Lenses and Corrections Software --Your thoughts

<p>With reference to the new dslr lenses and the software used to correct chromatic aberration, distortion and light falloff, I'd like to know what you think about them.<br>

It seems that optical flaws are frequently being dismissed, because you are now able to correct them with name-brand and off-brand software that is available today.<br>

Do you like the "look" of the corrected photograph? Can you tell it's been software corrected? <br>

I'm sure it makes a lot of sense for people who cannot afford top-shelf glass, and for those who use zooms, which are notorious for the above-mentioned optical flaws.<br>

One thing is for sure; this software makes r&d and mfg a lot less costly.</p>

 

<blockquote>

<p>I'm sure it makes a lot of sense for people who cannot afford top-shelf glass</p>

</blockquote>

<p>You need to disabuse yourself of the notion that there are <em>any</em> lenses that do not involve irresolvable contradictions in their design and manufacture. Some lenses do better than others, and the closest to the "perfect" lens will be likely be a slow, medium focal length, prime lens. Check the reviews at Photozone.de and you will see that there is no lens that does not have some kind of wart or compromise, although a few reduce those to an astonishing degree.</p>

<p>It's not just zooms, it's fast lenses, wide lenses, and so on and on. Like other lenses, even "top-shelf" lenses are made for specific kinds of shooting. In any given case, the designers will have done their best to reduce problems that hamper achievement of that particular goal, in the process, other corrections will be less emphasized of necessity. In such a way, an f/4 version of a "top-self" lens may outperform an f/2.8 version at that focal length in some particular way like resolution. The cheap little "kit" 50mm lens may actually be 'sharper' or have other optical advantages over a lens that has been tweaked to have an f/1.2 maximum aperture. Anybody who buys the latter is probably (or should be) aware that 'sharpness' is not the reason for buying an f/1.2 lens.</p>

<p>So, yes, the software helps. Usually the correction(s) is (are) not at all obvious, but the more you do, the more defects can emerge as a result of overworking the correction. Like so many things, if you can tell it's been done, then it's been done too much. If you correct chromatic aberration, vignetting, perspective, and a host of other things all in the same picture, it may well "show" in a bad way. The software doesn't make getting the best lens you can get unnecessary, but getting the "best" lens you can afford, doesn't mean you can avoid some software correction in critical cases.</p>

<p>Most of the time, little or no correction is necessary because the 'warts' often show up only in special kinds of shots. Shooting a little bird with a big telephoto in a tree is probably not going to show any barrel or pincushion distortion, although it may almost certainly be present. Shoot a brick wall, different story, so the software may be critical if you shoot brick walls for a living. The same principle goes for vignetting, chromatic aberration, and all the other warts that lenses have in the real world.</p>

<p> </p>

<p>Getting things right optically is always preferable to fixing it later in software.</p>

<p>Warping the image to correct for geometric distortion necessarily involves interpolating pixels, and therefore reduces sharpness to some degree. This is less of a problem the more MPs you have, but if you look at 100% there is often a bit of wavery blurriness particularly around sharp, contrasty edges after geometric distortion correction has been applied. It has been noted in some reviews that the Olympus micro-4/3 pancake lens seems soft around the edges (where geometric distortion is greatest) partly due to software corrections.</p>

<p>CA, I think, can be fixed in software with relatively little ill effect, but I'd still rather use lenses that correct it optically.</p>

<blockquote>

<p>You need to disabuse yourself of the notion that there are <em >any</em> lenses that do not involve irresolvable contradictions in their design and manufacture. </p>

</blockquote>

<p>Sure, but I don't think that's what the question was really about. No lens is perfect and sometimes it will be helpful to apply corrections during post-processing. You're also right that quite often some distortions aren't harmful to images -- if there are no straight lines in the picture, you aren't likely to notice a small amount of geometrical distortion. But the question as I read it concerned the trend among some manufacturers to produce lenses with fairly severe distortions and automatically correct them in-camera when producing JPEGs. For inexpensive consumer products it is an understandable engineering compromise that helps to keep costs down while producing images of acceptable quality, but from a more serious photographer's perspective I don't think it's really a viable alternative to better-quality lenses.</p>

<p>I asked this in 2008<br>

<a href="http://www.photo.net/casual-conversations-forum/00RtSl">http://www.photo.net/casual-conversations-forum/00RtSl</a></p>

 

<p>See also <a href="http://www.dpreview.com/articles/distortion/">A Distorted View</a>.<br>

The differences in their samples are ever so slight.</p>

Was there any way to fix these lens aberrrations in the darkroom during the great days of films ? The reason why I'm asking is because other that Perspective Control, I usually didn't give a damn.

<p>I'm not aware of any.</p>

<p>Regarding CA the theory was, that ideally it would be cancelled out by the different colour-sensitive layers being stacked on top of each other, rather than being in the same plane as with a Bayer sensor.</p>

<p>Scans of my slides reveal though, that this didn't work out most of the time.</p>

<p>I don't think it's a cheap vs expensive thing, necessarily. Doing some of this correction in firmware can allow the use of a lens that is smaller, and perhaps even better in some ways where it really counts. Don't forget that digital camera sensors have special needs that did not need to be addressed by lenses made for film cameras.</p>

<p>@Pierre Lachaine-- True, concerning the special needs of camera sensors. You either address it through optical engineering, or through software.<br>

I'm wondering if optical design will branch out; one for photochemical photography, and the other photoelectric. </p>

<p>@Pierre and/or Mark, what sort of "special needs of camera sensors" did you have in mind? I realize that sensors and film are different creatures in many ways, but the most obvious difference that might affect lens design is the greater reflectivity of sensors, which has been addressed in recent years by the development of new lens coatings.</p>

<p>This feature is now being built into Broadcast HDTV camera/lens combinations. It is known as CAC or chromatic abberation correction.</p>

<p>On looking over my own earlier post, especially given recent grammatical discussions here,</p>

<blockquote>

<p>Shooting a little bird with a big telephoto in a tree is probably not going to show any barrel or pincushion distortion, although it may almost certainly be present</p>

</blockquote>

<p>I am now wondering how the little bird got the telephoto into the tree in the first place. Perhaps it was an <a href="

Swallow</a>?</p>

<blockquote>

<p>I'm sure it makes a lot of sense for people who cannot afford top-shelf glass</p>

</blockquote>

<p>You'd be hard pressed to find higher quality lenses than those made by Zeiss for my Hasselblad.</p>

<p>The software that Hasselblad provides for use with its CFV-39 39-megapixel digital back includes correction functions, and they work quite well.</p>

<p>This does is not in any way associated with the design of the lenses. In fact the glass existed long before computer correction.</p>

<p>There's no such thing as a distortion-free lens.</p>

<p>- Leigh</p>