Patent: Canon EF 50mm f/1.4

[chromophore]

Funny no one is going to pick up the dead-easy edge and just stick IS in a 50mm prime. Sure you can do without it (currently everyone does) but given the consumer appetite for it evidenced in camera forums, someone could sell a lot of units if they would just do it. If you can do it economically in a $199 kit lens you can do it anywhere.

Uhh.. I'm quite sure that it's A BIT different thing to put IS to a variable aperture zoom that is at max f/3.5 in wide end and f/5.6ish at 50mm, than what it is to put it into 50mm f/1.4... So far closest is sony 50mm f/1.8 IS for crop, IIRC.

A decade ago I've been dreaming of a 50mm Image Stabilizer. After I replaced my old Canon F1.4 by a Sigma Art, I'm no longer sure I would buy one.

I can still dream, but the reality is that a Canon 50mm Image Stabilizer with the same picture quality of the Sigma Art would cost more than $ 1500.

"Why is it that there are no optically image-stabilized lens designs at f/1.4?"

This has to do with a number of difficulties related to the way lenses focus light at small f-numbers.

Lenses that gather more light do so by increasing the cross-sectional area through which light rays are refracted. When we look through the rear of a lens, we are seeing the exit pupil, the diameter of which is inversely proportional to the f-number. The exit pupil diameter must be sufficiently large to achieve a particular light-gathering ability, regardless of what is going on anywhere else in the lens design.

The way optical image stabilization works is to use a group of corrective lens elements that are free to react to movement of the system to compensate for vibrations in real-time. This correction works in large part because these movements are small and the degree of displacement required is also small; however, another factor contributing to resultant image quality is that, because these IS elements are always present in the optical path, they must also permit an acceptably uniform degree of correction across the image plane.

Consequently, the problem with adapting IS technology to very fast apertures is that, say, at f/1.4, the correction is much more difficult to achieve than at f/2 or f/2.8, where in the latter case (1) the image-forming marginal rays are not so oblique; (2) the aberrations dependent on image height are not as severe; (3) the size of the IS group does not need to be as large (and therefore have less inertia). Remember, at f/1.4, the exit pupil diameter is sqrt(2) times as large as at f/2, and the required IS group must therefore be at least twice as large in area and at least 2.8 times as massive (conservatively).

It doesn't make a lot of sense optically to try to implement IS in a very fast lens, because the challenge of correcting marginal ray aberrations adequately even without such a group is difficult enough; requiring that a hypothetical IS group operating at f/1.4, at least 3x as heavy, to maintain image quality, is not economical. Does that mean it isn't possible? No. But is such a design going to deliver excellent results, or do it at a reasonable cost? I doubt it.

 

[cayenne]

Me?

I'm gonna hold out, and save my pennies for the rumored new version of the 50L f/1.2 lens!!

;D

cayenne

 

[ajfotofilmagem]

Me?
I'm gonna hold out, and save my pennies for the rumored new version of the 50L f/1.2 lens!!
;D
cayenne

I suggest put their money into a retirement fund ... ???

May be many years before we see a new Canon 50mm F1.2 :

 

[PureClassA]

Honestly, with the 35L Mk II just announced, I don't think it will be another 12 months before we see a new 50L. One tends to bode well with the other with lens makers, as well as Canon.

Me?
I'm gonna hold out, and save my pennies for the rumored new version of the 50L f/1.2 lens!!
;D
cayenne

I suggest put their money into a retirement fund ... ???

May be many years before we see a new Canon 50mm F1.2 :

 

[bluenoser1993]

Is that front surface actually concave or just plano? I stared at it for a long time and can't decide.

Hold a straight edge to it.

 

[ajfotofilmagem]

Honestly, with the 35L Mk II just announced, I don't think it will be another 12 months before we see a new 50L. One tends to bode well with the other with lens makers, as well as Canon.

I'm gonna hold out, and save my pennies for the rumored new version of the 50L f/1.2 lens!!
cayenne

I suggest put their money into a retirement fund ... ???
May be many years before we see a new Canon 50mm F1.2 :

You say that the Canon 35L Mk II just announced? I lost an official announcement, or just saw a rumor that such a lens is ready to be produced?

As for the Canon 50mm... the F1.4 model needs an update urgently. And after that come a new F1.2 model.

 

[zim]

Would in-body IS therefore not be an advantage with faster lenses?

Edit: I don't mean as a replacement but a switch on if lens doesn't have IS , I believe in-lens IS has advantages

 

[Solar Eagle]

"Why is it that there are no optically image-stabilized lens designs at f/1.4?"

This has to do with a number of difficulties related to the way lenses focus light at small f-numbers.

Lenses that gather more light do so by increasing the cross-sectional area through which light rays are refracted. When we look through the rear of a lens, we are seeing the exit pupil, the diameter of which is inversely proportional to the f-number. The exit pupil diameter must be sufficiently large to achieve a particular light-gathering ability, regardless of what is going on anywhere else in the lens design.

The way optical image stabilization works is to use a group of corrective lens elements that are free to react to movement of the system to compensate for vibrations in real-time. This correction works in large part because these movements are small and the degree of displacement required is also small; however, another factor contributing to resultant image quality is that, because these IS elements are always present in the optical path, they must also permit an acceptably uniform degree of correction across the image plane.

Consequently, the problem with adapting IS technology to very fast apertures is that, say, at f/1.4, the correction is much more difficult to achieve than at f/2 or f/2.8, where in the latter case (1) the image-forming marginal rays are not so oblique; (2) the aberrations dependent on image height are not as severe; (3) the size of the IS group does not need to be as large (and therefore have less inertia). Remember, at f/1.4, the exit pupil diameter is twice as large as at f/2, and the required IS group must therefore be at least twice as large in area and at least 2.8 times as massive (conservatively).

It doesn't make a lot of sense optically to try to implement IS in a very fast lens, because the challenge of correcting marginal ray aberrations adequately even without such a group is difficult enough; requiring that a hypothetical IS group operating at f/1.4, at least 3x as heavy, to maintain image quality, is not economical. Does that mean it isn't possible? No. But is such a design going to deliver excellent results, or do it at a reasonable cost? I doubt it.

Thanks for explaining this.

 

[sleepnever]

I had a Canon 50 1.8 II and upgraded to the 50 1.4. Still wasn't doing it for me. Rented the Sigma 50 ART and fell in love, bought it. Is it heavier than the 50 1.4 from Canon? Yes it is. However the balance on my 5D3 is just right, so I don't mind. Also, the IQ is amazing.

So Canon would have to make the IQ at all fstops better than the Sigma and make it cost equal to or less. Otherwise the only advantage would be AF performance.

 

[alliumnsk]

Funny no one is going to pick up the dead-easy edge and just stick IS in a 50mm prime. Sure you can do without it (currently everyone does) but given the consumer appetite for it evidenced in camera forums, someone could sell a lot of units if they would just do it. If you can do it economically in a $199 kit lens you can do it anywhere.

Just buy a ILCE7RM2. It has IBIS, plus, the bokeh won't be clipped as much as on DSLR.

 

[Hjalmarg1]

Is 100mm compact ? Compared to a Sigma of Otus yep... It's going to be bigger than the current model, I smell the hint of a red ring !

Yes, compact compared to Sigma Art and Zeiss Otus.

It apears to be a RED ring lens rather than a consumer grade lens. 100mm is as long as the Sigma 1.4.
Nikon hs only two 50mm lenses f1.4 & f1.8....Bye-bye 50mm f1.2L

 

[Maximilian]

Is 100mm compact ? Compared to a Sigma of Otus yep... It's going to be bigger than the current model, I smell the hint of a red ring !

Yes, compact compared to Sigma Art and Zeiss Otus.

It apears to be a RED ring lens rather than a consumer grade lens. 100mm is as long as the Sigma 1.4.
Nikon hs only two 50mm lenses f1.4 & f1.8....Bye-bye 50mm f1.2L

Hi guys!

Sorry, but you make the same mistake I made some weeks ago with the 50/1.8 STM patent.

The 100.16 mm ist NOT the mechanical length of the lens barrel but the length of the optical formula, that means from front element to the image plane/sensor.
The lens barrel is about 6 cm long as stated by me @ reply #12 below.

So if I read the patent right (this time) the lens will have a lengh of approx. 62 mm (100.16-38.10).
So some 22% bigger compared to the old 50mm f/1.4 USM which has 50,5 mm.
And of course much smaller than the Sigma A (99,9mm).

This still leaves the question if this will be a 50/1.4 USM successor and therefore non L or not.
But I suppose Canon wil stay with their three 50 mm lenses (and the 50 mm Macro).

Edit: Because the market for a 400 - 800 €/$ 50 mm lens is too big to be ignored. And a L 50 mm will be way north 1k bucks not matter what aperture it'll have.

 

[JonAustin]

It seems "glaringly obvious" to me that Canon needs to produce IS-equipped updates to their 50- and 85mm non-L primes, to continue the 24/2.8 IS, 28/2.8 IS and 35/2 IS line that they released in 2012. But then, I don't have access to Canon's market research data, financials or product development roadmap.

The latest zooms are so good that I don't use primes often, but I have the 35/2 IS, 50/2.5 CM and 100/2.8L IS, and I enjoy using them when the need arises. I would love to replace the razor-sharp but "buzzy" little 50/2.5 with a 50/1.x-2.0 IS lens with all the technological goodies that Canon puts into its latest non-L primes.

 

[Gordonium]

Funny no one is going to pick up the dead-easy edge and just stick IS in a 50mm prime. Sure you can do without it (currently everyone does) but given the consumer appetite for it evidenced in camera forums, someone could sell a lot of units if they would just do it. If you can do it economically in a $199 kit lens you can do it anywhere.

Uhh.. I'm quite sure that it's A BIT different thing to put IS to a variable aperture zoom that is at max f/3.5 in wide end and f/5.6ish at 50mm, than what it is to put it into 50mm f/1.4... So far closest is sony 50mm f/1.8 IS for crop, IIRC.

A decade ago I've been dreaming of a 50mm Image Stabilizer. After I replaced my old Canon F1.4 by a Sigma Art, I'm no longer sure I would buy one.

I can still dream, but the reality is that a Canon 50mm Image Stabilizer with the same picture quality of the Sigma Art would cost more than $ 1500.

"Why is it that there are no optically image-stabilized lens designs at f/1.4?"

This has to do with a number of difficulties related to the way lenses focus light at small f-numbers.

Lenses that gather more light do so by increasing the cross-sectional area through which light rays are refracted. When we look through the rear of a lens, we are seeing the exit pupil, the diameter of which is inversely proportional to the f-number. The exit pupil diameter must be sufficiently large to achieve a particular light-gathering ability, regardless of what is going on anywhere else in the lens design.

The way optical image stabilization works is to use a group of corrective lens elements that are free to react to movement of the system to compensate for vibrations in real-time. This correction works in large part because these movements are small and the degree of displacement required is also small; however, another factor contributing to resultant image quality is that, because these IS elements are always present in the optical path, they must also permit an acceptably uniform degree of correction across the image plane.

Consequently, the problem with adapting IS technology to very fast apertures is that, say, at f/1.4, the correction is much more difficult to achieve than at f/2 or f/2.8, where in the latter case (1) the image-forming marginal rays are not so oblique; (2) the aberrations dependent on image height are not as severe; (3) the size of the IS group does not need to be as large (and therefore have less inertia). Remember, at f/1.4, the exit pupil diameter is sqrt(2) times as large as at f/2, and the required IS group must therefore be at least twice as large in area and at least 2.8 times as massive (conservatively).

It doesn't make a lot of sense optically to try to implement IS in a very fast lens, because the challenge of correcting marginal ray aberrations adequately even without such a group is difficult enough; requiring that a hypothetical IS group operating at f/1.4, at least 3x as heavy, to maintain image quality, is not economical. Does that mean it isn't possible? No. But is such a design going to deliver excellent results, or do it at a reasonable cost? I doubt it.

Pretty sure the diameter of entrance pupil is inversely proportional to the f-number. But is that necessarily true for exit pupil?

 

[cayenne]

Me?
I'm gonna hold out, and save my pennies for the rumored new version of the 50L f/1.2 lens!!
;D
cayenne

I suggest put their money into a retirement fund ... ???

May be many years before we see a new Canon 50mm F1.2 :

I thought I read here on CR that the new 50L was rumored to maybe coming out first Quarter or so of 2016?

C

 

[chromophore]

Pretty sure the diameter of entrance pupil is inversely proportional to the f-number. But is that necessarily true for exit pupil?

Yes. It is a necessary condition, but it is not strictly correct at very fast apertures. The actual relationship is trigonometric: D[exit]/(2*P'I) = tan(theta/2), where D[exit] is the exit pupil diameter, P'I is the distance from the exit pupil to the image plane, and theta is the angle of the incident light cone at the image center. But in the small-angle approximation, tan(theta/2) is approximately theta/2.

 

[Mr_Canuck]

Funny no one is going to pick up the dead-easy edge and just stick IS in a 50mm prime. Sure you can do without it (currently everyone does) but given the consumer appetite for it evidenced in camera forums, someone could sell a lot of units if they would just do it. If you can do it economically in a $199 kit lens you can do it anywhere.

I've always said the very same thing. Why not IS in everything? And why not a 50/1.4 IS just like the 35? Make it better than the old one (not super hard), and keep it mid-level. The 1.2 is for L and thousands.

 

[Gordonium]

Pretty sure the diameter of entrance pupil is inversely proportional to the f-number. But is that necessarily true for exit pupil?

Yes. It is a necessary condition, but it is not strictly correct at very fast apertures. The actual relationship is trigonometric: D[exit]/(2*P'I) = tan(theta/2), where D[exit] is the exit pupil diameter, P'I is the distance from the exit pupil to the image plane, and theta is the angle of the incident light cone at the image center. But in the small-angle approximation, tan(theta/2) is approximately theta/2.

Fair enough I can see how you get that relationship, but how does that relate to the f number?

 

[chromophore]

Pretty sure the diameter of entrance pupil is inversely proportional to the f-number. But is that necessarily true for exit pupil?

Yes. It is a necessary condition, but it is not strictly correct at very fast apertures. The actual relationship is trigonometric: D[exit]/(2*P'I) = tan(theta/2), where D[exit] is the exit pupil diameter, P'I is the distance from the exit pupil to the image plane, and theta is the angle of the incident light cone at the image center. But in the small-angle approximation, tan(theta/2) is approximately theta/2.

Fair enough I can see how you get that relationship, but how does that relate to the f number?

Please refer to the following article from Zeiss:

http://www.zeiss.com/content/dam/Photography/new/pdf/en/cln_archiv/cln35_en_web_special_bokeh.pdf

This should be a definitive and authoritative source for your question.

 

[Infraspace]

I have no idea what this is going to cost, BUT I WANT IT!