Fast Full Frame Unique Zoom Lenses

[AJ]

Yes, you got it. the light bounces in between the two elements and magic happens.

Whoa - this is making my head spin. It seems that for decades the objective was to reduce the reflection coefficients by means of coatings. But here we are talking about surfaces that are reflective on purpose.

I wonder what it does? I see from the one implementation that rays are shifted towards the optical axis. Wideangle gets the strongest shift. Maybe this simplifies the design beyond these exotic elements, allowing more possibilities (e.g. faster zooms). It's interesting that this technology is introduced in zooms, not primes.

I wonder what the "HM" stands for, in the diagrams?

In any case, this seems revolutionary. I'd love to read a whitepaper on this.

 

[neuroanatomist]

I wonder what the "HM" stands for, in the diagrams?

Half Mirror.

 

[AJ]

Makes me wonder how it affects bokeh

 

[neuroanatomist]

Makes me wonder how it affects bokeh

The users’ desire for good quality OOF areas of the image is called out in the patent.

 

[mattrobo]

Sigh. Because physics. Wide and standard lens designs are limited by the image circle diameter, telephoto designs are not. There is no point in a telephoto lens ‘for crop cameras’ since such a lens would work fine on FF. That’s why there are no such lenses.

Look at the Oly/OM 150-400mm, for example. It’s for m4/3 (2x crop) but it’s about the same size as the Canon 100-500 (and if OM made a FF camera, the lens would work on it).

As for wide lenses not making money, I really don’t get why some people seem to think they know better than Canon what lenses people will buy. I mean, Canon has led the ILC market for over two decades and dominates it today…but you know more about what lenses they need to make? LOL. You get to decide what lenses you want to buy, but Canon doesn’t care what you personally want.

 

[tpatana]

It's not that Canon didn't tell us, it's that Richard didn't pass along the full explanation. He hints at it, though, right at the start: "I had previously discussed these as quarter-wave optical designs...," in reference to the quarter wave plates (QWP) that convert linearly polarized light to circularly polarized light and back.

Details are in the patent, and the light manipulation you're missing is polarization, more than once in the optical path and both linear (POL) and circular (QWP2, QWP1), plus selective reflection/transmission.

View attachment 228270

Thanks, this now makes sense for most items. The items that make my head scratch are:

The sensor only receives horizontal polarization component of the light, so you basically have polarization filter inside your lens which you cannot remove.

But furthermore, switching from linear to circular (or back) should reduce light by 1 stop, so you wouldn't get much light in the end. Counting the light path for conversions: E D B B B

C (HM) probably also cut 1 stop on that first pass but I'm not sure so I'm leaving it out the calculation.

So that's 5 stops (or 6 if count HM) of light right there. Maybe I should go read the patent as this is interesting although I'm not yet sure if it's practical.

Also, QWP are fairly frequency sensitive so I'm wondering how much different the far ends of light spectrum will behave on that.