> Absolute BS. You have not come across enough photographers. A computer is NEVER better than a human at evaluating light, unless you are looking at a uniform 18% grey card.

Could you please substantiate your claim with any references or information? I am curious because I suspect you may be conflating dynamic range of human vision vs. sensor's dynamic range. Just because sensor's dynamic range isn't as wide as human vision, it has no bearing on sophisticated algorithms in today's cameras that determine the optimal exposure. Modern cameras have ridiculous processing power, scene recognition, face detection, and all kinds of algorithms that determine the best exposure. Please correct me if I misunderstood.

I use a Leica M10 and the auto mode almost always nails the exposure (I use full 3D metering).

In my view, "Manual Mode" in cameras is akin to "Manual Transmission" in cars, almost all super cars have automatic transmission for obvious reasons.

To add to the others, I frequently photograph raves and dance competitions that are strongly (but quite randomly) backlit. I dial it in on full manual to get proper exposure of the face, and then the lights streaming around the performer can come and go and without varying my exposure by a significant amount.

While wandering convention halls getting pictures of cospayers or suiters, the ambient lighting is consistent: I set to full manual and don't have to worry about the occasional light fixture in the background causing sensor confusion.

When I get home, I can bulk-process sets of images as a starting point, because they all have the same curves, then decide what I want to accentuate with each photo.

Source: Photography as a strong hobby for 35 years. I usually take around 20-30k photos/year at conventions and events, and the results are widely appreciated.

By default the computer will assume that it's shooting an 18% grey card and expose the picture accordingly. There are a basically two knobs you can tune, to change this behavior.

One, you can decide what parts of the pictures are used in this calculation and how they are weighted. These are your spot, matrix, and middle measurement modes.

Two, you can use a creative program, e.g., night photography. That will treat bright highlights differently, AFAIK.

However, the world isn't really an 18% grey card and how you expose the picture depends on your creative intent. The most obvious examples are high-key and low-key photos. In fact, photographers often use 18% grey cards to determine the light conditions and then lock the exposure and white balance down and shoot manual. This way, they have a precise reading and no surprises. I only do that when I do portrait photography because skin tones and therefore white balance are crucial. In most other cases, I look at the histogram to see if the exposure is correct and use auto white balance.

You can always adjust white balance when you shoot RAW in post processing. However, you cannot fix the exposure if you're missing the data. That's why I use manual.

You’re forgetting the third knob you can tweak - the exposure compensation dial. That’s arguably the most important control for the metering system!

If you understand how your camera’s metering system works (area selected, 18 percent gray etc) you can often work with the Metering system and compensate for when you know it’s likely to get it wrong - for example it’s not uncommon to dial a +1 if shooting a bright snowy scene, as most metering systems will be overly aggressive in trying to preserve highlight detail in the snow.

Similarly, if shooting in darker conditions dial in negative compensation.

Metering is pretty sophisticated in modern cameras, the exposure comp dial let’s you work with the tool and augment it’s measurements with your own human input.

Exposure comp plus histogram reviews with metering on is a pretty effective way to work a lot of the time.

All in all, I agree with your assessments. Its good to note that what you are trying to do is read the amount of light hitting the target. This could always be done with a ambient light meter placed in front of your subject (Studio).

The issue was the convenience of seeing it through the camera. The first automatic in camera meters were limited to high-key or low-key and tried to just make the picture's average lighting match the 18% grey card, a portion of it (center weighted or spot). Highly colored subjects could throw this off as the meters weren't always linear to all wavelengths. As technology advanced, matrix meters used multiple points to try and guess tricky lighting conditions. As more elements to the matrix were added, identifying where the subject became possible.

However, they all are just trying to guess at the level of light hitting the subject of the photograph.

Again, you are incorrect if you meant that multiple samples have fixed the in-camera meter foobar.

I am assuming you are well acquainted with colour management, so I will just leave a brief suggestion. If you further need more pointers, let me know and I would love to elaborate.

Read about reflectivity and their influence on metering.

> Could you please substantiate your claim with any references or information?

If you are dealing with a scene that has uneven illumination, for example, a sunrise or sunset, your camera metering will completely go nuts. The same happens when you are shooting snow.

Read about metering modes as one of the child comment had mentioned.

Personally even for exceptionally uneven lit frames why drop into manual? Point the spot meter to the area you want exposed, hold hallway down and readjust shot. This is so much faster.

For places that are controlled for light and movement and field (tripod camera pointed at a sunset)... sure use manual.

Read about 5 degree vs 1 degree spot. Also read about reflectivity of different colours. Lastly your in camera spot accuracy depends on the focal length of the lens sitting in front of the sensor. You will also benefit from Ansel Adams’s zone system. Read the book if you have time. I think you will like it.

This is why for challenging exposures, we use a hand held meter.

Good questions.

Look into metering modes and exposure compensation. In short, he's right.

It depends a lot on whether the scene is changing or the light is changing.

The problem with auto-exposure is that where you have a fast-changing scene with slower-changing lighting - especially fast-changing background scenery with constant lighting, the computer will change exposure when it shouldn't. This is a scenario professional photographers probably find themselves in rather more often than amateurs - think concerts, stadium sports, etc.

When the lighting changes more frequently than the scene then auto exposure can be helpful.

In this sense full manual settings could be viewed as a longer lasting "exposure lock" type function in a way. It's not that the computer-assisted metering is not useful - more that you want to get the metering locked in for the actual subject then stop it from changing.

I've often wondered how cool it would be if Google or Apple made the software for a DLSR camera? Judging by the results they can pull out of tiny sensors and tiny lenses, I'd be excited to use one.