I know the theory that people espouse about +1 and -1 for white and black, respectively, (I learned this early on in a photography school) but I wonder how much it applies in some cases in today's photography and to what one wants to do with an individual image. I know that when I focus on the brightest spot for exposure metering using the back screen and live view and a back button approach doing a +1 would not work out well for me. Thus, I do a minus exposure as needed. I don't generally blow out highlights in my images. I'd like to see how the theory holds with all types of shooting techniques and how mirrorless shooting might be affected. I am not disciplined enough to attempt this type of scientific experiment.

It is amazing how exposing to the right is kind of like a religion or belief that is cult driven...

Here is a good explanation I think...

Quote: "Most of a particular camera's extra dynamic range is relegated to the shadow detail. Most all cameras allocate only a small portion of the dynamic range to the highlights, and most cameras give you roughly the same amount in the highlights, with minor variations. So a camera that has 11 stops dynamic range will give you 3 stops above middle gray, and 8 stops below it, while a camera with 14 stops dynamic range will very likely also give you 3 stops above middle gray and 11 stops beneath it. More dynamic range will give you more ability to pull up shadow detail, while highlight headroom depends on exposure.

This meter calibration has nothing to do with the ability to pull down nominally overexposed highlights from a raw file. A color camera will typically have three color channels, and each of these channels is exposed differently, with the red and blue color channels usually being underexposed relative to the green channel by a stop or more. These red and blue channels are amplified during raw processing, and if you happen to have a bright, intense colored object in the scene, the raw processing may very well blow out this color. It's quite typical for blue skies and red flowers. But if the raw channels were not overexposed, they can be recovered. How much you can recover depends on the camera white balance and the light source, the model of camera, and many camera or raw processor settings, such as saturation, contrast, and the camera profile used, such as Landscape, Standard, Flat, Portrait, etc. It is possible to measure or estimate this color channel headroom and use that to adjust exposure upwards in certain circumstances.

If you want more highlight headroom, all you have to do is to dial in some negative exposure compensation. I do it all of the time!"

https://www.dpreview.com/forums/post/60302908

Thanks for the kind feedback.

It's true people can go off and do their own thing--but only if they can get the
equipment and supplies they need. No company is going to make a camera
because you or I want it. It will get made only if millions of people want it.

This is a big change. I remember a UHH post from a woman who had a "4 x 5"
camera made for her -- including a photo of the (very beautiful) camera.

But most of today's digital cameas use big custom chips (ASICs): Canon calls
its DIGIC, Sony has BIONZ, and Nikon has EXPEED. The captial investment
required to create a new ASIC is tens of millions of dollars. Nobody's going
to make one unless they can sell thousands and thousands of cameras.

I own many cameras for which I can no longer buy film (either roll or sheet).
I cannot make film stock or film: it requiers large, specialized equipment with
very close tolearances. Fortunately, sales of B&W film are booming (though
you wouldn't knwo it from UHH).

But other then film and lenses, if necessary, it is possible to make everything else
one needs for traditional photography: camera, developer, stop, fixer, enlarger, and
paper. Moreover, mechanical and electro-mechanical cameras can be maintained
and repaired fairly easily.

Photography is inherently optical, but it's only eletrical, electronic or chemical if
you want it to be.

The price of using digital technology is total depencence on camera manufactuers,
printer manufactuers, computer manufacturers, and their supply chains. But another
part of the price is extreme complexity: believe me, firmware for an embedded system
is complex, and ASICs are microprocessors are unbelievably complex.

Currently, only three types of photographic (e.g, non-industrial, non-security) digital
cameras are being produced:
* DSLRs
* mirroless with a screen on the back
* mirrorless with a screen inside (EVF)

That's not a lot of choices. If people continue to be lead around by the nose by marketing,
the DSLR may go away. The only camearas with opical viewfinders anymore are DLSRs.
So if they go away, so do optical viewfinders.

We live in a time of reduced choices and diminishing expectations. This is covered up by
golly-gee-whiz computer gadgetry and half-baked automation and a paid cheering section.

If someone claimed to have a portrait-painting machine, people would laugh.
But an automatic portrait-taking camera seems possible to the some people:
AF, AE, AISO..... Yeah, you can take drivers license photos that way.

A photographer is an essential ingredient in photography. The more closely the photographer
controls the process, the better the result can be.

But what Joe Consumer wants is easy and convenient. And his Uncle Bob just
likes fancy gadgets. So that's the direction the industry is going.

Nobody now alive knows how to cast bronze as well as the Shang Dynasty did.
And in not too many years, nobody alive will know how to take a good fine
art photograph. And life will go...just as it would without good artists or good
musicians. But the US no longer leads the world in photography, as it
did in the 1920s though 1970s.

Americans think they are individualists...but they are now corporatists. They drink
Coke, eat Whoppers, smoke Camels...and think that it's all their idea. And now
the ads and paid influencers are telling them to shoot mirrorless...

Not sure why you would constantly over exposure but I guess it would depend on what you are shooting and the metering method you are using. I can do as Gene states and spot meter on the brightest area I wish to retain data on in a scene and comfortably expose this area by slightly over 2 stops without losing data. This is based on calibration of my camera to my Sekonic meter using their software. But the key to me is that you are metering on the brightest area of highlights and the meter is looking to define that as 18 percent gray. If you did it using matrix metering of the whole scene you would get very different results. But doing what Gene and Bob say does work in my experience too.

On its face value, it has merit. However, there is a lot being left out of this statement. For ETTR to work best, spot metering should be used along with spot metering on the very brightest part of the scene. Secondly, one needs to know how their camera reacts to the exposure being pushed. I use ETTR with The Zone System almost 100% of the time. However, that was after extensive testing and profiling the camera and various lenses I use.

I recently purchased a "new" camera. I've yet to actually use it to photograph anything but exposure charts. I need to know how the camera responds to additional exposure. What I don't agree with in the statement is a fixed amount of additional exposure. That should be determined by the scene and the desired placement of the high luminance values.
--Bob

I can, but with reservations. See my thread on Severe underexposure.

Note that this image was underexposed about 4 stops but there was still a lot of information that could be recovered in the shadows:



It was taken with a full frame camera that has a wide DR of a scene with a relatively narrow DR. This was a convenient pairing of the camera's DR with the scene's DR.

As I stated in the thread, so much underexposure is not a safe practice. However, with modern cameras it's safer to underexpose a little than to overexpose and risk blowing the highlights. If you don't know what you are doing, ETTR is not a safe practice.