To ETTR or not to ETTR...?

Started 3 months ago | Discussions thread
TomFid Veteran Member • Posts: 3,115
Re: Yo PanOly - make this easy!

FingerPainter wrote:

TomFid wrote:

The point is, they should be. Any process that's not essential to the qualitative expression of the image is a good candidate for automation. Figuring out how much exposure can be given without clipping, subject to some constraints like how much shutter speed is needed, is something the camera can do more quickly than you or I can read a histogram and react.

It would be nice if they would be designed that way. A problem is how to teach the camera which hightlight details I care about and which ones are inconsequential. Can a camera recognize a specular higlight? Maybe there would be some way to let the camera query you about which highlights you want to preserve.

I think you're never going to teach it "which," but "how much" might be doable.

For example, it should be easy to compute the fraction of pixels that are at risk of clipping or falling far enough into the shadows to have bad S/N. Put that on a quasi-log scale (1%, .1%, .01%, ...). As a start, try to balance the losses on each end. To provide some user control, reinterpret the exposure compensation dial as a shadow/highlight tradeoff dial, that shifts things in favor of the shadows or highlights.

Of course, none of this is really needed if you have enough sensor DR and computational power. This may be a problem that makers should have solved ten years ago, whose time is now nearly past.

Post (hide subjects) Posted by
XRF
XRF
XRF
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow