General Photography - Techniques & StylesDiscuss the fundamentals of photography, photographic technique, infrared and macro shooting, and related topics here!
I have seen a few comments recently when someone is commenting on someone eles photo about over/under exposure. It got me thinking about what is considered to be correct exposure.
I consider there to be 2 kinds of perfect expsure.
1. Rendering the scene as it looks to the naked eye. Usually this is a low light shot and histogram is a fair way to the left... even missing some shadows. The scene looks as it did in real life.... often too dark for some reviewers.
2. Perfect histogram expsure. Histogram spread all the way from left to right and no clipping. High DR and no loss of detail. This renders low light scenes very very bright and nowhere near what they look like in real life.
I personally shoot with number 2 but quite often use PP to make it like number 1 (typical expose to the right).
BUT... when commenting on someone eles work is it correct to comment on the exposure or the birughtness? (assuming no loss in detai).
There is no correct exposure. There are exposures that the camera meter tries to achieve. Namely, 12% gray. If you shoot a sheet of white paper, the camera will meter to make that piece of paper look 12% gray.
For many things it works relatively well (depending on your subject and metering mode), but you have to understand what the meter is trying to do, and why you would want to override it. For example, the camera will try to make snow (or large expanses of white) into gray, so you have to add +1-2 EV to the shot to make the snow white. Similarly for black subjects, you tend to need to subtract -1-2 EV to retain the black without turning it gray.
After a while, you develop the skill to recognize when a photo is exposed 'properly'. It took me a while to recognize it, but I can usually tell pretty easily now.
One thing that you can do to test exposure is to meter off your hand or palm in the light that you are using. It's pretty close to neutral gray, and makes a handy (hah hah ) tool that is pretty difficult to misplace.
The subject, or part of the subject intended to be the eye-catcher, is the most important part to get exposure right - all else is secondary.
Fitting each pixel within the dynamic range of the histogram ensures no 'loss' of detail from all the pixels, which is fine if everything in the image needs to come out, though many images don't fit here because of black backgrounds, low-key moody shots, or high-key captures, etc...
The way it looks to the naked eye is of great importance, but hard to get right unless the screen is adequately bright and well calibrated. But what really matters is how it pops out to the viewer, not necessarily how perfect its histogram looks.
The correct exposure is the one that allows you to render the image as you want in rendered.
Unlike Dave, I believe that there is one correct exposure for any given scene, but it will vary from person to person depending on how they want to render the image.
Unlike Dave, I believe that there is one correct exposure for any given scene, but it will vary from person to person depending on how they want to render the image.
There is a correct exposure for a given level of light / ev. But there may be several different ev levels in a scene and thus not possible to expose all parts of the scene properly. For example the ground may be properly exposed but the sky is all white and overexposed. In this case the photographer has to make a choice.
I have been struggling with just the problem outlined by Andreas. I wanted to practice getting the exposure to match what my eyes saw in an admittedly wide dynamic range photo. Please don't judge the composition (it was only a practice for exposure) but in this case the after (HDR processed) shot was much closer to reality than what I could capture with one exposure alone.
First picture is the best I could manage with one exposure. The second is the result of merging 4 exposures. What do you think?
Last edited by mithrandir; 11-07-2009 at 01:45 PM.
I have been struggling with just the problem outlined by Andreas. I wanted to practice getting the exposure to match what my eyes saw in an admittedly wide dynamic range photo. Please don't judge the composition (it was only a practice for exposure) but in this case the after (HDR processed) shot was much closer to reality than what I could capture with one exposure alone.
First picture is the best I could manage with one exposure. The second is the result of merging 4 exposures. What do you think?
What a good example of proper hdr/tone mapping!
Regarding proper exposure, I think there are implicit definitions in the ISO specifications.
One simple definition relates to the absolute intensity of light & length of exposure time necessary to saturate a photo detection system; another relates to the light intensity & time needed to achieve a S/N ratio of 40. See wikipedia for ISO.
The correct exposure is the one that allows you to render the image as you want in rendered.
Unlike Dave, I believe that there is one correct exposure for any given scene, but it will vary from person to person depending on how they want to render the image.
Talk about contradiction did you mean: there is not one correct exposure for a scene, everyone will have a different opinion.
As an individual there may be one correct exposure but as a collective there are many possible correct exposures.
I also am a firm believer in natural looking HDR being very realistic to the scene as seen by the eye. On a side note... it's weird how in bright daylight the DR of the eye beats the camera (ie we can see the light and dark with detail in each) but at night the camera (with long exposure time) beats the eye. Where is the middle ground......
Thanks for the responses everyone.
The lesson in this. When Wheatfield comments on your exposure being wrong incorrect..... tell him to <insert your comment here>!
I have been struggling with just the problem outlined by Andreas. I wanted to practice getting the exposure to match what my eyes saw in an admittedly wide dynamic range photo. Please don't judge the composition (it was only a practice for exposure) but in this case the after (HDR processed) shot was much closer to reality than what I could capture with one exposure alone.
First picture is the best I could manage with one exposure. The second is the result of merging 4 exposures. What do you think?
I think the issue was with the window. If it weren't there, you would have been able to get the room to look with just one exposure, without worrying about blowing out the window light. The HDR shot looks better as you were able to capture detail in the near parts of the room without overexposing the window.
Maybe they should make a "draw your own graduated filters", sort of like etch a sketches, where you can use a little pen, make the shape of the highest part of the scene (here the window light) and then shoot the scene with that part darker! Hey, wait, can't you do that in photoshop? Have you tried shooting a RAW and then upping the exposure on everything but the window? I'm not good in PP so I couldn't tell you how, but there must be a way. Just wondering how that would look compared to an HDR shot.
I believe to get the most detail in any shot there is only one correct exposure value. If the highlights (usu skies) are outside the dynamic range of the sensor then using a polarising filter or graduated grey filter will help to bring them back within range. Ultimately you may need to bracket till such time that sensors have a dynamic range wider than slow B&W film (10 to 12 stops)
Whether the final image is low key or high key can be sorted later in PP. A perfect exposure will keep all your options open. You can always loose data you can't re-create it.