Keith over at Northlight uncovered a USPTO patent showing various optical designs for catadioptric lenses, better known as mirror lenses. The advantages to mirror lenses are reduced size, weight, and cost at long focal lengths.

Some of the historical disadvantages to mirror lenses tend to be low contrast, fixed aperture, and doughnut bokeh. That's not to say that Canon hasn't improved or eliminated those issues in their designs.

Optical designs that appear in this patent:

  • Canon RF 400mm f/3.6 IS
  • Canon RF 800mm f/5 IS
  • Canon RF 1200mm f/8 IS
  • Canon RF 1200mm f/10.5 IS
  • Canon RF 2000mm f/15 IS

Interestingly, a Canon RF 1200mm f/8 appears on my Canon RF lens roadmap. This patent may actually be part of future consumer products. However, I do have it reported as an L lens, so we'll have to wait and see on that one.

As Canon has shown with the RF 600mm f/11 IS STM and RF 800mm f/11 IS STM, they aren't afraid to bring slower long lenses to the masses. Not everyone can afford or justify the purchase of 5 figure big white lenses.

Some of our articles may include affiliate links. If you purchase through these links, we may earn an affiliate commission at no extra cost to you.

Go to discussion...

Share.

81 comments

  1. Agreed! I'm sure there is a market for compact and lightweight long télephoto lenses. Even at the cost of some image quality aspects.
    No matter what the photo snobs says, the donut bokeh do not kill an image. Especially if it's for a technical field or everything outside fine art.
  2. Donut bokeh is inevitable with this kind of design - but as has been said that's not necessarily awful. If it means the difference between a $900 lens and a $9,000 lens with otherwise equal sharpness, I'll take those donuts!

    It may be able to reduce donutting with lens corrections in software of course.
  3. They seem like interesting consumer lenses, I wonder what the price of the Canon RF 800mm f/5 IS Mirror lens would be compared to the RF 800mm F/11 IS STM? Would it be cheeper because it's mirror lens or more expensive because of the wider aperture?

    What would the Canon RF 2000mm f/15 IS be used for? I'm thinking it could be used for documenting distant birds in midday light on a tripod?
  4. The reason “fast” and “mirror lens” generally don’t go together is because at ratios faster than f/8 or so, the central obstruction gets too big and you lose a lot of contrast. There have been mass-produced Schmidt cassegrain telescopes (SCTs) as fast as f/6.3, but generally they’re f/10. Maksutov cassegrains (MCTs) are typically slower still, f/12-15 or slower.

    That said, the f/5 and f/3.6 are definitely intriguing. I wonder what sort of magic Canon has up their sleeve for these!
  5. Donut bokeh is inevitable with this kind of design - but as has been said that's not necessarily awful. If it means the difference between a $900 lens and a $9,000 lens with otherwise equal sharpness, I'll take those donuts!

    It may be able to reduce donutting with lens corrections in software of course.
    If you could use software to properly "fill" the bokeh hole in post then this would be awesome. I don't know if that's possible with a single lens design which our cameras are.

    This would make an ideal design for any pure-sky astronomy photos, since everything is focused at infinity and there should be no bokeh balls. In fact, some telescopes use this design but don't have the extra corrective lenses (shown in the patent) after the 2nd mirror, shortly before focusing on the sensor. So this would be even better than some consumer telescopes optically. And that rumored future R body with active cooling (to keep the sensor heat & noise low) might be a nice partner for astro work with these lenses. I can see the rumors starting about a hopeful "astro" R... mount camera in the future.

    You know, if Canon has gone this far, they could come out with a small RC (Ritchey-Chretien) lens design that would be amazing.

    What a great option for Canon to bring to the R mount! :)
  6. I have or had a old 600mm Sigma Mirror lens. When I got my EOS R, I put it on and tried it with focus peaking. I did not like the sharpness. I may try again with my R5, it has definitely better IQ. The problem is finding the time. I've owned 5 or 6 of the old mirror lenses, my 500mm Nikkor was fairly good. Most of the old ones were Sigmas.
  7. The interesting thing here is if they will have autofocus or will be manual focus only like the usual eBay special. A mirror lens with IS and AF would be quite interesting and you can always fill the donut holes yourself.
  8. What would the Canon RF 2000mm f/15 IS be used for? I'm thinking it could be used for documenting distant birds in midday light on a tripod?
    Not where I live, in general the atmospheric haze (heat haze) would reduce IQ to a milk bottle bottom most of the time.
  9. The interesting thing here is if they will have autofocus or will be manual focus only like the usual eBay special. A mirror lens with IS and AF would be quite interesting and you can always fill the donut holes yourself.
    Zero chance it would be manual focus.

    As for filling in the holes yourself, good luck with that!
    1620919384311.png
  10. If the lens is primarily used in astrophotography, bokeh is something you would never see.
    Do astro photographers shoot with 400-800mm focal lengths? My impression is they either go wide to catch a big slice of the sky, or long to, say, have the moon fill the frame.

    How much would IS help for astro, when the lens is mounted on a sturdy tripod anyway?
  11. This is very cool and long overdue. Not likely that these are MF, so Cats that are a full stop faster than those of old with IS and AF would be very interesting. I have quite a collection of old mirror lenses and some are very sharp (e.g. 500 mm Minolta MD and 1000mm Nikon). The challenges are always accurate focus and adequate stability, although the R5 IBIS helps a lot with the latter. There is another angle to the timing of this. The Chinese telescope industry is currently in shambles. Between lawsuits, lack of production, and whatever, you simply cannot buy a Meade or Celestron telescope these days. That 2000 f/15 would be quite small and would be useful in telescope land. Note that the biggest of these (the 1200 f/8) is still only 6 inches in diameter and still tiny compared to a big white.
  12. The interesting thing here is if they will have autofocus or will be manual focus only like the usual eBay special. A mirror lens with IS and AF would be quite interesting and you can always fill the donut holes yourself.
    The patent does have IS. Minolta made a 500mm f/8 mirror reflex lens with AF that works on some Sony bodies. I prefer doughnuts,

  13. Hmmm.. strange. First thought was a standard lens, but on seeing it was a mirror, disappointing. Useful for astro, focal lengths tend to be 1m or more. But IS? Long astro lenses will be mounted on heavy duty trackers which don't need it. And mmmm... donuts!
    Not unless Topaz can come up with an AI donut filler?

Leave a comment

Please log in to your forum account to comment