, blue is boosted substantially, and can clip. At Cloudy end, red is boosted substantially and can clip. Just how it is, all things vary. You have not mentioned clipping though, seemingly more concerned about the black end, but blacks hopefully ought to be black (left end of histogram). But we really don't know what we are discussing other than an "all images" general case, but I doubt the the rest of us can comprehend that.

Try to show us?

 

[aroy]

After reading this post, I tried various modes with my D3300. Here are the findings
. With Matrix metering, the exposure is totally scene dependent. Trees in bright sunlight tend to be fine. Evening shots where the sky is a part of the scene are under exposed, Indoors with flash under exposed.
. With centre weighed metering, the area metered is perfectly exposed. Indoors with flash the histogram shows 1.5 stop under exposed, but the scene looks perfectly exposed to the eye.
. Spot metering exposes perfectly in the centre. So you can meter the area you want, lock exposure and then shift and reframe for the shot.

My conclusion, is that if there is even a small area of bright light in the area of metering, then the body accounts for it, which leads to the rest of the scene looking under exposed. That is in tune with the philosophy that no pixels should be blown. When we examine the image, we want the main area to be perfectly exposed, which means that in case there is a bright area in the periphery, we do not mind if it is blown. The camera really does not know that and tries to accommodate the maximum brightness. Hence the best recourse is to meter for that area which you want properly exposed, and forget the bright spots.

 

[egosbar]

is exposure compensation set to zero

 

[Revet]

Yes, no compensation was being used

 

[Revet]

Yes, I did some experiments I that is pretty much what I found also. I will post some pictures in another thread.

 

[Revet]

First shot - Manual mode, Matrix metering, Iso 100, 1/25, f/9 (needle in middle saying it is a correct exposure)

Second shot - Same settings but used center-weighted metering.

Third shot - Aperture Priority - Iso 100, 1/15, f/9, matrix metering

Forth shot - Aperture Priority - Iso 100, 1/15, f/9, center weighted.

It does appear the using center-weighted metering gives me a half-stop increase in exposure over matrix (many other shots taken and this was pretty consistent). There is a little clipping in the 4th shot but I did shoot more of the driveway. It also appears that Aperture Priority did a better job in Matrix mode (shot 3) than Manual in Matrix mode (shot 1). Again I had taken many shots and it seems as though if I am in manual mode, center weighted metering gives me more shots with a correct exposure. But if I am in Aperture priority, Matrix mode seemed to be more consistently better. However, this was one scene shot at one time of day.

So, the lesson I learned here is:
1) nothing is perfect in life especially a camera meter
2) take lots of shots and know your camera!!
3) If you can, take a test shot and use exposure compensation
4) If all else fails, as long as your in the ballpark, Lightroom can fix it!!

 

[WayneF]

It also appears that Aperture Priority did a better job in Matrix mode (shot 3) than Manual in Matrix mode (shot 1).

You attribute the #1 / #2 shots to Manual Matrix vs automated Matrix, instead of to the presence of black car added in the center of #2?

Reflected meters : Large black presences in the metering area are expected to cause increased exposure. Not sure it is a biggie here, it was of course fairly dark before too.

Camera Aperture mode simply centers the meter automatically. How would simply centering the meter yourself in Matrix differ from the camera centering the meter in Matrix? It is not really about who centers the meter. It is about what is in front of the meter.

I am just teasing, but yes, it is a difficult exercise. But more odd things probably happen in Matrix than in Center metering.

In a flat scene, there really won't be much difference in matrix and center. Center is not necessarily always brighter, it depends only on what is metered in the view. Reflective meters are a difficult case, which need to be understood for their own characteristics.

But if high contrast (which this is), there can be differences. Center meters more in the center, and plays down anything abound the edges. Matrix area is influenced by the focus point, but it looks everywhere for excessively bright spots, and backs off a little for them. For example, 3 & 4.... the scenes are NOT the same scene, but center metering is overexposed and is clipping at the bright edges (edges were not much considered). But matrix did not even see the outer bright area, as it was not included in its view. It probably would have also backed off if it saw the bright edges. Such comparisons are pretty difficult. The first step is to insure all scenes are the same scene.

 

[Revet]

Nuts Wayne, When did my neighbors car drive in???? Then she had the nerve to get out of her car and enter my photo!!!! She is lucky I like her!! LOL Plus, I think a little lens creep occurred in shot 4. The best laid plans right!!!

Oh well I tried. Overall though, I have been shooting primarily in Manual mode for about 6 months and it seems to under-expose the images by a half to a full stop. If I am not rushed or if I am going to be taking many shots, I take a practice shot and use exposure compensation. I do think it is important to get a good feel for all settings on the camera so I think I will shoot in Aperture priority mode for a bit and also try center-weighted for a while also to see if I will get more consistent "on the money" exposures.

 

[WayneF]

I was impressed with your pictures. The scene possibly might not qualify for a National Geographic cover, but it was a difficult high contrast scene, and to me, the meter did an admirable job of capturing it. It comes out as a pretty picture, full range. The light seems great, I see no fault. In fact, I thought it was excellent, and certainly I don't see any inherent exposure defect.

But for critical comparisons, we do have to take extra steps to make sure the scenes are exactly the same (a careful tripod, and don't touch the lens zoom), else we are discussing other things than what we think. I do think that effort might have been better this time.

But of course, the bigger issue is normally due to reflected meters, just about how light reflects from different scenes. A predominately dark-colored scene (which does not reflect very well) will be overexposed. A predominately bright-colored scene (which reflects very well) will be underexposed. Just how meters work.

For example, direct flash will typically see a lot of dark background, beyond flash range, and we can expect frequent overexposure of foreground.
Or bounce flash indoors often sees a lot of light colored walls, tending to underexposure.
It is really about what the meter sees, and not about any accuracy problem.

So I'm thinking the real idea there actually was, 1) the assumption that the meter ought to always be correct, and 2) then which metering mode can better achieve this always accurate result? But the fallacy is, reflective meters are simply not always correct, simply not how it works. When we actually realize this, then we are way ahead, and then have a good chance to work around it.

But if we keep waiting for the camera meter to finally always get it right, it ain't gonna happen.

An incident meter can typically do that (a lot better), but in the case of this picture, how do we do that? In which light we we stand and meter it? Sun or shade? Wherever we stand, it will try to put that light around midrange. So again, it is usually up to us to figure out. Metering is a skill we develop. Point&Shoot is a misnomer.

For those who may not have seen it, I would offer How light meters work as a fast example of how reflected meters work. It shows obvious evidence which we simply have to acknowledge exists. This is simply how it works, every time.

And at the bottom of that page is a Kodak article about using light meters Accurate Exposure with Your Meter which says the same thing:

"A reflected-light meter reading is influenced by both how much light there is in the scene and how reflective the subject is. The meter will indicate less exposure for a subject that reflects little light, even if the two subject are in the same scene and in the same light. Because reflected-light meters are designed to make all subjects appear average in brightness, the brightness equivalent to medium gray, they suggest camera settings that will overexpose (make too light) very dark subjects and underexpose (make too dark) very light subjects."
​

Nothing new, this has always been true, since we invented reflective light meters. I know it is a lot of preaching, but the first absolutely necessary step is to understand and believe this is really how it works. The camera does provide the Exposure Compensation menu as the tool to to help work around this fact of life.

My own preference is Center Metering (easier to understand and predict), and I mostly use camera Manual mode only for indoor flash (when no one centers the meter), and I normally use camera A mode for outdoor snapshots. Because, the only advantage I can see for otherwise using camera Manual, and centering the meter manually myself, is that Compensation can be easily done that way, by not quite centering the meter. That seems a preference, and is easy, and a fine goal. But if we are always going to center the meter, it seems convenient to just let the automation do it.

 

[aroy]

In a high contrast scene it is always difficult to decide what and where to meter.
. If the main subject is in relative shadow, then metering on it will over expose the brighter areas. If exposing for bright areas, then the main subject is dark.
. If the main subject is in bright light then metering on it will under expose the rest, if exposing for the rest the main subject will get blown.

So you have to decide where to expose depending on what you want to see and what you do not mind blowing.

Modern sensors have helped a lot in resolving this dilemma. If the DR is within the sensor's range, just meter for the brightest and then recover the shadows in post processing. That is what I do normally with my D3300, unless the main subject is -5EV or less than the brightest part, in that case I let the brightest parts blow as they may not be necessary ( for example sky in back ground when I am shooting flowers).

If you want every thing properly exposed; from the darkest to the brightest; then HDR is your only recourse. Of course if in future we get sensors with a DR of 20EV and 24bit RAW most of the problems will be resolved. This is not an impossibility, as we have advanced fro a DR of 8EV to 14EV in a decade.

 

[kevy73]

I will swap between matrix and spot depending on the scene and what I am trying to achieve, but do spend most of my time on matrix. However when on matrix metering, it is essential you understand how the camera's meter works and it all revolves around neutral grey...

If you think about a bright sunlit scene, a bride - in white - your matrix metering mode is going to want to take the exposure DOWN to neutral grey. Same as if you had a dark skinned person in a dark suit in a dimly lit room, the matrix meter is going to want to bring the image UP to neutral grey.

In both cases the image you take - if you trust the meter will be either darker or brighter than you wanted it to be and wouldn't represent the actual light and scene you had in front of you.

So. You need to compensate. In the first example, the bright scene, bride in white, you actually need to over expose (according to the meter) the image to make it correct and with the dark scene, you need to under expose the image (according to the meter) to make it true to what you see.

How much you under or over expose all depends on how bright or how dark and it is something you learn and get used to.

Jerry Ghionis explains this a lot better than me I am sure if you can find one of his videos when he discusses exposure.

 

[Revet]

I guess the bottom line is that when I take a photo, the LCD display is just not large enough or it is too sunny out for me to tell how the exposure is. The histogram helps but sometimes the better photos are pushed to the right or left. It is always easy to correct in Lightroom and I can do it globally if a took a lot of shots. I understand what everyone has brought to light (no pun intended!!!) but I can tell you this, if I shoot in manual and I use matrix, 90 % of my shots (indoors without flash and outdoors in all conditions) are pushed to the left on the histogram. Searching on Google, it seems to happen to many other people too. I will see what my new D7100 does when I get it.

I did shoot some more photo's using center weighted and it does appear to get the exposure right more often than matrix. I wonder why I even started this post since I like non-human subject pictures slightly under-exposed anyway (the color's appear more appealing to my eye). I guess I just like things to work how they should 100% of the time!! A major character defect of mine that I recognize and work on every day!!

 

[WayneF]

I guess the bottom line is that when I take a photo, the LCD display is just not large enough or it is too sunny out for me to tell how the exposure is.

Sun is something else, but don't overlook that you can easily zoom in (and scroll around) on the rear LCD picture, to see it as large as you want.
Page 107 D3100 manual.

 

[Revet]

Very true, I use the zoom but not enough. I'll give it a try next time I want to check exposure and focus.

 

[aroy]

I guess the bottom line is that when I take a photo, the LCD display is just not large enough or it is too sunny out for me to tell how the exposure is. The histogram helps but sometimes the better photos are pushed to the right or left. It is always easy to correct in Lightroom and I can do it globally if a took a lot of shots. I understand what everyone has brought to light (no pun intended!!!) but I can tell you this, if I shoot in manual and I use matrix, 90 % of my shots (indoors without flash and outdoors in all conditions) are pushed to the left on the histogram. Searching on Google, it seems to happen to many other people too. I will see what my new D7100 does when I get it.

I did shoot some more photo's using center weighted and it does appear to get the exposure right more often than matrix. I wonder why I even started this post since I like non-human subject pictures slightly under-exposed anyway (the color's appear more appealing to my eye). I guess I just like things to work how they should 100% of the time!! A major character defect of mine that I recognize and work on every day!!

Matrix metering evaluates the light across the frame, and then decides; based on its data base; what is the best exposure. I think that the algorithm is a bit conservative not wanting to blow any highlight immaterial of its impact on the photo. On a uniformly lit scene, as in a landscape shot, it works fine most of the time, but if you have a lot of shadows and bright areas, then it underexposes.

Centre weighed exposure metering gives maximum weightage to the centre of the frame. So if the object you are concentrating is in the centre and is in shadow, it will expose the object correctly. The rest of the area may be over exposed. Similarly if the object is in bright light, the rest of the frame will be under exposed.

 

[Revet]

I went to the Zoo yesterday and shot about 100 photos. I used center weighted for all of them. I only had to adjust 2-3 in Lightroom for exposure. This was definitely what I was looking for!!