MTFs for Canon R 100-500mm, 600 f/11 and 800 f/11

AlanF

Hands. Face. Space.
CR Pro
Aug 16, 2012
7,345
7,146
IMPORTANT EDIT: Canon has out-of-date MTF charts listed on its USA site for some lenses. Thanks to alerts from subsequent posts, I have gone to the Japanese site https://cweb.canon.jp/eos/ and have revised the charts and conclusions. The opening post has been rewritten so as not to mislead and I have added data from the EF and RF extenders which are available only on the Japanese site. Canon can be a real pain in revealing data, which are hidden away in Japanese. They now no longer give stopped down data and I cannot find whether they use diffraction corrected values, but I guess that they do comparing some Sigma values with and without corrections with Canon's. I hope this compilation is useful to those considering purchases.
 
Last edited:

koenkooi

EOS R
CR Pro
Feb 25, 2015
1,449
1,247
I've heard it mentioned that the new ones are indeed diffraction corrected, but like you I can't find anything official from Canon about that.
 
  • Like
Reactions: tron and AlanF

tron

EOS R5
CR Pro
Nov 8, 2011
4,733
976
How can a lens be diffraction corrected?

I know they advertise this in PP but even so I do not believe that we can reclaim something that isn't there in the first place.
 

AlanF

Hands. Face. Space.
CR Pro
Aug 16, 2012
7,345
7,146
How can a lens be diffraction corrected?

I know they advertise this in PP but even so I do not believe that we can reclaim something that isn't there in the first place.
In theory, if you know the point spread function, you can reconstruct and some people use algorithms such as the Richardson-Lucy https://en.wikipedia.org/wiki/Richardson–Lucy_deconvolution, though this is difficult for something as complex as a photo. From what I have read, it seems that the Canon and other camera programs improve the microcontrast at frequencies that are slightly lower than the diffraction limit (details larger than the diffraction limit) but don't deal with the frequencies that are higher. People claim they can see an improvement but I never do.
 
  • Like
Reactions: Sharlin and tron

AlanF

Hands. Face. Space.
CR Pro
Aug 16, 2012
7,345
7,146
  • Like
Reactions: tron

Billybob

800mm f/11 because a cellphone isn't long enough!
May 22, 2016
124
247
If I'm reading the revised MTFs correctly, it seems the new lens matches or may be marginally better than the old one at 100mm (maybe slightly better contrast) but normal humans won't be able to tell. And the new lens at 500mm looks slightly worst than the old one at 400, which suggests to me that the new lens is likely as good or a bit better than the old one at 400mm and provides an extra 100mm to boot. Correct? Only you can decide whether that difference is worth an extra $500 (or $1000 depending how you measure). For me, 500mm is the minimum focal length for birding (I hate TCs), so the old lens is not a consideration. And if the IQ at 500mm is as close to the old lens at 400mm, the new lens is going to have better IQ at 500mm than any of the long consumer zooms including the Sony 200-600mm
 

Sharlin

EOS R
CR Pro
Dec 26, 2015
1,346
1,150
Turku, Finland
I know they advertise this in PP but even so I do not believe that we can reclaim something that isn't there in the first place.
The data is there, it's just spread around. As AlanF said, if you know (an approximation) of the lens's point spread function, you can get some of it back (although image sensor aliasing irrevocably destroys some of the signal, at least if the resolution (sampling rate) is not sufficient). Deconvolution is a standard technique in scientific imaging, including optical astronomy, where the optical characteristics of instruments, including the PSF, are usually more or less well understood.

Incidentally, this is also why redacting sensitive data like text or people's faces by naive box or Gaussian blur is definitely not a secure technique. With such a well-behaving function getting useful data back via deconvolution is not at all unlikely.
 

Maximilian

The dark side - I've been there
CR Pro
Nov 7, 2013
3,106
1,214
Germany
Would love to have them side by side in RL.
Esp. the AF speed of the new one is something I'd like to look at.
 

AlanF

Hands. Face. Space.
CR Pro
Aug 16, 2012
7,345
7,146
I have calculated the diffraction MTFs for lenses at different apertures. At f/11, MTF10 = 0.93 and MTF30 = 0.79; at f/16, 0.9 and 0.69; and f/22, 0.86 and 0.59. The f/11 lenses are basically close to diffraction limited.
 
  • Like
Reactions: Larsskv

AlanF

Hands. Face. Space.
CR Pro
Aug 16, 2012
7,345
7,146
I couldn't find plots of MTF vs f-number due to diffraction on the internet, so here are my calculated curves in case anyone is interested (MTF10 = 10 and MTF30 = 30 lp/mm as used by Canon, Sigma, Nikon etc as guides for contrast and resolution).

Mtf_Vs_std_f.jpg
 

Bert63

What’s in da box?
CR Pro
Dec 3, 2017
929
2,013
So, in a nutshell, the old tried and true 100-400L II outshoots all the new long lenses? Theoretically within its capabilities?
 

AlanF

Hands. Face. Space.
CR Pro
Aug 16, 2012
7,345
7,146
So, in a nutshell, the old tried and true 100-400L II outshoots all the new long lenses? Theoretically within its capabilities?
It's a little complicated. You have to take focal length into account as well when cropping. I'll add some more information when I have sorted it out to be digestible. The 600/11 is a bit of a wimp with a small front element and the 100-400mm and 100-500mm with extenders will outresolve it. The 800mm has about the same size front element as the two zooms and is competitive at longer focal lengths.
 
  • Like
Reactions: Bert63
Aug 9, 2020
5
2
Well, I have an RF 100-500 on order for my fabulous R5. I have been shooting the R5 with the 100-400 II and an adapter and also the 100-400 II with the 1.4 TC version II. Results are excellent. So when deciding to get the RF 100-500 I compared the above MTF charts of the 100-400 II and 1.4 TC @ 560 with the RF 100-500 @ 500. I am very pleased that the RF lens compares just about identically, with very minor differences, to the EF lens and TC combo @ 500, and is equivalent to the bare 100-400 II within its range. This is huge in my opinion. I will have 100-500 mm without the need to add extenders and essentially the same performance throughout the range. I am really looking forward to seeing what the reality is in every day use.