MountainDave

I use -1/3 outside most of the time if the sun is shining. I have studied histograms at standard exposure. When they are off, which isn't often, they are always off in the overexposed direction. There are occasions when I'll purposely overexpose as well such as a dark subject against a bright background. Recently I watched a video by Tom Mangelsen who said he just leaves his camera at -1/3 and often goes lower. His theory on this is that 1. Nikons run "hot." and 2. preservation of detail is paramount. It's pretty easy to brighten a dark image but once a portion is blown out from overexposure, it's gone for good. He also emphasized the importance of checking histograms. I use Canons but I think they tend to run "hot" as well. I think everyone should at least experiment with exposure compensation in various situations.

CHG_CANON Loc: the Windy City

Experiment and individual experience with specific equipment is the key. I was testing something earlier today in Aperture Priority where +1 in EC was needed for a dark subject in a well-lit room. Turning toward the window without changing the camera settings had just a few blinkies in the image of the window. If I was concerned about that second shot, I might leave as-is as the image file is RAW. I might also lower the EC to +2/3.

The point is that to assume all models from a vendor perform the same is a generalization that should be confirmed by personal experience with a specific model as well as the image format. My own experience with a 5DIII is +1 or more is appropriate. I say that only about my photography needs for my specific body. In a past life, I used a D7200 and I remember Rockwell's recommendation was +1/3 EC by default. That suggestion worked well although it directly contradicts your generalization.

TheDman Loc: USA

He's fantastic, isn't he? Bet you're the only person who clicked through and watched it.

MountainDave

In the situation you describe, I would use + something as well. My comments were directed to outdoor shooting. I'll repeat, everyone should experiment and check histograms.

TriX Loc: Raleigh, NC

While ISO invariant camera’s allow you to “save” inadvertently underexposed shots in post, I contend intentionally underexposing at base ISO with invariant cameras is not good practice, and limits both the DR of the A/D and can cause unintended consequences in post. I’m aware that there are those that disagree, but comparing histograms of the same subject “properly” exposed with that of the subject underexposed and brought up in post will elucidate the differences.

BebuLamar

Still more exposure (without burning out the highlight) would result in lower noise.

WJB Loc: Salisbury, MD

I still "bracket" a lot of shots, when possible. Seems to work out for me.

BebuLamar

As for ISO invariance. I did a test with the camera set at f8, 1/8 sec at ISO 100 and and ISO 1600. Adjusting the exposure in ACR to +4 for the ISO 100 yields very similar results. However, if I set the exposure for f/8 and 1 sec for the ISO 100 shot it would have much less noise.

Weddingguy Loc: British Columbia - Canada

Yes . . . according to the following facts . . .

A major difference between the digital and film worlds is the grain and noise. Film photography has grains; digital photography has noise. While film has grains in the highlights, digital has grains in the shadows. This is one of the major aspects that determine how we expose an image in the digital world - ETTR.

ETTR (Expose to the Right ) vs ETTL (Expose to the Left)

ETTR refers to “Expose to the Right” and ETTL refers to “Expose to the Left. ” ETTR means to slightly overexpose the image and ETTL means to slightly underexpose the image.

ETTL was popular in the days of film photography for the simple reason that grains were prevalent in the highlights. Therefore, exposure was shifted slightly to the left (generally 1/4th to 1/3rd of a stop) to produce a cleaner image devoid of grains.

ETTR is a digital phenomenon because digital noise occurs in the shadows and not in the highlights. The images are now slightly overexposed to reduce, minimize, or eliminate the noise factor in the shadows.

fetzler Loc: North West PA

Expose correctly and adjust in post. It is important not to blow the highlights. It possible to extend the range a bit by making a pseudo HD image if you shoot raw. Make a set of tiffs in past with exposures of 0 ex and +/- 1 ev and then use HD software for the final image.

For those familiar with the Zone system. Remember you light meter puts everything in Zone 5 (18% refelectence). For snow +2-3 ev from the snow reading is required.

johntaylor333

ISO invariance is a very complicated issue and I recommend an excellent article - https://photographylife.com/iso-invariance-explained.

The point here is not to underexpose for the sake of it but in situations where you end up with a choice of correct exposures being very long shutter speeds or very high ISO, the idea is that, for an ISO invariant camera, you underexpose by multiple stops and end up with an acceptable shutter speed and a sufficiently low ISO. The noise when you edit in Lr, Ps or ACR is lower, for an ISO invariant camera, that it would have been for the "correct", higher, ISO.

PhotosBySteve

I agree that underexposing an image for the purpose of high saturation is not a good idea. If an image is underexposed, it is difficult to recover shadow detail, if desired, without excessive noise. That being said, I expose my images dependent on what type of dramatic effect I am attempting to achieve. I sometimes underexpose, to purposely eliminate some shadow details. I overexpose at times for varying reasons. In general, I usually expose to 18% gray and will quite often push it to the right, short of blown out highlights, to achieve more shadow or subject detail. It always depends on the situation. I use the histogram mainly to control my exposure levels within the bounds of the effects I am attempting to capture. I will also do this at times with post processing in mind when I take the shot, when I want a result that the camera cannot fully capture at the time.

selmslie Loc: Fernandina Beach, FL, USA

Habits carried over from using Kodachrome are a good starting point for digital. But Kodachrome highlights did not blow out quite as abruptly as they do with digital.

The single most important precaution you should take with digital is to not blow your highlights. Your camera's histogram helps you a little but the highlight warnings are even better because they not only tell you that you might be blowing some highlights but where in the image this is happening. What neither one can reliably tell you is which color channel is getting blown first (often green). When only one color is blown, recovering the highlights might result in some suspiciously tinted highlights (often magenta). Rather than wonder about that, play it safe.

To permanently bias your exposure by +/- 1/3 stop presumes that you are applying that compensation to the reflected light reading that the camera reads. While it may well be that your particular camera's reflected readings are off a little, it would take some careful testing with a gray card to confirm that. Unless you approach this in a scientific manner it may be more trouble than it's worth.

A 1/3 stop change in exposure is virtually indistinguishable from a 1/3 stop change using the Exposure slider during the raw conversion. The change in saturation is minimal and the effect on noise is undetectable. However, a 1/3 stop reduction in exposure might mean the difference between an irretrievably blown highlight and one than can be recovered.

There are some basic considerations that you may be aware of but they are worth mentioning:

1. Reflected readings are influenced by the reflective properties of the scene as well as by the light illuminating it. An incident reading is less biased.
2. In broad daylight, the sun is a constant light source. You should not need a meter of any kind. In fact, there are lots of situations where you can get away with not metering. See Exposure value
3. The sky, particularly during sunrise and sunset, behaves more like a light source rather than a reflector, especially if the sun or moon is included in the scene.
4. The most saturated parts of an image (after the raw conversion) occur in the middle zones, less than two stops above or below middle gray. That's not the middle gray that your camera sees but rather the midpoint between a JPEG value of 0 (maximum black) and 255 (maximum white) or 127 (the middle of Zone V in Adams's Zone system). A couple of zones darker and the colors look dull. A couple of zones lighter and the colors start to look washed out like pastels.

The primary goal of HDR is to bring the highlight colors down and/or the shadows colors up into the more colorful and saturated image zones by combining two or more shots that capture the extremes more faithfully. HDR is demonstrably more effective in scenes with wide dynamic ranges than highlight and shadow recovery in a single image.

But HDR is only necessary for a scene with an extreme dynamic range like sunrise and sunset. Most daylight scenarios have a narrower dynamic range for which highlight and shadow recovery in a single image is enough, especially at low ISO and with a camera that provides ISO invariance.

David Kay Loc: Arlington Heights IL

Photocraig

Yes, donmikes,
Since only the SUPER Wizards ever could manipulate Kodachrome at all and for that matter, any slide film, we needed to try stuff like that. With our PP options today and RAW capture, the closer we start with accurate exposure, the more room we have to process the file.

While there's much to learn in the digital world, there is some constructive unlearning to do, too.