lens chromatic abberation setting in canon 80D camera menu

[ivans75]

hey guys, i was just playing around the camera to learn deeper of its feature, i came across "lens abberation correction" in one of the menu setting. If i clicked on it, it shows the lens that is on at the moment (50mm f1.4) and it says "correction data available" but I have the Chromatic abberation, peripheral illumination and distorition sub settings disabled.

Before I am writing this, I took a bit of search about CA and it tells me it is that nasty discolor or wrong color effect around the edge of an image through a lens in which light spectrum don't bend properly on the pentaprism. And when I saw that "correction data available" does it mean the lens i am using has CA?

I did few tests on f1.4 to f9 on AV mode iso 200 (manual white balance) in my dining room (with 2 fluoro lights) and i didnt see any of the nasty shadow of pink or purple on the edge of the image, infact it looks fine to me unless I am wrong?

Questions are:
should i enable the ca, pi and dist. setting in the menu?
does it mean my lens has a minor CA when it says "correction data available"?

Not sure about the other 2 (peripheral illumination and distortion) i dont see anything illuminated out of control and nothing distorted so not sure what these are lol

Thanks in advance

- - - Updated - - -

oh one more thing, should I check on this setting before I use a new lens? I also read that CA is mostly happening in telezoom, what is your opinion on this? And what cause CA? Knock on the lens ?

[Tannin]

Nope, you are right Ivan. Nothing to see with that lens unless you get really OCD about it.

All lenses have CA. The only question is how much, and under what conditions. If you do want to go OCD about things, look for CA in the corners of images shot wide open. (Actually, don't waste your time doing that. Life is too short. I doubt that you'd find much in a decent 50mm lens anyway: normal lenses, especially normal primes, are among the easiest of lenses to design and manufacture.)

The "correction data available" you see simply means that the 50/1.4 is one of the 30-odd lenses for which Canon have helpfully provided lens profiles stored in the camera's ROM. You can download and install profiles for many other lenses also. (Possibly not non-Canon models: Canon sometimes like to pretend that you will only ever use Canon lenses.)

More than you ever wanted to know about lens abberations: https://www.lensrentals.com/blog/2010/10/the-seven-deadly-aberrations/

I would normally leave all of those settings off. From memory, they mostly apply to the in-camera JPG, not the raw, and in any case they can all be applied later (if desired) using DPP or Lightroom or Photoshop or whatever. They also have an effect on performance, though that's unlikely to matter in practice. There are various subtle complications associated with most (all?) of them too.

Probably the only "fancy" image correction function I would use is the Auto Lighting Optimiser. This is great for avoiding blown highlights and getting correct exposure most of the time, but be aware that it overrides the your own exposure settings to some extent, meaning that exposure compensation may not work as expected.

All of this is explained in detail in the camera manual, but you probably don't have the full manual. When I bought my 7D II recently (in Australia from official Australian stock) it came with a 180-page "basic instruction manual". Compare with the 5D IV I bought from a Hong Kong grey market vendor, which came with the full manual - all 650-odd pages of it! But you can always download the full manual from Canon if you do not have it.

[arthurking83]

I'm pretty sure that these types of features, are only relevant to the user if they use the manufacturer's raw software after the images have been downloaded to the computer .. OR if you shoot jpg.

That is, those in camera settings will be applied to the jpg images during the write from raw data to jpg, or, they are applied as a notation in the raw file to tell the computer software that this feature is enabled, and to render the raw file with that feature enabled.

So .. if you use any thirdparty raw software like Ps or Lightroom or whatever, the settings in camera(on the raw file) are pretty much useless.

CA is where you get strange colour shifts are harsh contrast edges.
So in a landscape shot with trees/leaves at the very edges of the frame, you may see purple or green fringing where the edges of the trees/leaves are set against a bright sky.

Peripheral illumination will be vignetting. 50/1.4 will vignette badly when set to f/1.4. All lenses vignette a little at any wide open setting(small aperture number). this reduces as aperture is stopped down.
99.9999% chance you won't actually see much on your camera tho. 50/1.4 is a full frame lens, and 80D is an APS-C sensor. your sensor only sees a small central portion of the lens image circle. So you don't see this 'periphery illumination' effect much.
It may be so slight, but you need to look hard, you may see it.
If you do want to test to find out if you can see it, you need to set camera to spot meter. Set lens to wide open(in this case f/1.4) Point to a plain colour sky(white or blue) but must be plain colour .. no clouds rendered. If you have clouds, it's best to defocus to see this effect.
Camera should be set to 0Ev exposure compensation too, but maybe -0.3Ev may show the effect a bit too.
Then when you're looking at the image in your software notice any difference in the brightness between the centre of the image and the very edges of the image(corners will show it more). You still may not see it, but the software will!
Somewhere in your software should be an indicator to display the RGB values for the pixels on where the cursor is located at the time.
Can't help as all software is different, but it will show this in almost all software. If you find this feature in your software, then point the cursor to the centre of the image. As an example it may show RGB values of (say) 130, 140, 125. This would be a normal (0Ev) exposure set fo values for RGB. But if you now point the cursor at the edge of the frame, you may notice that the RGB values may only be something like (say) 90, 89, 98. So brightness values for the RGB set at the edges is down by a small amount.
Note that RGB values at 255 for all three = white(usually blown highlights). And RGB values of 0 for all three = black(or lost shadows).

50mm lens will have close to zero distortion. If not close to zero, then less than 1% barrel for sure. 0.5% barrel is about the norm for cheap 50mm lenses.

So unless you use Canons' DPP software, or shoot jpg(only) in camera, I'd say turn it all off!

[ivans75]

Wow alot to digest here i ll take it slow and i ll do a test as arthur suggested. Tony thanks for bringing the image type. Yes i shoot in raw and never in jpg. I am leaving them off now thanks again guys. Very good information here

[arthurking83]

Wow alot to digest here i ll take it slow and i ll do a test as arthur suggested. ....

try the sky test for vigneting only.

Don't bother for the CA test tho.
I doubt very much you will see any CA from a 50mm lens.
In fact there are different types of CA.
The type you correct for in camera for the lens is really called lateral CA. Think of it as 'across the image' or the CA is spreading across the frame .. lateral.

The other is called Longitudinal CA. Longitudinal is up and down ways .. or more accurately 'into the image'. May fast f/1.4 type lenses show this.
Another term for it is bokeh fringing. It shows up (usually) as green/magenta appearance in the DOF.

But the lens correction will be for the lateral CA only. Bokeh fringing is hard, if not impossible to correct for.

if you have any zoom type lenses, they're usually better for seeing CA. The cheaper the better(make that worse) they show it up.

[tandeejay]

here is what CA looks like (this was taken on a nikon 55-300 at 300mm which has plenty of CA at the long end)...

132329

and selecting the "Remove Chromatic Aberration" option in Lightroom:

132330

That's basically what the CA setting in your camera would be doing, but I believe it only applies it when it creates a jpeg. But as others have said, prime lenses tend to be designed to reduce that. Much harder (and more expensive) for the manufacturers to fix in a zoom.

[ivans75]

try the sky test for vigneting only.

Don't bother for the CA test tho.
I doubt very much you will see any CA from a 50mm lens.
In fact there are different types of CA.
The type you correct for in camera for the lens is really called lateral CA. Think of it as 'across the image' or the CA is spreading across the frame .. lateral.

The other is called Longitudinal CA. Longitudinal is up and down ways .. or more accurately 'into the image'. May fast f/1.4 type lenses show this.
Another term for it is bokeh fringing. It shows up (usually) as green/magenta appearance in the DOF.

But the lens correction will be for the lateral CA only. Bokeh fringing is hard, if not impossible to correct for.

if you have any zoom type lenses, they're usually better for seeing CA. The cheaper the better(make that worse) they show it up.

hi arthur

just took few tests to the bright blue sky and indeed i dont see any problems
i took samples from 50mm and my new 70-200mm

i also took samples from my cloth hangers in backyard with cloths hanging of course, and no i dont see any problems and yet the camera is saying correction data available on both lenses
It might be nothing as Tony has stated above. I read from few online articles they are actually intended if you shoot in jpg or both raw n jpeg. I am still having them off as we speak

[arthurking83]

so your 70-200 is the f/2.8 type I think .. if so, from memory it's CA is very good. low to nothing(significant)
But a cheap $200 type 50-200mm zoom will probably show CA. expensive lenses are expensive for a reason. The glass used is as good as it gets(up to a point taking into account $'s)

Some 50mm f/1.4's are cheap .. and by cheap I mean at about $400 cheap.
They can show a small amount of CA too, but you have to look hard to see it.
An image of (sun or sky)light through tree leaves can show this, but I'm sure only if you zoom the image to 100% pixel view or more .. and we don't really do this in normal photography.
Camera can make a difference if you can see this CA, or not, as well. Faster aperture lenses usually show this type of aberration up more than slower aperture lenses(again depends on camera body).

eg. the Canon 50/1.4 will display more CA than the Canon 50/1.8 will .. it's just the way glass is, and why the 50/1.4 is more expensive than the f/1.8 lens. But if Canon wanted CA on the f/1.4 lens to be completely zero, it's going to cost a lot more(back to the price point comment earlier).

if you compare Canon's $400 50/1.4 to Sigma's 50/1.4 Art(or Zeiss 50/1.4 Otus) .. you see that Sigma and Zeiss is much less, or even zero CA. BUT!! the price difference($1000 for the Sigma and $5K for the Zeiss lenses) .. is a lot more than Canon's $400 or so.

TDP Canon 50/1.4 vs 50/1.8 CA comparo. (https://www.the-digital-picture.com/Reviews/ISO-12233-Sample-Crops.aspx?Lens=115&Camera=979&Sample=0&FLI=0&API=2&LensComp=989&CameraComp=979&SampleComp=0&FLIComp=0&APIComp=0)

Take note of that site and bookmark it. it gives you 'a good idea' on how different lenses compare to each other.
It is important to note that I used the term 'good idea' because of one simple point .. we don't shoot photos of black and white test targets for our normal photos! We take photos of .. <insert favourite subjects here> and in those normal photos we take, there are a lot more elements that override any insignificant CA issues in the lens.

In that link, if you have the mouse hover away from the image it shows the f/1.4 image, if you hover the mouse over the image it changes to the second compared lens(the f/1.8 lens here) so you can graphically see the differences easily.
So if you look at that comparo, and think to yourself why not get the f/1.8 lens if it has much less CA(note the purple/magenta colouring in the f/1.4 lens images at f/1.4) .. you need to set some variables.
Change the aperture in the f/1.4 lens to f/1.6 and f/2 and watch it get much better.
Saying that tho, there is a reason that you may see many comments that the nifty fifty is a must have lens .. and for good reason, it makes good pictures(as you can see).
But imagine at night, in low light shooting fast moving subjects(eg. kids) you need shutter speed there. The f/1.4 allows you almost 1 extra stop of shutter speed, or 1 stop lower ISO by comparison .. and you won't see that CA in low light on people .. so your choices are what do you want/need.

Because the CA is not going to really have a lot of impact on your photos, it's best to leave that setting off in camera. Then(and this is easy) if you see it in any photo, you hit a small tool in the software to remove it ... for any image that needs it.

back to the tree leaves test. If you can, take a shot of the bright sky, and have some tree leaves or thin branches at the corner of the image. Then look at the edges of the leaves or branches. It's common to have CA there is many lenses. The edges are usually worse for CA than the centre fo the lens image circle.
This is where it can affect landscapes, and again using that click tool in software helps. Some lenses are much worse than others at that test tho, and that's where you may see a lot of comments re CA issues.
Also note that resizing a very large(eg 24Mp file) down to say 1000pixels will automatically reduce or eliminate CA on the tree test. The reduction factor will make them less noticeable .. the problem is still there, but much much harder to see because you can see individual pixels to see the CA! ;)

[Tannin]

The other thing that Arthur didn't emphasise is compromise. As AK says, a quest to eliminate one or other of the seven or so major lens aberrations mean spending more money, often a lot more money. But it also inevitably involves technical compromise in other directions. These top-class 50mm lenses Arthur mentions: take a minute to look up how big they are and how much they weigh. But it gets worse.

Essentially, a lens starts as a relatively simple design with a small number of elements. Then the trouble starts. A simple lens (any simple lens, no matter how well made or how well designed) has aberrations. So the designer adds an extra element to correct (for example) barrel distortion. More cost, more weight, and extra glass. Every time light passes through a surface (such as air to glass or glass to air, or from one kind of glass to a different kind of glass), some light is reflected. You get a loss of contrast and a loss of detail. So, from a resolution and contrast point of view, a perfect lens has the smallest possible number of elements. A "perfect" lens in this sense will also have horrible distortion, CA, and all the other aberrations we do not like. Everything you do to correct one of those faults adds other, different kinds of fault. Fix your CA but add a little distortion. Sort out the distortion at the cost of a little less sharpness. And so on. And on. And on.

This is why all lenses these days are designed using vast amounts of super-computer time, even (or perhaps especially) the cheapest ones - all in an attempt to find the best compromise between a dozen different competing requirements.

If you want to get fussy and technical about minor lens faults (don't!) you will need very deep pockets to pay for your obsession and a strong back to carry it. You will also need to give up zooms - for while it is impossible to design and make a perfect prime, it is impossible to make a zoom that is anywhere near perfect.

In reality, modern lenses are miraculously, jaw-droppingly good. Not one of us here is skilled enough to really deserve the lenses we own and use every day. But, thanks to the amazing technology and skill of the people at Tamron and Nikkor and Zeiss and Canon and Sigma and Tokina and Swarovski and Leica, for a few hundred dollars (or sometimes a few thousand) we are able to buy lenses much, much better at their tasks than we are at ours.

Celebrate that!

- - - Updated - - -

PS: an example. The Canon 50mm/1.2 is legendary. It's absurdly heavy and expensive, of course. It is regarded as a superb tool for things like intimate portraits: sharp and clear as you like, with truly beautiful, dreamy out-of-focus backgrounds very hard to match. Take all that on faith (I do, it's not my sort of photography and I'm not qualified to have an opinion).

But it also has major faults and limitations, in particular bad focus shift. Cameras always AF wide open and shoot at whatever the set aperture is set to. In the 50/1.2 (and some other lenses, but especially this one) changing the aperture shifts the focal plane such that you get back focus. This is not something that can be adjusted, it's a consequence of the optical design of the lens. At some apertures and subject distances, focus-recompose doesn't work properly and off-centre AF points don't either.

Why on earth did Canon let that flawed design out the door? Because the 50/1.2 also has great strengths and, in the hands of a skilled photographer who works within its limitations, can produce images superior to pretty much any alternative. "Fixing" the focus problems by changing the optical design would also "fix" the benefits. It's not a question of extra money. If Canon though that they could make it perfect by redesigning it and spending more, they would have already done so. They are never afraid of charging a lot for a lens. No, Canon decided that the focus weirdness was a price worth paying for the other qualities the lens has. If you don't want to put up with them, that's fine. Buy a different lens - which will of course have other limitations of its own.

[arthurking83]

Whoohooo! someone else posting long replies .. :th3:


.... while it is impossible to design and make a perfect prime, it is impossible to make a zoom that is anywhere near perfect.

....

What caught my eye out here was the term 'impossible' used there.

Almost.
Some fun facts about Zeiss .. BTW who make some very good lenses too!
They make a nice 70-200 T2.9 lens. Technically it'll be an f/2.8 aperture, but the video world don't care about aperture they care about Transmission .. hence it'll be a T2.9 lens

The fun facts:
Any normal well designed 70-200/2.8 lens will generally be about 200mm long overall, it won't change physical dimension with zooming or focusing. That's standard protocol now for any self respecting 70-200/2.8
They all weigh in at about 1.5kg. They'll all use 77mm filters, and usually their maximum diameters will be just a tad more at about 85-ish mm for the diameter of the lens body.

Zeiss on the other hand do things 'a bit' differently .. a bit here meaning gigantically a bit
First up is the length .. only 50mm more at 250mm long, and it doesn't really need it, as it's only a 200mm lens .. but that's what Zeiss do anyhow.
Then there's the additional girth. it's filter size alone is 95mm and it's girth is over 100mm.
Then there's the weight. An added 25% weight isn't considered enough by Zeiss, they like to do things BIG! .. so they add another 1.3kg, just to make you think twice .. two hundred times more ... about thinking of getting this lens. That's nearly 3kg in weight.
And to totally burst the bubble you may have been living in thinking you could get one of these close to perfect zoom lenses, comes the price. 2 x the price of the most expensive of the 4 major brand names in this category isn't enough to slug you with .. neither is 5x the price premium.
Zeiss like to think biggerer is besterer, so they've both biggerer and bested by a long shot .. 10x the price of the most expensive 70-200/2.8(which should be the Nikon version. Canon are usually cheaper, so the price multiplier factor is even more(I think)

But if you thought that $2-3k for such a lens is a major slug, Zeiss want $20K for theirs .. and that US$20K .. convert that to Aussie Lira .. and it works out to nearly Au$30K .. and then add GST to it :D The tax component is going to cost more than the Aussie price of the Nikon lens!!

But, it won't have a trace of CA, it's parfocal(will maintain focus even if zoomed(which is very rare for zoom lenses), it's IQ will be 'near on perfect' but at what price .. weight, size and literally price(cost) .. so you may not want CA, or you may want sharper IQ, or whatever .. but how much are you willing to pay for it too.

[ivans75]

Arthur, i certainly wont even if i have the money for it :D

Did few test on leafs towards the blue sky and nope no purple lining on both lenses too. I just did a check again and i think i must have mistaken, the 70-200 doesnt show "correction data available" but it instead it says "no data available" so i must have been day dreaming the other day. The 50 1.4 does say that but then again not a single tests Are showing ca. Must be minor or due to the apsc sensor of 80d. But yup i am happy with these lens i have

[Tannin]

Woaah there! Obviously I didn't explain well enough.

The fact that your camera has pre-loaded correction profiles for one lens and not for a different lens means absolutely nothing except that one of your lenses happens to be one of the 30-odd samples Canon had room to pre-load on the camera. They could equally well have chosen 30 different ones. Whether they pre-load the 30 most popular lenses or the 30 lenses they happened to think of first, I don't know, but I'd put my money on the former. Canon provide correction profiles for all their lenses (not sure how many that is, but it would be hundreds) but there is only room in the camera for about 40 profiles at any one time. So they provide about 30 common ones to get you started and leave space for you to add whichever other Canon lenses you own, if you can be bothered. Most people don't worry about it, as lens correction is a near-universal function in raw converters such as Lightroom anyway.

[arthurking83]

..... Most people don't worry about it, as lens correction is a near-universal function in raw converters such as Lightroom anyway.

This is my understanding of it too! ;)