Overexposure vs Underexposure

[ronlane]

Okay so this is a multi-part question/discussion. From some of the discussions that I have read, it seems that most photographers would rather overexpose a little than underexpose. It is my understanding that this is because it is easier to process the shadows with this case.

Is this correct?

The second part of this discussion is about LR post processing. Is the most of your processing involve lowering the exposure?

It seems like in my processing, that I hardly ever have to lower the exposure, most of the time I have to raise it 0.36 (roughly 1/3) or more at times.

The third part would be this. If you got what the camera says is proper exposure in the camera, how often do you have to adjust the exposure in post? Is it typically increase or decrease?

 

[Derrel]

I think the over or under issue depends on a couple things. First off--the "new-sensor" Nikons have very wide dynamic range that they can bridge, and they also have amazing highlight recovery ability in their RAW files. So...with those kinds of cameras (D3-series, D700, D800, D7000 and other new APS-C Exmor-generation sensors) I think in RAW capture mode, the best thing to do is to overexpose a bit, and then "pull" the highlights back down in post processing.

With older cameras, like my Nikon D2x or Canon 5D classic, the dynamic range the sensor can handle is significantly, noticeably much LESS. With those cameras, and with cameras of that older era, blowing highlights was a terrible No-No, and often resulted in images that could not be saved in blown highlight areas.

I've argued with a few people who dismiss this point of view, but I own three d-slrs, which I have used within the last year, to greater or lesser degree: a Nikon D2x, a Canon 5D, and a Nikon D3x. The best way to expose with these three cameras is as described above; over-exposure with the D2x or 5D is often disastrous; with the D3x, the sensor performance is so amazing that overexposure is not an issue. If you own a 5D and think it can handle the same DR as a newer-generation Nikon, you're full of ***+. Period.

If you consistently have to raise the exposure .36 EV or whatever, maybe it would be good to dial in + .3 EV on the exposure compensation, or if your camera allows for it, make an "exposure offset", or whatever your camera maker/model calls it. THis will calibrate, or offset, the meter, so that a 0.0 =/- match needle reading is already biased the .33 EV it needs to be.

 

[SCraig]

Paraphrased from Here. Read that article for more detail.

A 12-bit image is capable of recording 4,096 discrete values. Since each f-stop change allows 1/2 the light of the previous stop, and if we assume a dynamic range of 5 f-stops, the brightest f/stop will contain 2,048 discrete tones, the second f/stop half that or 1,024 discrete tones, the third half that or 512 discrete tones, the fourth half that or 256 tones, and the fifth f/stop half the previous or only 128 discrete tones.

So in the brightest areas of the shot you have 2,048 discrete tones of "White" whereas in the shadows you only have 128 discrete tones. If you limit the dark parts of the image and get the light parts right where they are on the verge of blowing out by effectively shifting things to the right as much as possible, then you will have more discrete tones in the shadows than if you had allowed them to go all the way down to "Black".

 

[ronlane]

I think the over or under issue depends on a couple things. First off--the "new-sensor" Nikons have very wide dynamic range that they can bridge, and they also have amazing highlight recovery ability in their RAW files. So...with those kinds of cameras (D3-series, D700, D800, D7000 and other new APS-C Exmor-generation sensors) I think in RAW capture mode, the best thing to do is to overexpose a bit, and then "pull" the highlights back down in post processing.

With older cameras, like my Nikon D2x or Canon 5D classic, the dynamic range the sensor can handle is significantly, noticeably much LESS. With those cameras, and with cameras of that older era, blowing highlights was a terrible No-No, and often resulted in images that could not be saved in blown highlight areas.

I've argued with a few people who dismiss this point of view, but I own three d-slrs, which I have used within the last year, to greater or lesser degree: a Nikon D2x, a Canon 5D, and a Nikon D3x. The best way to expose with these three cameras is as described above; over-exposure with the D2x or 5D is often disastrous; with the D3x, the sensor performance is so amazing that overexposure is not an issue. If you own a 5D and think it can handle the same DR as a newer-generation Nikon, you're full of ***+. Period.

If you consistently have to raise the exposure .36 EV or whatever, maybe it would be good to dial in + .3 EV on the exposure compensation, or if your camera allows for it, make an "exposure offset", or whatever your camera maker/model calls it. THis will calibrate, or offset, the meter, so that a 0.0 =/- match needle reading is already biased the .33 EV it needs to be.

Thanks Derrel, I'm still not convinced it is the camera that needs to be calibrated. It may be the operator. I think what I need to do is get together with a local photographer and go on a shoot to compare my SOOC and post to theirs.

 

[ronlane]

Thanks Scott. I'm going to have to chew on that a little bit to understand it. I can admit that it is WAY over my head.

 

[hirejn]

If you're looking to get this settled, you won't. Exposure is as subjective as art. I personally practice shooting to the right when possible, meaning I place the highlights as close to the right edge of the histogram as possible without clipping, when there is a highlight. I do this by spot metering the highlight and opening up to the maximum range my camera can capture, which is about 2.8 stops. The reason people STTR is because the highlights in digital files hold more detail than shadows, so we want to push the exposure to where there's more to play with.

Most of the time I don't adjust exposure in post; I simply move the white and black clipping points. Raising the exposure in post creates noise, so I want to avoid that. This isn't a problem because I'm rarely off by more than 1/3 stop from where I want to be. I consider 1/3 stop inconsequential, but an adjustment of 2/3 stop or more is something I consider a repair because that can easily be done in capture. Perfect exposure is a matter of understanding light, metering and using the tools available correctly. In most situations I can get perfect creative exposure, with full detail in highlights and shadows, in one click, and so should any pro. Perfect exposure is done and taught throughout the industry every day.

I rarely use what the "camera says" is correct exposure. The camera is capable of telling you only one exposure, and that is the correct exposure to render the subject medium gray. Sometimes it uses algorithms to appear to be a bit more creative, but I'd rather not rely on the camera to make those decisions. When possible, I use a hand-held meter, either in incident mode or spot mode off of a highlight, to get the exact exposure and then decide if I want that or want to make creative decisions. I also use the Sekonic L-758's averaging functions to find the dynamic range of the scene to be sure I can capture it in one shot. For this to work you must use the meter and Sekonic software to calibrate the meter to your camera's dynamic range.

 

[Derrel]

Thanks Derrel, I'm still not convinced it is the camera that needs to be calibrated. It may be the operator. I think what I need to do is get together with a local photographer and go on a shoot to compare my SOOC and post to theirs.

Well, it could be, Ron! I wonder too if the "plus" exposure you need to add is just the factory's default "highlight protection"? 1/3 of an EV is not that far off. I am assuming you are saying the images typically need + .36 EV based on so-called proper histogram of the files in Lightroom or PS? Or are you judging this by how the images appear on a specific monitor?

1/3 of a stop is not a whole lot; it "might" be as I said, a built-in factory bias, a design decision, based on an average scene, or a specific Tone Curve that the camera is set to. When reading a camera-generated JPEG file, or in the field and looking at the histogram on the back of the camera, the Blinkies, and the RGB channel indicators, and so on are ALL made up from the embedded JPEG, which DOES take into account the Tone Curve (the degree of contrast the camera's processing engine applies to the data) the camera is set to. So...it's "possible" that a number of camera set-up factors could be affecting the overall light metering the camera is giving you.

Exposure metering, and exposure setting needs to take into account the Tone Curve's impact on the way the capture data is going to be made into a photo. On the SOOC JPEG files, I think having a bias toward having them come out a bit DARKER, is a good general compromise for JPEG images.

 

[KmH]

Thanks Scott. I'm going to have to chew on that a little bit to understand it. I can admit that it is WAY over my head.

What it said is that fully 1/2 of the image information is in the brightest stop of exposure. The concept is the basis of ETTR - Expose To The Right.
Each dimmer stop of exposure has 1/2 of the luminosity info that remains, which is why underexposure makes image noise more visible.

Capture every photon you can.

Note that ETTR only works if the dynamic range in the scene can be captured by the image sensor in your camera.
Indeed, knowing the dynamic range your camera is capable of capturing is a critical specification you need to be aware of.

 

[pixmedic]

I usually shoot to the right.
sometimes though, it can be inconveniencing having to turn a few times to get something that is on my left side.
I kinda feel when that happens, it would have just been better to shoot to the left.
but, rules are rules.

 

[480sparky]

I expose to the right only because my right leg is shorter than my left.

 

[kundalini]

I tend to hang to the left, but then I tend to ETTR when possible in camera. In post, will bring down Exposure and bump up Fill and Blacks to taste.

 

[ronlane]

Thank you all. I will continue to read and digest this more as I practice. Part of my problem may have been how I had changed the camera to back button focusing. When I did this, I put in the wrong setting where I lost my exposure lock. I wasn't able to set the exposure for the correct portion of the composition. I have read the manual and changed this to where I have bbf and can exposure lock with the shutter button. We shall see how this does for me this weekend.

 

[weepete]

I find I'm almost allways slightly under exposing my shots and have to lighten them (somewhere in between 0.2 and 0.7 normally) in post. I frequently shoot in non ideal conditions but I'm not entirely sure whither I'm dailing in EC manually and then forgetting or whither I'm screwing up with the metering mode. I do know that my camera can produce properly exposed shots as there are times when it looks ok and I don't need to adjust so I'm pretty sure its me that's messing up a bit. The way I understand ETTR is basically your more likley to introduce noise by underexposing and pulling exposeure up than the opposite, so its best to ETTR as long as you don't blow out the highlights and plan on editing.

Most of my time in post is taken up with sharpening and tone curves or masking.

 

[Stevepwns]

I usually shoot to the right.
sometimes though, it can be inconveniencing having to turn a few times to get something that is on my left side.
I kinda feel when that happens, it would have just been better to shoot to the left.
but, rules are rules.

I tend to follow this set of ideas. But I am left handed and get confused. So I just hit the shutter and pray it works out.

 

[kundalini]

..... so its best to ETTR as long as you don't blow out the highlights and plan on editing.

That's why I prefer to set the LCD display to "Highlights" rather than Histogram. It must be my old eyes because I can't read a thing (text) within the length of my outstretched arms without reading glasses on. I do not wear glasses while shooting and often forget to bring a pair along. So with the LCD set to "Highlights" or "Blinkeys", I can tell immediately what is blown out in the frame. If a part is blown out and intregal to the image, I can make a quick setting adjustment so that part of the frame is still recoverable. I think of the Highlights vs Histogram much like an analogue clock vs a digital. If I walk past an analogue, I can tell you the time of day straight away. If I walk past a digital clock, I have to read the time, which may or may not tell me what I want to know as I cruise by.

Sorry, didn't mean to tilt the conversation.