April 26, 2018, 08:19:47 AM

Author Topic: A Prototype Full Frame Mirrorless From Canon Exists [CR1]  (Read 24811 times)

AvTvM

  • Canon EF 400mm f/2.8L IS II
  • *********
  • Posts: 3165
Re: A Prototype Full Frame Mirrorless From Canon Exists [CR1]
« Reply #120 on: March 16, 2018, 10:15:14 AM »
AFAIK, that was possible because film behaves differently than digital sensors. Film chemistry will react with light hitting it at any angle, while digital sensors need light to hits close to perpendicular, and that requires more distance between the back element and the sensor.

... all it requires is a little bit more thought and innovative solutions to the design of microlenses over digital imaging sensor. Leica has finally solved it by now (after messing around for about 10 years).

canon rumors FORUM

Re: A Prototype Full Frame Mirrorless From Canon Exists [CR1]
« Reply #120 on: March 16, 2018, 10:15:14 AM »

Vilmos

  • PowerShot SX60 HS
  • **
  • Posts: 3
  • 5D Mk IV
Re: A Prototype Full Frame Mirrorless From Canon Exists [CR1]
« Reply #121 on: March 16, 2018, 10:34:39 AM »
I agree with and accept all your arguments - Talys, Antono Refa, Isaacheus.
So the conclusion is that the most probable outcome could be an FF MILC with unchanged EF mount, unchanged flangeback distance, and a continuation of the current lens line.
Which I would be very, very happy with.
However, those workhorse primes and zooms that are currently awkwardly large, probably won’t get any smaller.

Throwing out the mirror and replacing the prism with an EVF do not seem so big deal, I wonder what takes this much time for Canon? Further refining DPAF? Sensor warming? Battery life?

Talys

  • EOS 5DS R
  • ******
  • Posts: 1452
  • Canon 6DII
Re: A Prototype Full Frame Mirrorless From Canon Exists [CR1]
« Reply #122 on: March 16, 2018, 12:44:51 PM »
I agree with and accept all your arguments - Talys, Antono Refa, Isaacheus.
So the conclusion is that the most probable outcome could be an FF MILC with unchanged EF mount, unchanged flangeback distance, and a continuation of the current lens line.
Which I would be very, very happy with.
However, those workhorse primes and zooms that are currently awkwardly large, probably won’t get any smaller.

Throwing out the mirror and replacing the prism with an EVF do not seem so big deal, I wonder what takes this much time for Canon? Further refining DPAF? Sensor warming? Battery life?

Not to take credit for it, because it was someone else's idea (maybe ahsanford mentioned it?)  but the smart move, I think, would be EF-S.  That way, you could mount an EF lens, or an APSC EF-S lens, and if you do the latter, the camera goes automatically into crop mode.  To their credit, Sony does a great job with this. 

It opens up some really good, cheap EF-S lenses like 10-18, which, in crop mode gives you about a 16-35 equivalent field of view (especially as it pertains to things like tripod sunsets and lake shots, where you're going to stop down anyways), but at a bargain of a price, and at a good enough IQ for a lot of people who only casually use that wide FL's.  And, it opens up some great all purpose lenses that aren't heavy, like EFS 18-135, which is a nice video lens and walkaround lens, especially if you're not too worried about the corners or obsessed with the best possible IQ.

I think there is another benefit of a recessed sensor -- I am using an A7R3 right now, and the sensor is frighteningly close to the mount.  When I do lens swaps in the field with Canon lens/bodies, I do them pretty quick and it's a "snap-snap" type of deal.  I'm a little scared on the Sony that if I am slightly off center, I'll scratch the massive sensor that's right there.  Perhaps an unfounded fear; I don't know if anyone has ever done that.

Antono Refa

  • EOS 7D Mark II
  • *****
  • Posts: 551
Re: A Prototype Full Frame Mirrorless From Canon Exists [CR1]
« Reply #123 on: March 17, 2018, 06:46:40 AM »
AFAIK, that was possible because film behaves differently than digital sensors. Film chemistry will react with light hitting it at any angle, while digital sensors need light to hits close to perpendicular, and that requires more distance between the back element and the sensor.

... all it requires is a little bit more thought and innovative solutions to the design of microlenses over digital imaging sensor. Leica has finally solved it by now (after messing around for about 10 years).

Could you, please, provide a source to back this claim, preferably hard data rather than marketing?

Don Haines

  • Canon EF 800mm f/5.6L IS
  • ***********
  • Posts: 6760
  • posting cat pictures on the internet since 1986
Re: A Prototype Full Frame Mirrorless From Canon Exists [CR1]
« Reply #124 on: March 17, 2018, 11:40:36 AM »
AFAIK, that was possible because film behaves differently than digital sensors. Film chemistry will react with light hitting it at any angle, while digital sensors need light to hits close to perpendicular, and that requires more distance between the back element and the sensor.

... all it requires is a little bit more thought and innovative solutions to the design of microlenses over digital imaging sensor. Leica has finally solved it by now (after messing around for about 10 years).

Could you, please, provide a source to back this claim, preferably hard data rather than marketing?
Just thinking here, no hard data.......


Lenses, for both film and digital, are designed to focus on the film plane. The longer the focal length, the more perpendicular the light is. Conversely, the wider angle the lens is, the greater the angle that light will hit the sensor, and once you get past a certain angle, the light does not make it to the bottom of the well.

There are two ways to deal with this, make the well shallower (BSI) and redirecting the light. Ideally both.....

Just about everyone uses micro lenses to deal with this and direct the photons down into the well....  if your camera design has a short flange distance, you would be more aggressive with your micro lenses, so in the end result, does it really make a difference?

As an aside, look at the new designs for prime lenses in the 50mm and under range. Almost all of them are BIG! My assumption is that this is give more space to bend the light, allowing shallower angles and lesser aberrations. If we are going FF for image quality, and the lenses required for that are also big, and the required distance between the last element and the sensor is also big, then it becomes a question of do we put that distance in the body, or do we put it in the lens? Either way, we are not talking small.....
« Last Edit: March 17, 2018, 11:44:10 AM by Don Haines »
The best camera is the one in your hands

AvTvM

  • Canon EF 400mm f/2.8L IS II
  • *********
  • Posts: 3165
Re: A Prototype Full Frame Mirrorless From Canon Exists [CR1]
« Reply #125 on: March 17, 2018, 01:40:07 PM »
AFAIK, that was possible because film behaves differently than digital sensors. Film chemistry will react with light hitting it at any angle, while digital sensors need light to hits close to perpendicular, and that requires more distance between the back element and the sensor.

... all it requires is a little bit more thought and innovative solutions to the design of microlenses over digital imaging sensor. Leica has finally solved it by now (after messing around for about 10 years).

Could you, please, provide a source to back this claim, preferably hard data rather than marketing?

not sure whether this qualifies as "marketing material" or not ... but was what i could find on the quick ... http://gmpphoto.blogspot.co.at/2016/01/the-future-of-sensor-technology-at-leica.html

there are quite a number of possibilities to handle small flange focal distance well ... from lens formula to sensor design ... including thickness of sensor cover glass / AA filters and micro lens array ... shape and positioning of micro lenses, eg using "tilted" microlenses towards sensor edge and/or progressively varying microlenses' focal lens between center of sensor and edges etc.  ...

Sony FE lenses with lots of empty barrel space towards camera are not a "natural law" but solely a consequence of Sony's very poor choice of lens mount ... E-mount was designed and optimized for APS-C sensor image circle .. but then pressed into forced FF service as an aftherthought ...

it is definitely possible to build less complex, smaller sized and lower priced lenses for FF mirrorless systems than Sony FE lenses. AT least for the most frequently used focal lengths. For longer tele lenses, physical size is determined by entrance pupil ... which in turn defines diameter of front element. But anything up to about 100mm focal length could be built nice and small and with excellent IQ at affordable prices.

Antono Refa

  • EOS 7D Mark II
  • *****
  • Posts: 551
Re: A Prototype Full Frame Mirrorless From Canon Exists [CR1]
« Reply #126 on: March 17, 2018, 02:13:35 PM »
AFAIK, that was possible because film behaves differently than digital sensors. Film chemistry will react with light hitting it at any angle, while digital sensors need light to hits close to perpendicular, and that requires more distance between the back element and the sensor.

... all it requires is a little bit more thought and innovative solutions to the design of microlenses over digital imaging sensor. Leica has finally solved it by now (after messing around for about 10 years).

Could you, please, provide a source to back this claim, preferably hard data rather than marketing?

not sure whether this qualifies as "marketing material" or not ... but was what i could find on the quick ... http://gmpphoto.blogspot.co.at/2016/01/the-future-of-sensor-technology-at-leica.html

It's a good explanation, I'm not waving it out of hand.

What I was hoping for is hard measurements, ideally comparing vignetting on film & the new sensor, and whether it loses light when aperture is wide open.

there are quite a number of possibilities to handle small flange focal distance well ... from lens formula to sensor design ... including thickness of sensor cover glass / AA filters and micro lens array ... shape and positioning of micro lenses, eg using "tilted" microlenses towards sensor edge and/or progressively varying microlenses' focal lens between center of sensor and edges etc.  ...

I'm all for it. I traded in the EF 15mm f/2.8 when I bought the EF 11-24mm f/4L, and miss it a bit and tempted to buy the 8-15mm f/4L.

it is definitely possible to build less complex, smaller sized and lower priced lenses for FF mirrorless systems than Sony FE lenses. AT least for the most frequently used focal lengths. For longer tele lenses, physical size is determined by entrance pupil ... which in turn defines diameter of front element. But anything up to about 100mm focal length could be built nice and small and with excellent IQ at affordable prices.

As time passed, I moved on from 300mm on crop to 200mm on FF, and even my crop has gone from 24-105mm to 24-70mm, so I do hope to see Canon releasing a small FF MILC with small lenses up to 100mm.


Edit: I started with 35-70mm on film & 18-55mm on crop, and was busy thinking the focal lengths were limiting. Now I regret I didn't just take more photos.
« Last Edit: March 17, 2018, 02:18:43 PM by Antono Refa »

canon rumors FORUM

Re: A Prototype Full Frame Mirrorless From Canon Exists [CR1]
« Reply #126 on: March 17, 2018, 02:13:35 PM »

Talys

  • EOS 5DS R
  • ******
  • Posts: 1452
  • Canon 6DII
Re: A Prototype Full Frame Mirrorless From Canon Exists [CR1]
« Reply #127 on: March 17, 2018, 04:11:21 PM »
AvTvM, you say that Sony chose poorly with E mount for FF,  because it was a crop mount repurposed for full frame. Now, I'm not disagreeing with you, but can you describe how a different mount would result in smaller, better, or cheaper lenses?  Only 2 things really matter in a mount design: throat diameter and flange size. Based on your knowledge, what would this superior mount change and why?

Here's the thing;

1. Sigma or Tamron have said that there are challenges that a small throat introduces. Also, I have read that this makes ulrawides much harder to build. But if you widen the throat, you make the camera body larger. You need more vertical and horizontal space, as the fingers are already cramped and the Songs essentially leave no space as it is.

2.  The focal flange distance can't be shorter for lenses that people would like shorter, without along other compromises, as I understand it. So 16-35, 24-80, 70-200 2.8 and popular lenses like 85 1.4 aren't getting any shorter from sensor to the furthest glass element, because this is determined by the optical formula, which requires so much space between the sensor and the first element.

3. Would not any change from E mount kill compatibility (even with adapter) with existing crop lenses? 

4. You refer to Leica lenses as a company that has 'solved ' the size issue. But these are manual focus, non image stabilized super expensive lenses. Can you cite any examples that are mainstream workhorse zooms, that have been shrunk, have great IQ, and are cheap?




BillB

  • EOS 7D Mark II
  • *****
  • Posts: 481
Re: A Prototype Full Frame Mirrorless From Canon Exists [CR1]
« Reply #128 on: March 17, 2018, 04:39:49 PM »
AvTvM, you say that Sony chose poorly with E mount for FF,  because it was a crop mount repurposed for full frame. Now, I'm not disagreeing with you, but can you describe how a different mount would result in smaller, better, or cheaper lenses?  Only 2 things really matter in a mount design: throat diameter and flange size. Based on your knowledge, what would this superior mount change and why?

Here's the thing;

1. Sigma or Tamron have said that there are challenges that a small throat introduces. Also, I have read that this makes ulrawides much harder to build. But if you widen the throat, you make the camera body larger. You need more vertical and horizontal space, as the fingers are already cramped and the Songs essentially leave no space as it is.

2.  The focal flange distance can't be shorter for lenses that people would like shorter, without along other compromises, as I understand it. So 16-35, 24-80, 70-200 2.8 and popular lenses like 85 1.4 aren't getting any shorter from sensor to the furthest glass element, because this is determined by the optical formula, which requires so much space between the sensor and the first element.

3. Would not any change from E mount kill compatibility (even with adapter) with existing crop lenses? 

4. You refer to Leica lenses as a company that has 'solved ' the size issue. But these are manual focus, non image stabilized super expensive lenses. Can you cite any examples that are mainstream workhorse zooms, that have been shrunk, have great IQ, and are cheap?

There seems to be a chicken or the egg issue here concerning price, assuming there is any technical solution to making small high quality full frame lenses, zoom or prime.  To get low prices you have to have high volume to spread fixed design and production costs over more units, but to get high volume you have to have low prices (assuming that price is a main factor in determining demand.).  Opinions expressed in this forum to the contrary, I don't think there is all that much demand for primes, pretty much regardless of price or size.  So that leaves zooms, and the question becomes the minimum size achievable for high quality, attractively priced, zooms.

AvTvM

  • Canon EF 400mm f/2.8L IS II
  • *********
  • Posts: 3165
Re: A Prototype Full Frame Mirrorless From Canon Exists [CR1]
« Reply #129 on: March 17, 2018, 05:02:38 PM »
@Talys - i dont have the slightest knowledge in optics design / engineering. I am just an observer.

Based on my observations of different lens mounts/lenses and related parameters / performance characteristics may guesstimate for an optimal FF mirrorless lens mount is around
* as big a hole as possible ... 50mm throat width and
* FFD not too short ... around 22-24mm - definitely NOT as short as 18 mm 
:-)

I "believe" (since i cannot do it myself i'm just a believer, so no hard facts .. )  this to be the best possible MILC lens system platform allowing for compact AND high-IQ AND affordable lenses in the most used focal range between 20mm and 100mm ... which is all i really care for. :-)
« Last Edit: March 17, 2018, 05:06:13 PM by AvTvM »

Rocky

  • EOS 5DS R
  • ******
  • Posts: 820
Re: A Prototype Full Frame Mirrorless From Canon Exists [CR1]
« Reply #130 on: March 18, 2018, 11:06:06 AM »
@Talys- Flange distance is the distance between the front of the len mount to the sensor (or film).  It will affect the optical design in BIG TIME for short focal length lenses, may be up to 75mm, depends on the actual flange distance. Think about a simple, symetrical single element convex lens. The center of the lens is the focal length of the lens. For Canon EF mount, the flange distance is 44 mm. How do we do the 28 mm wide angle len? The lens will be inside the camera body and will be hit by the mirror.The answer is a reverses telephone design. To put in in a simple way, a concave lens put in front of a convex lens. The conex lens is placed beyond the flange such that it will not be hit by the mirror and the convex lens is way in front of the convex lens.Then you end up with  large lens. in real live, it will end up with anywhere from 6 to 8 element for all kinds of optical corrections. Throw in the AF and IS, then you will have a HUGE lens.
Image sensor likes to have close to vertcal incident angle. That also put a big restriction on the system. Leica M choose 28.7 mm flange distance for the compatability of ALL Leica lenses since 1930's via a 1mm thick adapter. The original screw mount is 29.7 mm. This is choosen for mechanical reason. That is the thickness of the Leica body that will allow the usage of reusable film cartrige. It is doing "okay" withthe M8 (APSC-H) with software correction for dark corners. It is not good enough for the full frame sensor. They come up with off-set micro-lens to correct the problem caused by the large incident angle.
So the EF mount(44 mm flange) is a "Blessing In the sky" for Canon to minimize the effect of incident angle. EF-S lense take advantage of the smaller mirror of the APS-C camera. The EF-S lenses get deeper into the camera body to allow easier lens design
Fast forward to EF-M with 18 mm flange. I have no idea where this is from. May be that is the minimum body thicknesss to fit eveything in. May be that is the minimum distance to allow various adapter to be made for vintage lenses, including Leica. Anyway, we will think that making a 22 mm lens will be a piece of cake. NO. The incident angle come to mess up things again. If you look at the EF-M 22/2 closely, you notice that the front elenment is a lot more than 22 mm from the sensor and the back element is a lot bigger that the front element. Also the surface of the front element is convave. All of them is to minize the problem caused by incident angle. Even with all these, the EF-M 22/2 is still -1.5 to -2.0 stop dark corner. In fact, the 22/2 is one of the best "small" lens and still priced at $300.
Leica lenses are small and expensive due to the following reasons: exceptionally tight tolerence and quality control, low volume, no AF and no IS, Some lenses have the rear element go outside of the lens mount
 

Talys

  • EOS 5DS R
  • ******
  • Posts: 1452
  • Canon 6DII
Re: A Prototype Full Frame Mirrorless From Canon Exists [CR1]
« Reply #131 on: March 18, 2018, 01:55:06 PM »
@Rocky - I understand the issue of incident angle of light striking the sensor to be the reason why Sony has had so much trouble producing full frame ultrawide angle lenses. 

Also, as I understand it, for longer focal lengths, the flange focal distance simply can't be reduced.  The variations of the optical formulae being used all have a pretty significant distance between the sensor and the first glass element, so if they shorten the mount, they must lengthen the lens by the same amount to create the same space between glass and sensor.

The two factors I was driving at, was this:  For all the people who want Canon to abandon EF and go to a mirrorless mount, I ask, why?

1. The camera+lens package isn't going to get much shorter for large aperture bread-and-butter zooms.  It IS going to get shorter for low-end consumer lenses that have smaller variable apertures, but do we care, in the context of $2,000+ cameras?  It's also going to get smaller in terms of pancakes, but again, do we care?  These are small lenses anyways.

2. The lens diameter is not going to get smaller beyond the first glass elements if you reduce the throat diameter on most lenses that are produced that are large aperture.  A Sony E mount 100-400 or 70-200 may have a throat diameter of 46mm, which accommodates the nearest glass element just fine.  But the furthest glass element and pretty much all of them in between are exactly the same size as a Canon, ending with glass taking a 77mm filter.  So who do we benefit with a small throat?

Again, predominantly, small, variable aperture zooms, where the entire lens barrel can miniaturized because none of the glass elements need to be very large. 

For people who use any of the lenses where the diameter of the lens is larger, the smaller throat is a burden, because the camera/lens design becomes like the Sony, where the lens tapering to the mount constrains finger space (the alternative is to make a camera that and feels cavernous for the smaller lenses that don't have large barrels).

I think all of this makes sense in something like EF-M, where the market is for lower price points and portability.  But it makes a lot less sense in full frame, where most of the people who are buying it and using it in the way it was intended are going to also invest in expensive glass.

I think Canon has it right at the moment: make APSC small and light, and make full frame ergonomically correct.

AvTvM: I take your point that you only care about focal lengths up to 100mm and you want Canon to make a camera that caters to you.  But you already know that Canon is never going to do that, and nor is any other company that makes full frame cameras, because what you describe excludes a HUGE portion of their professional and enthusiast customer base.

So, the lens mount -- the throat diameter and FFD -- and ergonomic design of the camera will be designed for somebody in mind, with compromises so that it can be usable for everyone else.  Sony's E mount design is to make the camera body small at any cost, and to use that as a marketing push, and to a large extent, for the segment that really wants that, they have been successful.  To an extent, they have even been somewhat successful with people who want more ergonomically fit cameras, by selling them a battery grip that at highway robbery prices.

I do not want this for my camera.  I don't want a tapering lens design (certainly not to the extent of Sony), or a body that MUST have a grip to be usable, even with relatively small lenses like 24-70/2.8.  I do not want an interchangeable lens point and shoot; I want something that feels right whether it has an 8-15mm or a 100-400mm.

I can accept that EF may not be the optimal design, but the question is, how much smaller to make the throat diameter, and how much shorter to reduce the FFD?

AvTvM could well be right that 50mm would be an ideal throat diameter.  But I don't want to throw out all my lenses, or wait for new lens designs, just for 4mm.  I don't think that reducing the FFD is helpful for the same reason: I wouldn't support it unless they could reduce the total package size by enough to make a meaningful benefit for enough focal lengths of professional-grade options, and I don't think that is in the cards.

jolyonralph

  • EOS 7D Mark II
  • *****
  • Posts: 575
    • Every Other Shot
Re: A Prototype Full Frame Mirrorless From Canon Exists [CR1]
« Reply #132 on: March 18, 2018, 03:31:14 PM »
Sigh, so much misinformation again.


Shorter flange distance allows design of lenses suitable for short, medium or long flange distance (yes, you just put the rear element recessed inside the lens if you need to)

Longer flange distance only allows design of lenses with longer flange distance.


In what world is that more flexible than a shorter flange distance?

@Rocky - I understand the issue of incident angle of light striking the sensor to be the reason why Sony has had so much trouble producing full frame ultrawide angle lenses.

I'm not saying that Sony hasn't had problems but there is no lens design that works currently on the Canon EF mount that would not work just as well on the FE mount with the suitable adaptor and/or alteration to lens mount.  So, there's absolutely no sense in your statement linking the shorter flange distance to this.

And tell me, if the incident angle of light issue is so bad then how do we have the spectacular Sony Zeiss FE 55mm f/1.8 and 35mm f/2.8 lenses?  Before you say this is a problem that can't be solved go and actually try the lenses.

There are some people here with an agenda to poison the atmosphere about possible new mounts for a Canon FF Mirrorless because they don't want to see a fall in the usefulness/value of their EF lens collection.

I can understand why you would want to do this, but please, what you're saying is simply misrepresenting the facts.
Jolyon Ralph

Cameras: 5DSR, A7RII, 5D III, EOS M6/M5/M3, Mavic Pro, DXO One.  Oh, and more lenses than I care to count.

canon rumors FORUM

Re: A Prototype Full Frame Mirrorless From Canon Exists [CR1]
« Reply #132 on: March 18, 2018, 03:31:14 PM »

Talys

  • EOS 5DS R
  • ******
  • Posts: 1452
  • Canon 6DII
Re: A Prototype Full Frame Mirrorless From Canon Exists [CR1]
« Reply #133 on: March 18, 2018, 03:50:16 PM »
@jolyonralph - You should read my post a little more carefully. I did not say that a shorter FFD is a limiting factor.  Rather:

1. On the big lenses with wide apertures, you don't get to make them any smaller. To the contrary,  any space you saved o. The mount side needs to be added right back in on the lens side, making for 10 longer lenses and one smaller body.

2. The lenses that really benefit were not huge to start with, and are generally not pro lenses.

3. Small throat and small flange mean tapering lens and ergonomic grip issues.

4. I'm not sure which market wants $2000-$4000 camera bodies and small, cheap, crappy lenses with small and/or variable apertures.

As to the ulrawide issue, I have read in several places that the E mount makes ultrawides more difficult to build, and Ilit seemed believable to me.  It doesn't personally affect me, as I don't own a lens wider than FF 16mm anyhow, nor do I desire one, though It is notable that Sony's 16-35 was released only last year.

neuroanatomist

  • CR GEEK
  • ***************
  • Posts: 22546
Re: A Prototype Full Frame Mirrorless From Canon Exists [CR1]
« Reply #134 on: March 18, 2018, 04:16:03 PM »
And tell me, if the incident angle of light issue is so bad then how do we have the spectacular Sony Zeiss FE 55mm f/1.8 and 35mm f/2.8 lenses?  Before you say this is a problem that can't be solved go and actually try the lenses.

There are some people here with an agenda to poison the atmosphere about possible new mounts for a Canon FF Mirrorless because they don't want to see a fall in the usefulness/value of their EF lens collection.

I can understand why you would want to do this, but please, what you're saying is simply misrepresenting the facts.

Speaking of misrepresenting the facts, have a look through the mount of those two lenses you mention...what do you see?  A recessed rear element.  In other words, the problem of high incident light angles was 'solved' by avoiding them with an empty spacer tube at the back of the lens.
EOS 1D X, EOS M6, lots of lenses
______________________________
Flickr | TDP Profile/Gear List

canon rumors FORUM

Re: A Prototype Full Frame Mirrorless From Canon Exists [CR1]
« Reply #134 on: March 18, 2018, 04:16:03 PM »