So after working with film for a while with only a cursory technical understanding, I've started seriously shooting digital and I'm going buck-wild with the technical side of digital photography. I've recently encountered the "expose to the right" principle: digital sensors have to encode a logarithmic phenomenon into a linear space, so the amount of detail devoted to each successive stop drops exponentially from the brightest stop downward (please correct me if that's wrong or incomplete).

The advice I see stemming from this observation seem to center around minding your histogram (which my weirdo Leica M-D doesn't support...) to push your exposure as close to the right as possible without clipping.

The thing is, this sounds suspiciously similar to the film photography principle of exposing your shot to capture details in the shadows, letting film's silky-smooth nonlinear saturation handle the highlights, and then adjusting in the darkroom.

I've seen advice to use spot or center-weighted metering on the brightest point of a scene with an extreme upward exposure compensation to push things to the right. But why not do the film thing and expose for the shadows and let the highlights do what they will?

As I write this I realize highlight clipping is probably a factor, but humor me: maybe there are other reasons I'm not thinking of.


2 Answers 2


It's a matter of giving attention to not discarding what can't be recovered; if you expose film for shadows, you're risking some "silky-smooth nonlinear saturation" being applied to the highlights. If you you expose digital for shadows, you're risking hard, off-the-edge-of-a-cliff clipping, which is a very decisive loss of information that you probably want to avoid.

A key point in ETTR is that you meter for the brightest part of a scene for which you want to retain information. This may not be the actual brightest part of a scene, such as specular highlights that will be saturated for any reasonable rendering of the scene that you have in mind. You let the shadows "do what they will" because the loss of information there can at least be managed with noise reduction, etc.

  • 1
    \$\begingroup\$ Also, If I recall correctly, film has about 2.5 stops more flexibility in the highlights than digital has. \$\endgroup\$
    – MikeD
    Commented Mar 6, 2017 at 12:42

But why not do the film thing and expose for the shadows and let the highlights do what they will?

Film and digital sensors react converse to extremes of light:

With negative film, you have a sharp cutoff in the shadows, but an asymptotic behaviour in the highlights. So you need to expose for the shadows where you want to keep detail and take care of the highlights in development.

Digital sensors, OTOH, have an absolute cutoff in the highlights, while the shadows are less well defined, it just gets noisier if you brighten them in post. So you have to take care of your highlights when exposing. At the same time, you'll want to expose as bright as possible while still keeping detail, to gather as much information as you can, which you then can pull down to a correct exposiure in post processing.

  • \$\begingroup\$ One useful method of ensuring highlights remain generally usable on a capable digital camera is to consistently set your metering to underexpose by ⅔ to 1 stop then recover the brightness when developing the raw file, all the while metering "normally". It goes without saying that you should understand the limits of the camera's noise and shadow detail. \$\endgroup\$ Commented May 3, 2018 at 0:08
  • \$\begingroup\$ @NickBedford consistently underexposing is about the opposite of the ETTR idea and really bad for noise levels. \$\endgroup\$
    – ths
    Commented May 3, 2018 at 6:39
  • \$\begingroup\$ It's not ETTR, but it's the digital equivalent of consistently overexposing film and "metering for the shadows" on film then handling the overexposure in scanning and printing. Consistently metering under on a digital camera is one way of shooting quickly without needing to check your histogram all the time, while ensuring highlights are less likely to clip. It's not bad for noise levels if your camera can handle it and you understand the point of it. \$\endgroup\$ Commented May 4, 2018 at 1:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.