[Kunzite]

This is already being done with on the sensor PDAF. Refining that technology is IMO more logical, than going back to a separate AF system.
Doing it the other way around - making a SLR which would use the main sensor for AF (and other things) will reduce cost and calibration issues. Perhaps still not good enough in the immediate future, but a viable idea.
Unless there's something fundamentally wrong with the on-sensor AF which I'm not aware of.

[JimmyDranox]

Interesting. Very interesting. I've tried to read the automated translation of this article, and I have some conclusions.

This way, the K mount will be preserved in the mirrorless era.

More than this, some downside of the mirrorless cameras will not be present using this mirror in a DSLR. That's about image degradation in the corners, and lack of PDAF in mirrorless (as I understand from that translation).

One thing is unclear. If the maximum transmission can be 100%, what about maximum reflection? If the mirror can change the characteristics very quickly, as they say, and the reflection can be raised to more than 80%, an OVF could also be used.

Anyway, I think that's a good news. And I hope that this innovation will be developed to production very quickly.

[zosxavius]

This is one thing that is often overlooked with on sensor PDAF: FF/BF issues due to poor calibration are completely nonexistent. In light of this I have to agree that aggressive pursuit of this technology is the way to go.

[Zygonyx]

...
More than this, some downside of the mirrorless cameras will not be present using this mirror in a DSLR. That's about image degradation in the corners, and lack of PDAF in mirrorless (as I understand from that translation).
...

downsides of mirrorless / EVF :
- increased power consumption (most of the time), hence reduced batterylifecycle
- still too low frame-refreshment rate ; and consequent disturbing imprecision and oscilations mainly under artificial light
- no way to look inside your finder, once your battery is exhausted... too bad for nature shooter/watchers

---

Last edited by Zygonyx; 11-29-2014 at 01:29 PM.

[derekkite]

Isn't this about hybrid ovf? Allow enough light to get to the sensor for focus peaking, but keeping the benefits of a mirror?

[beachgardener]

probably for the KS-2.

[Mistral75]

This is already being done with on the sensor PDAF. Refining that technology is IMO more logical, than going back to a separate AF system.
Doing it the other way around - making a SLR which would use the main sensor for AF (and other things) will reduce cost and calibration issues. Perhaps still not good enough in the immediate future, but a viable idea.
Unless there's something fundamentally wrong with the on-sensor AF which I'm not aware of.

Nothing really wrong but you have to completely redesign your lenses.

Read for example what Mr Kazuto Yamaki, CEO of Sigma (KY here below), told to Dave Etchells (DE) and William Brawley(WB) from Imaging Resource in an interview made during Photokina when asked whether the 18-300mm f/3.5-6.3 DC Macro OS HSM might come to Sony E-mount or µ4/3 mount:

Sigma Q&A with Kazuto Yamaki: Why no Micro Four Thirds lenses, and is a full-frame Foveon feasible?

KY: No, because it's made for the conventional DSLR. If we made it for the mirrorless or compact system camera, we'd have to change its optical design for focus tracking.

WB: Like you said [note: earlier in the interview], those would be live view platforms, so would need a different focus system.

KY: Yeah, it would have to be a stepping motor or voice coil motor.

DE: Oh, good point.

KY: In the case of the mirrorless camera, the focus has to always track the subject, because of the "wobbling" [in the US, we refer to this as "hunting", moving back and forth to determine the point of focus], always moving a little bit. In this case, the focus glass must be very lightweight.

WB: The elements that are moving have to be lightweight enough to move fast.

KY: Yeah, and also there are not many motors which have both strong power and compact size, so the lens element has to be very light.

DE: Very, very light. And so the design of that lens, I mean, that would be a completely different optical design.

KY: Completely different.

DE: Yeah, different optimizations could be more expensive. Or might sacrifice quality.

KY: It cannot be expensive, but it [would have to be] totally changed, a different design and construction.

And I definitely think Ricoh are not prepared to completely redesign the whole line of Pentax K-mount lenses.

[dakight]

I don't know much about it but I'm skeptical. Even if the transmission can be 100%, which I seriously doubt, it is still another optical element in the path of the light and that inevitably changes it and usually not for the better. Improvements in electronic viewfinder imaging are more likely to be the solution to the shortcomings of the mirror over the long term.

[Mistral75]

(...)

One thing is unclear. If the maximum transmission can be 100%, what about maximum reflection? If the mirror can change the characteristics very quickly, as they say, and the reflection can be raised to more than 80%, an OVF could also be used.

(...)

Like I wrote in the original post, if the transmission rate can be adjusted between 50 and 100%, and considering for the sake of simplification the light loss through absorption as negligible, the reflection rate will vary from 50 to 0% and its maximum is therefore 50%, i.e. minus 1 EV.

[LensBeginner]

Replacing OVF by EVF would actually sign the end of PENTAX as an historical and ethymological full sensed acronym...(*)
This would be actual nonsense marketingwise, also because it would probably mean K-mount's death => in japanese : a dobble HARA-KIRI !


(*) : The Pentax (derived from PENTAprism refleX) camera was the first camera to incorporate a pentaprism viewfinder and a reflex mirror system in 1952...

My only cameras are a K100Ds (pentaMIRROR) and K-01 (live view only), how's that for a seppuku? ;-)

[JPT]

This is already being done with on the sensor PDAF. Refining that technology is IMO more logical, than going back to a separate AF system.
Doing it the other way around - making a SLR which would use the main sensor for AF (and other things) will reduce cost and calibration issues. Perhaps still not good enough in the immediate future, but a viable idea.
Unless there's something fundamentally wrong with the on-sensor AF which I'm not aware of.

I don't think there's anything wrong with PDAF on the sensor, but even though it shares a name with the PDAF systems on DSLRs, there may be some significant differences in the way they work. Take the Sony A99, for example. It has three ways of focusing: CDAF, on-sensor PDAF (102-point) and the dedicated PDAF sensor using reflected light (19-point). The A99 can only acquire focus using the 19 points on the dedicated PDAF sensor. It uses the 102 on-sensor PDAF to track objects that move outside or between the 19 points once focus has been acquired. The A99 was released in 2012, so at least at that time, there was a difference in how the systems worked.

Metering on SLTs is done using the imaging sensor and there is no need for a dedicated metering sensor. In my experience, the K-01 metering is very consistent compared to the K20D I had before. I've read something similar in other reviews. So it seems that this make another component redundant without compromising capability.

A did a quick search to find out about front and back focus issues on SLTs. They seem to have just as many problems as DSLRs. Since all DSLRs can focus in live view now, I don't know why no manufacturer has been able to make the camera calibrate itself with a lens. You would just put it on a tripod and point it at a target. The camera would go through a routine of comparing focus at all PDAF points and figure out the adjustment itself. I'm sure it would fail sometimes, but then it would just inform you to choose a better target and try again. Am I being unrealistic here?

[Mikesul]

I don't think there's anything wrong with PDAF on the sensor, but even though it shares a name with the PDAF systems on DSLRs, there may be some significant differences in the way they work. Take the Sony A99, for example. It has three ways of focusing: CDAF, on-sensor PDAF (102-point) and the dedicated PDAF sensor using reflected light (19-point). The A99 can only acquire focus using the 19 points on the dedicated PDAF sensor. It uses the 102 on-sensor PDAF to track objects that move outside or between the 19 points once focus has been acquired. The A99 was released in 2012, so at least at that time, there was a difference in how the systems worked.

Metering on SLTs is done using the imaging sensor and there is no need for a dedicated metering sensor. In my experience, the K-01 metering is very consistent compared to the K20D I had before. I've read something similar in other reviews. So it seems that this make another component redundant without compromising capability.

A did a quick search to find out about front and back focus issues on SLTs. They seem to have just as many problems as DSLRs. Since all DSLRs can focus in live view now, I don't know why no manufacturer has been able to make the camera calibrate itself with a lens. You would just put it on a tripod and point it at a target. The camera would go through a routine of comparing focus at all PDAF points and figure out the adjustment itself. I'm sure it would fail sometimes, but then it would just inform you to choose a better target and try again. Am I being unrealistic here?

Your calibration idea is intriguing. I wonder if would add too much to the cost?

[JPT]

Your calibration idea is intriguing. I wonder if would add too much to the cost?

I think all the necessary hardware is already there in currently DSLRs. I guess the cost would be in developing the firmware, which may have a lot of hidden complications that I'm not seeing.

I think that focus alignment is something that beginners are rarely aware of when they step up from compacts. By the time they figure it out, they may have taken years worth of pictures which are not as sharp as they could have been. I found this with my K20D and even when I realised, the process of calibrating each lens was tricky. I'm still not sure I ever got it completely right with some lenses. When I got the K-01, the problem simply disappeared.

Anyway, it seems that Sony SLT has the same potential issue as DSLRs, so this patent isn't a solution in itself.

[JimmyDranox]

I don't think there's anything wrong with PDAF on the sensor, but even though it shares a name with the PDAF systems on DSLRs, there may be some significant differences in the way they work. Take the Sony A99, for example. It has three ways of focusing: CDAF, on-sensor PDAF (102-point) and the dedicated PDAF sensor using reflected light (19-point). The A99 can only acquire focus using the 19 points on the dedicated PDAF sensor. It uses the 102 on-sensor PDAF to track objects that move outside or between the 19 points once focus has been acquired. The A99 was released in 2012, so at least at that time, there was a difference in how the systems worked.

Metering on SLTs is done using the imaging sensor and there is no need for a dedicated metering sensor. In my experience, the K-01 metering is very consistent compared to the K20D I had before. I've read something similar in other reviews. So it seems that this make another component redundant without compromising capability.

A did a quick search to find out about front and back focus issues on SLTs. They seem to have just as many problems as DSLRs. Since all DSLRs can focus in live view now, I don't know why no manufacturer has been able to make the camera calibrate itself with a lens. You would just put it on a tripod and point it at a target. The camera would go through a routine of comparing focus at all PDAF points and figure out the adjustment itself. I'm sure it would fail sometimes, but then it would just inform you to choose a better target and try again. Am I being unrealistic here?

I think that the main problem is not technical or financial, but psychological. How can someone sell a camera as an autofocus one, if every user must do calibration tests with every lens, before use it. Some lenses have small difference in focus for different distances, especially cheap zooms. A calibration test for such a lens will require time, patience and skills, and most people would not want to do this.

I'm sure that every manufacturer makes this kind of tests, to make lenses as parfocal as possible, and camera as reliable as can be in their focusing system, but still, their is a small margin of error.

[Mistral75]

(...)

(*) : The Pentax (derived from PENTAprism refleX) camera was the first camera to incorporate a pentaprism viewfinder and a reflex mirror system in 1952...

Huh...

What about the Zeiss Ikon prototypes from 1936-37? What about the Contax S and Rectaflex Standard (Telemaco Corsi, Italy), both from 1947?



The Asahiflex I (prototype: 1951; mass production: 1952) was 'only' the first Japanese 35mm SLR. And there was no pentaprism in the Asahiflex; the first Asahi Optical SLR with a pentaprism was the Asahi Pentax from 1957.

By the way, the name 'Pentax' was originally a registered trademark of Zeiss Ikon. It was taken by Asahi Optical in 1957

If you want a mirror-linked world's first from Asahi Optical, just pick the successful instant-return mirror of Asahiflex IIB in 1954 -this world's first was the first of many from Asahi Optical in the field of SLR.