ETTR For Low Key images

[Steve48]

For a low key end result, assuming I am shooting raw, is it appropriate to expose to the right, and then lower the exposure in post? I use a flat Picture control in camera (nikon d810) in an attempt to get an accurate live histogram, and then use the liver histogram to set the exposure.

[digitaldog]

It's always appropriate to expose optimally for the kind of data you're capturing which is kind of what ETTR (an old acronym that should go away) implies. Expose optimally for raw data which differs from exposing for a JPEG where you'd ETTR because you don't have a better means to evaluate optimal exposure for raw. Tools like RawDigger that provide a true raw histogram allow this without ETTR coming into play.
https://www.rawdigger.com/howtouse/deriving-hidden-ble-compensation

[kpz]

It's always appropriate to expose optimally for the kind of data you're capturing which is kind of what ETTR (an old acronym that should go away) implies. Expose optimally for raw data which differs from exposing for a JPEG where you'd ETTR because you don't have a better means to evaluate optimal exposure for raw. Tools like RawDigger that provide a true raw histogram allow this without ETTR coming into play.
https://www.rawdigger.com/howtouse/deriving-hidden-ble-compensation

This is a wonderful link. Thank you for posting it.

[Garnick]

For a low key end result, assuming I am shooting raw, is it appropriate to expose to the right, and then lower the exposure in post? I use a flat Picture control in camera (nikon d810) in an attempt to get an accurate live histogram, and then use the liver histogram to set the exposure.

I couldn't agree more with Andrew concerning "optimal" exposure.  As in the B&W days and the analog "Zone System", one would always try to capture as much information as possible in the negative.  Then leave it to the developing (processing) of the film to make sure the field work was maintained and the negative ready for the next step - printing.  The same approach should be used for digital exposures and processing, albeit a totally different set of tools, all based on traditional darkroom procedures.  For instance - Photoshop Unsharp Masking to enhance sharpening of the image.  I was producing unsharp masks for B&W negatives to increase the sharpness in the print long before I started using Photoshop.  One more tool that Photoshop brought out of the darkroom and introduced to those who had perhaps never been aware of such a procedure. 

All of that aside, here's another excellent article about Optimal Exposure by Bob DiNitale right here in LuLa.  Here's the link to Bob's site and the page where you can find that article "Optimal Exposure", as well as a lot of other very helpful information.  You can purchase the whole article also, which I would recommend -- http://www.bobdinatale.com/bobs-books/ .

Gary         

[Bart_van_der_Wolf]

For a low key end result, assuming I am shooting raw, is it appropriate to expose to the right, and then lower the exposure in post?

Hi,

By simply using the exposure meter in the camera for a low contrast scene, you are likely to expose the highlights in such a way that the Raw date still has a lot of headroom before it starts clipping. Therefore it will improve the S/N ratio of the entire range of image data if you allow more photons to expose the image, by 'overexposing' relative to the regularly metered exposure. The scene contrast determines how much 'hotter' the exposure can be, still without clipping the highlights. A tool like Rawdigger will allow to fine tune how far you can increase exposure before clipping the highlights. It is not impossible, but hard to do this with the camera generated histogram which is based on the image's JPEG thumbnail. One way you could get close to optimal, is by (spot-)measuring with the camera meter the highlights that should remain unclipped, and add approx. 3 stops of exposure bias.

Once you have the optimally exposed image (which will have an image thumbnail that's too bright) you reduce the brightness of the exposed image to normal levels, with the controls for that (often labeled "Exposure", which is a bit of a misnomer) in your Raw converter. The resulting imagedata should have a better Signal to Noise ratio, leading to cleaner and usually more robust imagedata for postprocessing. Do note that some Raw converters are limited in the number of equivalent Stops that the apparent exposure can be reduced. Also, some Raw converters apply a tone curve, assuming an average exposed image. Pulling "exposure" in postprocessing, works best with a relatively 'linear' tonecurve response.

Cheers,
Bart

[digitaldog]

Therefore it will improve the S/N ratio of the entire range of image data if you allow more photons to expose the image, by 'overexposing' relative to the regularly metered exposure.

Or to put it another way: Therefore, it will improve S/N ratio of the entire range of the image data if you allow more photons to expose the image by optimally (correctly) exposing the image relative to the incorrect metered exposure by the camera.
Bypassing (understanding) how easy it is to fool meters and how wrong they can be is something film photographers have done forever and hopefully digital photographers can/will do as well. Meters can be dumb and fooled, photographers who understand how they can be fooled and how to expose for any media can optimally expose any media. Takes some testing. And ignoring not only dumb meters but dumb LCD screens and their dumb Histograms that lie.