Canon 70-200L IS USM vs 100-400 IS

[Alex_B]

A 3X fixed aperture zoom is theoretically going to outperform a 4X variable aperture zoom. My experience with 1.4X teleconverters is more positive than my experience with 4X zooms. But, like I said, I'm not a Canon guy so I'll bow out.

But even with Canon you are right here .. both lenses are supposedly very good pro lenses. But the 100-400 L is built for versatility, and the 70-200 L 2.8 is more built for image quality.
The 100-400 L is a bit soft at the 400mm end, at least the one I had in my hands. That is why I decided for the 300mm L 4. Also got a TC 1.4x now and will post images once I find some time to do so

From what I heard, the 70-200 L 2.8 is supposedly one of Canon's best telephoto zoom lenses. Of course primes are an alltogether different story.

 

[EOS_JD]

want to buy mine? ... but I won't sell it

Also, regarding camera filed of view (that makes your composition) and regarding the effect of camera shake relative to the size of your image (this makes your image quality), a 200mm lens on 1.6 crop behaves just as a 300mm lens on 35mm film or full frame sensors: Camera shake of the same amplitude will cause the same motion blur in the final print (if print size is the same from the crop camera and the full frame).

I understand what you are trying to say but sorry I can't agree.

That's like saying to a 5D owner better shooting faster than 1/200th on your 200mm lens just in case you want to crop your image. that's all the crop cameras do is provide a crop from what you'd get from a FF sensor.

 

[EOS_JD]

here's an image from a 300f4L IS & 1.4x (sorry I don't shoot wildlife much)!

 

[fmw]

You are both right. Alex is right in terms of angle of view and EOS JD is right in terms of focal length. It can get to be quagmire of semantics sometimes.

Alex, my comments weren't based on any experience with the particular lenses. They were based on a general knowledge of optics and the limitations of lens design that all lens designers face regardless of the name of their employer. We don't always have to have experience with the lens to have an idea how it should perform.

 

[lostprophet]

I've owned 2 100-400mm lenses and both died on me, would never own another one in a million years. Now have the 70-200mm F2.8 IS and I love it. I get sharper shots with a 2x on it than I did with the 100-400 as well

 

[Alex_B]

I understand what you are trying to say but sorry I can't agree.

That's like saying to a 5D owner better shooting faster than 1/200th on your 200mm lens just in case you want to crop your image. that's all the crop cameras do is provide a crop from what you'd get from a FF sensor.

I agree my posting is only valid if you do not crop but stick to original image dimensions.

Usually the smaller sensor also provides a higher pixel density. But all this does not matter in this discussion anyway, since we talk about final prints.

Let us further assume that the pixel-resolution of both the 1.6 crop and the FF sensor are sufficient for say a 30cmx20cm print. With sufficient I mean that megapixels are not the limiting factor to the quality here.

If this is the same, then camera shake of a the same amplitude on the 300mm lens translates into a smaller motion blur amplitude on the sensor than with a 400mm lens.
However, since we combine the 300mm lens with the smaller sensor and the 400mm lens with the larger sensor, then you get the smaller absolute motion blur on the smaller sensor and the larger blur on the larger sensor.
If we now use the full image from the sensors in both cases to produce that 30x20 print, then motion blur in the final images will be of similar amplitude.
The magnification from the smaller sensor to 30x20 is larger than the magnification from the larger sensor to 30x20, that means also the motion blur is magnified more.

If you shoot with a full frame camera with a say 200mm lens, and you plan to crop as you suggest (which means more magnification for the final print), then you should indeed use shorter exposure times to reduce motion blur resulting from camera shake.

BTW, I do exclusively shoot full frame, and I crop a lot. So this is not just theory.

 

[Alex_B]

here's an image from a 300f4L IS & 1.4x (sorry I don't shoot wildlife much)!

Very nice image by the way

Which TC do you use? Canon or Kenko or anything else?

 

[EOS_JD]

Thank Alex. It was the Canon 1.4x. I used to have a horrid Kenco 2x and didn't like it at all.

I'm still not overly convinced although I do see your theory.

How often do you print that big though? My largest images are generally around 19x13 (A3+ of my R2400) and I've never had issues with sharpness. My last was a 105mm shot at 1/20th sec and it looks amazing blown up.

I accept though that in some shots it would make a difference but not so sure I'd see it at my size of images.

Cheers
Jim

 

[usayit]

I've owned 2 100-400mm lenses and both died on me, would never own another one in a million years.

In what way did the 100-400mm lenses fail? Just curious because this is the first time I've heard anyone posting that high of failure rate.

Btw... just for clarification... canon 1.4x with 100-400 will still maintain AF on a 1 series body. I'm not sure if that includes the 5D.

 

[EOS_JD]

Yes I should have stated that. Only 1 series bodies keep AF up to f8. All the rest it's f5.6 I believe.

 

[lostprophet]

In what way did the 100-400mm lenses fail? Just curious because this is the first time I've heard anyone posting that high of failure rate.

Btw... just for clarification... canon 1.4x with 100-400 will still maintain AF on a 1 series body. I'm not sure if that includes the 5D.

1st one I had became very loose on zooming and Canon UK couldn't find a fault with it on the 3 times it was sent to them. The 2nd one went out on a cold day and a week later fungus had started, I'd used many other lenses on colder / wetter days and never had a problem. And that just put me off that lens. That said the current versions are different from the original, they have a much smoother focussing movement.

As for focussing at F8 it can be "made" to work on non series 1 bodies with a bit of tape, but I wouldn't recommend it just in case something goes wrong

 

[Alex_B]

Thank Alex. It was the Canon 1.4x. I used to have a horrid Kenco 2x and didn't like it at all.

I have not experience with the 2x, but I guess there it was not Kenko to blame but thew fact that it was 2x ... I have the latest 1.4x from Kenko, and it appears to have more coner sharpness that the Canon euqivalent on full frame. Following tests, this does not really matter on crop sensors. If you use the Canon TC and are happy this would support those tests.

How often do you print that big though?

Oh, you don't even neet to print on paper, you can even see it when you "print" on screen and zoom in a bit. Looking at the actual pixel resolution will mislead you though. But you could realise it if you just look at one quarter of the image filling your screen in both cases.

This all becomes more visible and more extreme of course the more you crop. If you crop to just 10% of the visible area, it will be quite a large effect. In particular if in the uncropped version you are already at the borderline.

My largest images are generally around 19x13 (A3+ of my R2400) and I've never had issues with sharpness. My last was a 105mm shot at 1/20th sec and it looks amazing blown up.

OK, now I see the problem why you are having trouble believing me

the inverse focal length as the shutter speed for telephoto of course is only rule of thumb for a specific print size looked at from a specific distance, assuming a typical amount of camera shake. I just guess you have a hand more steady than the average guy, and also IS will of course increase feasible shutter speeds. So my guess is that you are always way off any problematic case, so you cannot see the difference.

Of course, if you still do not believe in the principle of what I say (not talking of absolute numbers), then I could try to illustrate it with some sketch drawn to scale for both cases, small and large sensor and the two lenses. Of course that would take some time as I'd have to start a graphics programme and start drawing

But I have the feeling you understand what I meant anyway ...