This is where Ian Burley, DPNow's editor and founder, shares his unique thoughts and impressions on subjects that he hopes will be of interest to others on the subjects of digital photography and other related or loosely related topics! You can follow DPNow Editor's blog on Facebook and Twitter, too.

Rate this Entry

Over and under exposure - which is least destructive?

Submit "Over and under exposure - which is least destructive?" to StumbleUpon

Submit "Over and under exposure - which is least destructive?" to Google

Posted 12-06-09 at 04:17 PM by Ian
Updated 23-06-09 at 04:44 PM by Ian

Conventional photographic wisdom suggests that you should avoid over exposure at all cost. Roasting highlights, we're told, can't be recovered and your picture will be irrevocably ruined.

And of course there is a lot of sense in such advice. But should we take this advice with a pinch of salt in the digital domain?

Below are two shots of the same scene taken within a fraction of a second. One was taken 2/3rds of a stop under the camera's metered normal exposure and one was taken 2/3rds of a stop over the ideal exposure. That's a separation of 1.3 stops.

Using the camera RAW files, I have manually normalised the images as best I could in Adobe Photoshop Lightroom, using exposure, highlight recovery, and black level adjustments only. The results were re-sized and exported (sRGB colour space) for upload to my DPNow gallery.

Image 1 (above)

Image 2 (above)

I'll reveal the answers in my next blog update and discuss why over-exposing slightly might be preferable to under-exposing for safety.

Posted in Learn about photography, Photoshop, Photoshop Lightroom

Views 2935 Comments 13

« 11,000 images catalogued and counting... Main Expose to the left or to the right! »

Total Comments 13

Comments

	I beieve that the image 2 was the one overexposed becuse of the results in postprocessink the sky and clouds. In both cases, the shadows did not suffer a lot because of a relatively low contrast scene, with nearly similar EV values. Horacio.
	Posted 12-06-09 at 04:58 PM by Horacio

	Hi Horacio! Which picture is best, overall, in your opinion?
	Posted 12-06-09 at 05:13 PM by Ian

I'm not really sure what I want to say about this. Essentially though it struck me that the idea of under or over exposure is a difficult one and because the camera says the image is correctly exposed, that doesn't mean to say everything in the shot is correct. Evaluative metering used by cameras most of the time is really only giving an average of what it "sees" and even spot metering is really only based on an 18% grey.

So underexposing is a good rule of thumb if there are bright highlights in the frame, and its up to the photographer to "read" the frame in order to compensate for bright or dark elements of the shot. However another rule of thumb has always been "Expose to the right" this being a reference to the Levels curve. I've often tried this with varying degrees of success, but its not exactly the same as underexposing for the highlights. However with a scene that is fairly even tonally throughout, exposing to the right works better.

I'm not really bothered which of these shots was under or over exposed. I don't think it matters too much when, as I do, I use raw. The software allows me to tweak every shot in order to get what I feel is the optimum exposure. I'm currently going through a set of shots where everyone has had some degree of correction, the camera simply can't do by itself

If I was to guess I'd say the second shot had been overexposed. I only say this based on my knowledge of what Lightroom does to clouds when the Recovery slider is used. For me there is a little more form in the cloud. I do however admit I may be wrong

Posted 12-06-09 at 05:20 PM by Stephen Stephen is offline

	Evaluative metering breaks up the image into segments and then statistically evaluates the values, using an exposure setting that works best for the majority of the segments. It's important to stress this because it is different to an 'average' system, where small areas of big density difference will skew the exposure, but will be ignored in evaluative metering.
	Posted 12-06-09 at 05:27 PM by Ian

	"Best" is very subjective in this instance. I think its impossive to give a definitive answer to this. Which kinda proves that getting it right in the camera is not as important these days as it used to be when using slide film.
	Posted 12-06-09 at 05:28 PM by Stephen

Quote:

Originally Posted by Stephen

If I was to guess I'd say the second shot had been overexposed. I only say this based on my knowledge of what Lightroom does to clouds when the Recovery slider is used. For me there is a little more form in the cloud. I do however admit I may be wrong

I'm a bit confused by this

If the second shot has more cloud detail, shouldn't that be the under-exposed one, with the over exposed one having lost detail in the bright cloud areas?

Posted 12-06-09 at 05:30 PM by Ian Ian is offline

Quote:

Originally Posted by Ian

Evaluative metering breaks up the image into segments and then statistically evaluates the values, using an exposure setting that works best for the majority of the segments. It's important to stress this because it is different to an 'average' system, where small areas of big density difference will skew the exposure, but will be ignored in evaluative metering.

Nevertheless, one can't expect evaluative metering to always get the "correct" exposure. It is simply working out an optimum setting for the camera based on certain light information in the different zones of the frame

Posted 12-06-09 at 05:30 PM by Stephen Stephen is offline

Quote:

Originally Posted by Stephen

I completely agree. I just wanted to point out the differences. In the end, we're simply talking about the amount of light the camera is letting through to the sensor overall.

Posted 12-06-09 at 05:33 PM by Ian Ian is offline

Quote:

Originally Posted by Ian

I'm a bit confused by this

If the second shot has more cloud detail, shouldn't that be the under-exposed one, with the over exposed one having lost detail in the bright cloud areas?

In my experience cameras often don't give great detail and form to cloud. Clouds can be enhanced with the software esp ACR in Lightroom and CS3/4 The recovery slider and esp the clarity slider do a great job in putting more detail in. The fact is that with the first shot, if I'm right, there were no over exposed areas so you left it. With the second shot you used Recovery and this added a bit more form to the cloud area. Just my theory you understand

Posted 12-06-09 at 05:36 PM by Stephen Stephen is offline

Quote:

Originally Posted by Stephen

OK - good point.

Posted 12-06-09 at 05:37 PM by Ian Ian is offline

	Don't want to get into "handbag bashing" but to me number two has slightly more detail in the sky (but not significantly so), then again number one as more oomph! That's as technical as I get gentlemen jo
	Posted 12-06-09 at 09:09 PM by jojo

I agree here with Horatio the scene is relatively flat so talking under or over exposure has little relevance, the image could be retrieved under or over exposed.
Its scenes with very hard light & high contrast that will perhaps benefit to underexposure to save the highlights. In any case as Stephen suggest what is correct exposure anyway. When I studied the Ansel Adams zone system it very soon became evident correct exposure was what the photographer chose correct exposure to be. Putting zone 5 where he or she wanted. Its still a good maxim in contrasty light to ensure highlight detail by under exposure, shadow can usually be retrieved. In very contrasty lighting I often underexpose by as much as two stops, although now I have invested in a Cokin Z pro filter system I will use graduated neutral density filter 2 stop or 4 stop, for landscapes and expose more for the foreground. Getting it correct in camera or as close as possible is another good maxim, or at least what is right for you.

Patrick

Posted 12-06-09 at 11:22 PM by Patrick Patrick is offline

	Yes, I agree - the lighting was too flat to make this experiment work to its fullest extent. But it does show that we shouldn't be too afraid of over-exposing slightly, at least not in these conditions. I've re-taken the picture in more contrasty, sunny conditions. See my next blog update...
	Posted 14-06-09 at 12:58 PM by Ian