Chromatic aberration?

[jjd228]

I understand what this is. Can someone tell me why the Canon t4i has the CA correction setting turned OFF by default? From what I have read this setting does a really good job so what would be the reason to leave it off?

 

[480sparky]

Faster processing in-camera.

 

[jjd228]

Faster processing in-camera.

So if I'm not in a rush you'd recommend turning this on?

 

[Ysarex]

This is a software correction. It applies only to the camera generated JPEGs. Canon is able to make it work well because they're familiar with the camera's optics. If however you use an unfamiliar optic (eg. get an adapter to use an old non-Canon lens from 50 years ago) then this software algorithm may not work so well and turning it off could be an advantage.

Joe

 

[TCampbell]

The feature is LENS-SPECIFIC. If it has no data for that lens then it can't correct for it. Canon would likely pre-program all the Canon lenses, but not any third party lenses.

You can fix CA in post-processing. I use Aperture, which does support CA correction.

Technically you can fix it with any software that allows you to separate the "red" and "blue" channels out of the image. Basically you create a red, green, and blue channel version of the image and then fractionally resize the blue to be a touch bigger and the red to be a touch smaller and then re-merge them.

As Joe explains... RAWs are never edited in-camera. So if you're shooting RAW there's no point trying to turn it on because it'll still be ignored.

 

[Gavjenks]

Short answer: in-camera software isn't powerful enough to do the best quality CA-correction, and taking shortcuts could destroy data and make it impossible to do a better job later. So if shooting JPEG only, it depends on how much post processing you want to do. If shooting RAW, it doesn't matter what option you choose, and if shooting RAW+JPEG, turn it on.

Long answer: Chromatic aberration is very complex. It's not simply an issue of RGB channels being at different sizes, because real light does not come in discrete red, blue, and green wavelengths only. Complications:

A) Each type of color sensor in your camera actually records across a whole range of wavelengths, some of which may overlap with the other sensors' ranges, etc. etc. The result is that the RGB channels can be distorted in weird ways such that the same object might be different shapes in each channel.
B) Also, some information is lost forever by poor optics and may have to be "invented" in order to make the image look better. For one example, if chromatic aberration causes an object to entirely fall outside of the sensor in one channel but not the others. It wouldn't be recorded at all and would have to be extrapolated.
C) The 3-dimensional shape of the scene will affect amounts of CA, etc., and your camera doesn't know what the 3-dimensional structure of the scene is precisely.
D) The amount of CA often does not vary evenly across the lens. It will interact with spherical aberration, etc. in complicated ways.
E) Channels will be at different focus from each other, so one channel will usually be sharper than the others. Re-aligning them alone will not fix this, and there is no way to perfectly or objectively fix softness.

Correcting CA, therefore, can be very complex. Fixing these things requires powerful algorithms that use a large amount of processor power and may even require some amount of human decision-making to create the best possible end product. It would be impractical to do this type of highest quality processing in-camera. So if you want the best end product, it should done in post processing instead, and the camera shouldn't touch it at all.

This only applies to JPEGs, though. So if you're shooting JPEG only, you need to decide: "do I want to do more post processing and get a higher quality correction? Or do I not care about absolute pristine quality and would rather save time by using the simpler, rougher correction in camera?"

If shooting RAW only, it doesn't matter, because in-camera correction won't be applied to the image anyway, even if you have it turned on, and your converter software will do a better job.

If shooting RAW+JPEG, then turn on your in-camera correction. Because you would only ever use the jpegs themselves for previews or if you wanted a quick and dirty product, and quick and dirty CA-correction fits the bill. Or if you want the higher quality processing, you would be using your RAW, so there are no drawbacks.