ef 100-400mm f/4.5-5.6L IS II with eos 80D: Issues?

[YuengLinger]

Hi, all. I finally made peace with the 80D and think it's great for family, street, still life/food photography. (And I swear, one of these days I'm going to take videos!)

However, the one lens it does not behave well with is the ef 100-400mm II. Other lenses work great on it after AFMA'ing, and the 100-400mm works great on my 5DIII.

It is only this specific combination that produces consistently poor IQ, usually with what seems to be the type of CA you'd see with a poor tele-extender, and a lack of sharpness throughout the image.

Neither lens nor body have firmware updates available at this time.

If anybody else is using this same combination, could you please tell me your experience? Thanks!

 

[haggie]

Hi YuengLinger,
I will shortly get the EF 100-400mm L IS II USM to also use with my 80D. So I am quite interested in your findings, and experiences with this combination.

As I do not have the lens yet, I cannot give you any experiences. But I did some 'research' on this combination and here are some links that may give you some images to compare with yours.

https://www.dpreview.com/forums/post/58290619
http://www.fredmiranda.com/forum/topic/1454422/0
http://drkrishi.com/canon-ef-100-400mm-f4-5-5-6l-is-usm-ii-hands-on/comment-page-1/

Just to be clear, the discussion(s) in those thread have no bearing on your question. The images may be just a help for you to identify the problem.

 

[YuengLinger]

Thanks, haggie. In theory, this should be a really fun combo.

I'm going to do a series of tests and comparisons over the weekend, as controlled as I get them. If I can't resolve the issue, I'll be contacting Canon Monday.

In the meantime, any insights from those who have used the combo would be much appreciated.

 

[slclick]

I too await your findings as the 80D is on my 2nd body short list and as I recently picked up the 100-400 Mk2. If there is a true issue that will cross that body off my list until a fix is available.

 

[Plainsman]

I had soft images with the 70D/100-400 Mk I.
The lens though was very sharp with the 750D.
The 70D went back to Canon who seem to have fixed the problem.

 

[YuengLinger]

Hi YuengLinger,
I will shortly get the EF 100-400mm L IS II USM to also use with my 80D. So I am quite interested in your findings, and experiences with this combination.

As I do not have the lens yet, I cannot give you any experiences. But I did some 'research' on this combination and here are some links that may give you some images to compare with yours.

https://www.dpreview.com/forums/post/58290619
http://www.fredmiranda.com/forum/topic/1454422/0
http://drkrishi.com/canon-ef-100-400mm-f4-5-5-6l-is-usm-ii-hands-on/comment-page-1/

Just to be clear, the discussion(s) in those thread have no bearing on your question. The images may be just a help for you to identify the problem.

Haggie--I've been poring over the images in your links this morning, as it's raining out. Very helpful.

One thing that is surprising in many of the images is the low ISO being used. I still tend to expose to the right, which lowers shutter speed, so I go a little higher with ISO. I think I'm learning with the 80D that this is not only unnecessary, but it might be counter productive.

The fredmiranda images are a bit of a mixed bag, but I think I can attain IQ pretty close to the osprey series. The humming birds? On an ef-s 55-250mm STM? Pretty amazing. First, the sharpness and color, second the bokeh. Very good post processing too, and, at ISO 800, I'd guess some selectively applied NR (the background especially!) and sharpening. Just shows that budget lenses in the right hands can do wonders!

AS for the drkrishi shots, these are much closer to the results I'm getting.

I double checked my AFMA yesterday, and my adjustments seem right on a target. I will say that I have been influenced by the precision AF on the 5DIII, and using single point on the 80D for slightly moving birds with narrow necks and relatively small heads is not easy. It is very easy to have focus miss the target--but when it does it, the sharpness is good. More AF points works in some cases, but with lots of brush and reeds right behind birds, not so easy.

I am going to try again with these insights. I've shared my RAWs with a local bird photographer, he sees no hardware problem.

Part of what is going on is likely that I am so used to high ISO on the 5DIII for boosting shutter speeds that I need to be more careful with a cropped 24MP sensor. Images with the 5DIII and the 100-400mm II are consistently clear and sharp, but I am still getting used to the longer effective focal length, and what that does to motion blur--so I tend to go even higher with shutter speed than necessary.

slclick, I've struggled with the combo, trying the 80D some months ago, giving up on it because of problems I was having with the 100-400, but then trying it again because with lenses in the 16mm to 70mm I was seeing excellent results when reviewing portraits. (Note that my first copy of the 80D did have problems with AI Servo, which on the other two, including my current, works very well.)

Here is an 80% 20% crop of an image with no other edits. I've chosen one with people as it has less chance than my bird shots of being influenced by operator error. ISO is high because a moment before I was doing birds in flight.

ISO 1250, f/5.6, 1/2500th, 271mm

 

[haggie]

Thanks for sharing your findings and thoughts, YuengLinger.
With the AFMA being OK I think your assessment that Single Point AF could be of influence is correct.

Earlier today I remembered something I have read and what could also account for 'soft images' with the 80D.
Apparently, due to the 80D's high pixel count and the fact that it is an APS-C sensor, the sharpness decreases above f/8. I cannot reproduce the physics behind it, but I think it had to do with diffraction increasing with smaller aperture (i.e higher aperture values) in combination with a high-pixel count on APS-C.
This means that if you find the aperture was above f/8 in a considerable number of your 'soft images', this may also be at hand.

 

[Pippan]

You may be right about the diffraction with such small pixels. Actually even f/8 is beyond where diffraction makes a difference on an 80D; anything smaller than f/5.6 with that sensor will result in some diffraction. All my lenses on my 80D are softer at f/8 than at f/5.6 (except for the long ends of the 55-250 STM and 18-55 STM).

 

[haggie]

Pippan wrote:

Actually even f/8 is beyond where diffraction makes a difference on an 80D; anything smaller than f/5.6 with that sensor will result in some diffraction.

As I said, I do not know (the details of) the physics behind the decrease in image quality due to diffraction. But when visible diffraction already exists at apertures f/5.6 (i.e. numbers above f/5.6), than that is a shocking fact.

I do not know at what pixels count /pixel size the phenomenon of diffraction improves to levels where for instance the image quality at f/8 still is not soft due to the influence of diffraction.
But increasing the numbers of pixels of a camera like the 80D to a number where the higher pixel count actually causes a decrease in image quality is just plain stupid. Both from the user perspective and from the design perspective.

Up until now I had no real position in the "megapixel debate" that shows up every now and then.
But if this phenomenon (i.e. of soft images due to diffraction already showing at a wide open aperture) is mainly due to the increased pixel count of the sensors in APS-S cameras over time, then in my opinion this decrease in image quality is too high a price to pay for a higher pixel count.

I argue this design choice of one of the sensor's parameters would qualify as a straightforward design error in relevant cameras from Canon's "APS-C system".

 

[YuengLinger]

Pippan wrote:

Actually even f/8 is beyond where diffraction makes a difference on an 80D; anything smaller than f/5.6 with that sensor will result in some diffraction.

As I said, I do not know (the details of) the physics behind the decrease in image quality due to diffraction. But when visible diffraction already exists at apertures f/5.6 (i.e. numbers above f/5.6), than that is a shocking fact.

I do not know at what pixels count /pixel size the phenomenon of diffraction improves to levels where for instance the image quality at f/8 still is not soft due to the influence of diffraction.
But increasing the numbers of pixels of a camera like the 80D to a number where the higher pixel count actually causes a decrease in image quality is just plain stupid. Both from the user perspective and from the design perspective.

Up until now I had no real position in the "megapixel debate" that shows up every now and then.
But if this phenomenon (i.e. of soft images due to diffraction already showing at a wide open aperture) is mainly due to the increased pixel count of the sensors in APS-S cameras over time, then in my opinion this decrease in image quality is too high a price to pay for a higher pixel count.

I argue this design choice of one of the sensor's parameters would qualify as a straightforward design error in relevant cameras from Canon's "APS-C system".

So...The diffraction shows more on the 80D than the 7DII because of more pixels? Still trying to get out into good light for side by side of 5DIII and 80D, which, I think is relevant and "fair," because photographers with a FF often like to have a cropped sensor camera for wildlife and simply for a cheaper backup camera.

 

[Mikehit]

Pippan wrote:

Actually even f/8 is beyond where diffraction makes a difference on an 80D; anything smaller than f/5.6 with that sensor will result in some diffraction.

As I said, I do not know (the details of) the physics behind the decrease in image quality due to diffraction. But when visible diffraction already exists at apertures f/5.6 (i.e. numbers above f/5.6), than that is a shocking fact.

I do not know at what pixels count /pixel size the phenomenon of diffraction improves to levels where for instance the image quality at f/8 still is not soft due to the influence of diffraction.
But increasing the numbers of pixels of a camera like the 80D to a number where the higher pixel count actually causes a decrease in image quality is just plain stupid. Both from the user perspective and from the design perspective.

Up until now I had no real position in the "megapixel debate" that shows up every now and then.
But if this phenomenon (i.e. of soft images due to diffraction already showing at a wide open aperture) is mainly due to the increased pixel count of the sensors in APS-S cameras over time, then in my opinion this decrease in image quality is too high a price to pay for a higher pixel count.

I argue this design choice of one of the sensor's parameters would qualify as a straightforward design error in relevant cameras from Canon's "APS-C system".

The diffraction limit is the point at which the narrower aperture does not cause a corresponding increase in sharpness across the DOF range. The DOF is actually nothing to do with making an image worse.

So, for example, on a given sensor size* you may have a diffraction limit of f11 for a 8MP sensor, but you have a diffraction limit of f 9 for a 20MP sensor. This does not mean that the image at f5.6 is any worse on the 20MP sensor.
It does not even mean that the effect of diffraction makes the image worse at f16. It just means that there is no improvement by narrowing the aperture at that point.

I have spoken to several experienced photographer and they say that diffraction is not even an issue at f22 - it matters only to pixel peepers and measurebators.


*if you compare a APS-C with FF, you may see a difference in the aperture at which diffraction limit is noticeable but that is only because for a given viewing size, you are magnifying the APS-C image more. You would see exactly the same issue if you crop the FF image to the same FOV as APS-C.

 

[Pippan]

Pippan wrote:

Actually even f/8 is beyond where diffraction makes a difference on an 80D; anything smaller than f/5.6 with that sensor will result in some diffraction.

As I said, I do not know (the details of) the physics behind the decrease in image quality due to diffraction. But when visible diffraction already exists at apertures f/5.6 (i.e. numbers above f/5.6), than that is a shocking fact.

I do not know at what pixels count /pixel size the phenomenon of diffraction improves to levels where for instance the image quality at f/8 still is not soft due to the influence of diffraction.
But increasing the numbers of pixels of a camera like the 80D to a number where the higher pixel count actually causes a decrease in image quality is just plain stupid. Both from the user perspective and from the design perspective.

Up until now I had no real position in the "megapixel debate" that shows up every now and then.
But if this phenomenon (i.e. of soft images due to diffraction already showing at a wide open aperture) is mainly due to the increased pixel count of the sensors in APS-S cameras over time, then in my opinion this decrease in image quality is too high a price to pay for a higher pixel count.

I argue this design choice of one of the sensor's parameters would qualify as a straightforward design error in relevant cameras from Canon's "APS-C system".

So...The diffraction shows more on the 80D than the 7DII because of more pixels? Still trying to get out into good light for side by side of 5DIII and 80D, which, I think is relevant and "fair," because photographers with a FF often like to have a cropped sensor camera for wildlife and simply for a cheaper backup camera.

Not because of more pixels per se, but because of the smaller size of the pixels (though of course more pixels on the same sized sensor means they will necessarily be smaller). But you have to look quite closely at the image to notice diffraction softness until you get to very stopped down apertures (I have printed a 12"x8" photo taken at f/29 on my 80D/18-135USM and while I could see the softness, it still looked OK and my sister, whom I gave it to in a frame, thought it looked amazing). I don't know if diffraction's the cause of your issue but it may be something to consider.

 

[Sharlin]

So let's get some facts straight. All else (including aperture) being equal, a 24Mpix sensor will capture more detail than a 20Mpix sensor. Equivalently, after being downsampled to the same output size, a 24Mpix image will be sharper than a 20Mpix one. At the diffraction-limited aperture (DLA; eg. f/5.6 or so with the 80D) diffraction just barely starts to limit the theoretical resolution of a sensor. It takes a considerably narrower aperture before diffraction has completely eaten away the resolution advantage of a higher-megapixel sensor.

 

[privatebydesign]

Diffraction is a constant.

Diffraction is related to aperture alone, not sensor or pixel size.

For a same sized output a smaller sensors output is enlarged more so the same amount of diffraction is more apparent.

Diffraction is not related to pixel size.

Take two sensors the same size, one with 10MP one with 50MP, both will be impacted by diffraction the same amount. Exactly the same amount.

Diffraction will degrade the output of both sensors the same amount.

A sensor with more pixels will always have more detail (assuming anything is in focus).

Smaller pixels reaching a Diffraction Limited Aperture (DLA) at wider apertures is not a bad thing and is not a limitation.

An f5.6 DLA sensor will resolve more detail at f10 than an f10 DLA sensor.

 

[Pippan]

Um, I'm obviously not understanding the term Diffraction Limited Aperture then. Can you explain what it is, why it is not a bad thing and why an f5.6 DLA sensor will resolve more detail at f10 than an f10 DLA sensor? Or point me to some info?

 

[Mikehit]

Um, I'm obviously not understanding the term Diffraction Limited Aperture then. Can you explain what it is, why it is not a bad thing and why an f5.6 DLA sensor will resolve more detail at f10 than an f10 DLA sensor? Or point me to some info?

It is largely an issue in landscape photography.
You have an aperture of f4 and focus on a lamp post 10 away. Only a small region around the lamppost will be in focus.
If you want to have a wider amount of the image in acceptable focus (within the DOF) then you decrease the aperture and more and more of the picture becomes acceptably sharp.
There comes a point where if you open the aperture any more then more of the scene may be within the DOF but no more of the image appears sharp because diffraction is offsetting the increased DOF. In other words diffraction is limiting the benefits of smaller apertures.
Go much further and even the parts that were within the DOF and sharp also start to blur.

So you will see that 'limiting' does not mean 'ruining the quality' as many people make our.

And as said above a couple of times, there can be a very big gap between the aperture at which diffraction limits the benefits of aperture and the point at which is degrades the image.

 

[Mikehit]

Diffraction is a constant.

Diffraction is related to aperture alone, not sensor or pixel size.

For a same sized output a smaller sensors output is enlarged more so the same amount of diffraction is more apparent.

Diffraction is not related to pixel size.

Take two sensors the same size, one with 10MP one with 50MP, both will be impacted by diffraction the same amount. Exactly the same amount.

Diffraction will degrade the output of both sensors the same amount.

A sensor with more pixels will always have more detail (assuming anything is in focus).

Smaller pixels reaching a Diffraction Limited Aperture (DLA) at wider apertures is not a bad thing and is not a limitation.

An f5.6 DLA sensor will resolve more detail at f10 than an f10 DLA sensor.

True. But the complication is that DOF is dependent on image size and viewing distance - print it larger and view from the same distance the DOF decreases. And diffraction blurring becomes more obvious for the same reason.

So if you take the same scene, framed the same way with APS-C and FF and print both to the same size you will see the effect of diffraction sooner with the APS-C image because you are magnifying the image more.
Take two identical size sensors and look at a 10MP age and 50MP both at 100% and the 50MP image will make diffraction more obvious. Look at the m both at the same viewing size and there is no difference.

This is where people get confused and some people who know better talk BS.

 

[Pippan]

Um, I'm obviously not understanding the term Diffraction Limited Aperture then. Can you explain what it is, why it is not a bad thing and why an f5.6 DLA sensor will resolve more detail at f10 than an f10 DLA sensor? Or point me to some info?

It is largely an issue in landscape photography.
You have an aperture of f4 and focus on a lamp post 10 away. Only a small region around the lamppost will be in focus.
If you want to have a wider amount of the image in acceptable focus (within the DOF) then you decrease the aperture and more and more of the picture becomes acceptably sharp.
There comes a point where if you open the aperture any more then more of the scene may be within the DOF but no more of the image appears sharp because diffraction is offsetting the increased DOF. In other words diffraction is limiting the benefits of smaller apertures.
Go much further and even the parts that were within the DOF and sharp also start to blur.

So you will see that 'limiting' does not mean 'ruining the quality' as many people make our.

And as said above a couple of times, there can be a very big gap between the aperture at which diffraction limits the benefits of aperture and the point at which is degrades the image.

Thanks Mike. I do understand those things but I've been under the impression that diffraction, or at least the effects of diffraction, begin to occur at wider apertures the smaller the pixels are, and so I'm confused by Private's explanation that diffraction is dependent only on aperture and that an f/5.6 DLA sensor is sharper at f/10 than an f/10 DLA sensor. I'm happy to read it though as I'm using an 80D with its 3.7μm ϕ pixels. I suppose it must be that the higher resolution provided by the smaller pixels more than offsets the effects of diffraction.

 

[Mikehit]

so I'm confused by Private's explanation that diffraction is dependent only on aperture and that an f/5.6 DLA sensor is sharper at f/10 than an f/10 DLA sensor. I'm happy to read it though as I'm using an 80D with its 3.7μm ϕ pixels. I suppose it must be that the higher resolution provided by the smaller pixels more than offsets the effects of diffraction.

That is in my other post.
PBD is correct that the blur due to diffraction is identical for any sensor. Smaller pixels enable you to zoom in closer and see the diffraction, but within that blur the smaller pixels will still have more detail even if it is smeared a bit.
It is a bit similar to the noise issue where you can have a sensor with more noise but more detail.

 

[BobG]

Hi YuengLinger,
I will shortly get the EF 100-400mm L IS II USM to also use with my 80D. So I am quite interested in your findings, and experiences with this combination.

As I do not have the lens yet, I cannot give you any experiences. But I did some 'research' on this combination and here are some links that may give you some images to compare with yours.

https://www.dpreview.com/forums/post/58290619
http://www.fredmiranda.com/forum/topic/1454422/0
http://drkrishi.com/canon-ef-100-400mm-f4-5-5-6l-is-usm-ii-hands-on/comment-page-1/

Just to be clear, the discussion(s) in those thread have no bearing on your question. The images may be just a help for you to identify the problem.

I had a close look at the drkrishi.com series, and to me it looks as though the focus is in front of the bird, the white brown wagtail image shows this fairly clearly as you can see the earth in front of the bird is in focus. I tend to suspect operator error with some of these photos.