[BigMackCam]

I never delved into it but I though Canons dual pixel af was about the sensor?

Yes, and so far as I'm aware, Canon's DPAF is only available in live view for DSLR models - which makes sense, since when focusing through the viewfinder, the mirror is in the way...

[Kobie]

I never delved into it but I though Canons dual pixel af was about the sensor? What you are describing above sounds very different and very Pentax.

I meant similar regarding what they're trying to accomplish. It's a different means to an end result but the ideas both work the same way.
I think Pentax's way of doing this will be better due to the separation of the 2 focus sensors that are matching an image at different planes, it either lines the two images up or doesn't, there is no in between or "almost nailed focus" which in theory, should be fast to focus, incredibly accurate and reduce hunting.

[StarTroop]

The second patent is a new AF system that uses 2 AF modules and 2 mirrors which are offset and facing different angles. The interesting things (to me) is that the new AF system is directly tied to the exposure system with linked AF points and dual Phase detect sensors that operate both while initially acquiring focus and (if I read this correctly) during the actual time the exposure is taken. IT's basically a full time dual linked Exposure/AF system. It seems that it's very similar to Canons dual pixel AF but possibly with a unique mirror and dual AF sensor placement. The benefit of this system is full time AF with real time tracking using Phase Difference AF (Strengths of Contrast Detect AF: Accuracy, and Strengths of Phase Detect: Tracking and Speed) by combining both using 2 separate modules in parallel that see the same image in 2 different ways regardless if the mirror is down or up by lining up the 2 images (the Difference in Phase) so they match. This allows full time Phase AF in live view, movie recording and using the viewfinder.

What exactly do you mean when you say dual linked Exposure/AF system? I understand that later you describe basically an alternate way of doing dual pixel AF, but when you say "exposure system" I'm imaging the light meter. Does what you're describing imply that the scene exposure could be measured during exposure itself, like with the old OTF metering?

[Larrymc]

What exactly do you mean when you say dual linked Exposure/AF system? I understand that later you describe basically an alternate way of doing dual pixel AF, but when you say "exposure system" I'm imaging the light meter. Does what you're describing imply that the scene exposure could be measured during exposure itself, like with the old OTF metering?

Wonder if we are getting a glimpse of the K-New A/F and metering system??

[mee]

I never delved into it but I though Canons dual pixel af was about the sensor? What you are describing above sounds very different and very Pentax.

Yes. Canon's method uses two rectangular photodiodes (or mini-pixels, my term) that are housed under a single microlens to make a single square 'pixel'. Theres a bit more detail on this thread: Re: How many cross-type AF points?: Nikon Z Mirrorless Talk Forum: Digital Photography Review

The Pentax method appears to skip doing this and instead use two different AF modules at different angles to calculate the phase. I'm not sure its truely DPAF-level but it seemingly has similar qualities.


Please don't ask me questions, as that is the extent of my knowledge on this topic.

[Kobie]

What exactly do you mean when you say dual linked Exposure/AF system? I understand that later you describe basically an alternate way of doing dual pixel AF, but when you say "exposure system" I'm imaging the light meter. Does what you're describing imply that the scene exposure could be measured during exposure itself, like with the old OTF metering?

In the patent it's mentioned that there's 2 different exposure levels that happen during camera operation. One during AF with the mirror down, then it changes when the mirror flips up due to the current properties of the mirror system and how much light comes after the mirror/AF sensor then the Imaging sensor. By using this new system, there's essentially 2 mirrors at differing angles. The AF sensor can read exposure and send that info to the imaging sensor so there's a consistency of the real exposure value regardless of the mirror position. It seems that they've come up with a way to use a sensor in multiple ways. The AF sensor needs light to work, it can send the exposure value that it sees to the imaging sensor which then checks it's own exposure value in relation to what both AF sensors are showing for the more direct light hitting them and averages the difference for a more accurate exposure.
Remember, this is no matter what position the mirror is in. And we've all read the threads before "Why is my exposure different in Live View than it is in the View Finder?". This will eliminate that problem since the light information will be the same no matter which mode you're using (no more components that inadvertently block some of the light like a translucent mirror strip so the AF module gets some light with the mirror down etc). It should also allow extremely low light AF to work.
At least this is my take on what the patent shows and says. They don't really explain it 100% in non engineer speak but I work with engineers every day so I can navigate pretty well through the wording.

[phoebus]

In the patent it's mentioned that there's 2 different exposure levels that happen during camera operation. One during AF with the mirror down, then it changes when the mirror flips up due to the current properties of the mirror system and how much light comes after the mirror/AF sensor then the Imaging sensor. By using this new system, there's essentially 2 mirrors at differing angles. The AF sensor can read exposure and send that info to the imaging sensor so there's a consistency of the real exposure value regardless of the mirror position. It seems that they've come up with a way to use a sensor in multiple ways. The AF sensor needs light to work, it can send the exposure value that it sees to the imaging sensor which then checks it's own exposure value in relation to what both AF sensors are showing for the more direct light hitting them and averages the difference for a more accurate exposure.
Remember, this is no matter what position the mirror is in. And we've all read the threads before "Why is my exposure different in Live View than it is in the View Finder?". This will eliminate that problem since the light information will be the same no matter which mode you're using (no more components that inadvertently block some of the light like a translucent mirror strip so the AF module gets some light with the mirror down etc). It should also allow extremely low light AF to work.
At least this is my take on what the patent shows and says. They don't really explain it 100% in non engineer speak but I work with engineers every day so I can navigate pretty well through the wording.

I read it slightly differently.. the two light paths are similar to the current system:

1. Focus through the semi-mirrored section of the viewing mirror, reflected via a folding mirror behind it down to the PDAF sensor
2. Exposure (and possibly image interpretation - face detect?) from the sensor on the prism

The patent makes a lot about the stepped positioning of the focus pixels and adjacent exposure pixels: talking about using the exposure pixels to determine when to read the focus pixels. My take on that is; the stepped focus pixels give effectively double the resolution, while keeping the sensing areas large as it would be for single resolution. The exposure pixels on the focus silicon allow the adjacent focus pixels to be read as soon as enough photons have been received. I interpret that as allowing focus updates at an optimal interval for the lighting conditions: bright conditions allow much quicker sampling. There is also some mention of using the exposure sensor as part of the process.

There again, I might have completely misread the rather woolly Google translation.

[house]

This sounds rather expensive.Or what do you think?

[Kobie]

This sounds rather expensive.Or what do you think?

Well, as of what's available right now...
A camera that has a built in intervalometer that allows over 400 interval shots, can track the stars, 9 multi exposures, allows you to pan and rotate the sensor to adjust composition, lets you convert a jpeg into a raw file, is freeze proof, dust proof/fully weather sealed etc sounds really expensive, but we have that as Pentaxians and it's not that expensive. So even if it sounds like it is, doesn't mean it will be.

[Kobie]

I read it slightly differently.. the two light paths are similar to the current system:

1. Focus through the semi-mirrored section of the viewing mirror, reflected via a folding mirror behind it down to the PDAF sensor
2. Exposure (and possibly image interpretation - face detect?) from the sensor on the prism

The patent makes a lot about the stepped positioning of the focus pixels and adjacent exposure pixels: talking about using the exposure pixels to determine when to read the focus pixels. My take on that is; the stepped focus pixels give effectively double the resolution, while keeping the sensing areas large as it would be for single resolution. The exposure pixels on the focus silicon allow the adjacent focus pixels to be read as soon as enough photons have been received. I interpret that as allowing focus updates at an optimal interval for the lighting conditions: bright conditions allow much quicker sampling. There is also some mention of using the exposure sensor as part of the process.

There again, I might have completely misread the rather woolly Google translation.

That does make sense. Maybe I interpreted that part incorrectly lol.

[jcleary47]

Are these patents new enough where this technology probably isn't in the K-New camera? Or is it not really fair to try to deduce that just from patents alone?

[Kobie]

Are these patents new enough where this technology probably isn't in the K-New camera? Or is it not really fair to try to deduce that just from patents alone?

It could be either or. When a patent gets filed, it could be for the idea with no functioning proof of concept/prototype. Or it could be that they've already created it, tested it, it works and now they want to protect the invention. There's a few things that lead me to believe it will be in the K-New.
1) They mention that it's a BSI sensor and word around the campfire (Internet) is that it's going to use a new gen BSI Sony sensor.
2) Images of the K-New show a larger AF module which makes sense for this patent.
3) The previous patent for the hybrid viewfinder system would need a different AF system to be a worthwhile implementation.
4) The camera has been delayed. Most likely due to implementing this new AF system and the new SR system and testing to make sure there's no I'll effects.
But, they could scrap some of it. They better not though, they NEED this! WE need this lol

[Mistral75]

Remember that the protection granted by a patent begins on the day of its filing. Inventors are therefore lead not to file their patent requests too early.

The first one (voice coil motors for sensor stabilisation) was filed on 16 January 2015, therefore it should have been put in use by now.

The second one (focus detection device) was filed on 20 June 2016. The same answer more or less applies.

[Kobie]

Remember that the protection granted by a patent begins on the day of its filing. Inventors are therefore lead not to file their patent requests too early.

The first one (voice coil motors for sensor stabilisation) was filed on 16 January 2015, therefore it should have been put in use by now.

The second one (focus detection device) was filed on 20 June 2016. The same answer more or less applies.

Not necessarily. It would take time for testing and all that fun stuff.
For the 1st one, perhaps as SRII.

For the AF, no. It would require a sensor that Pentax hasn't had until supposedly now for the K-New.

Here's a list of camera releases by year from 2015 onwards.
Pentax K-S2 (2015)
Pentax K-3 II (2015)
Pentax K-1 (2016)
Pentax K-70 (2016)
Pentax KP (2017)
Pentax K-1 II (2018)

None of those cameras have that AF system.

[bobell69]

With the new IBIS System and the redesigned focus system., I'm just wondering if we may see a Hybrid Camera coming out from Pentax. No one has come out with a DSLR that can also mimic a mirrorless camera. I can only imagine the benefits of having a camera that offers both of the advantages of a DSLR and a Mirrorless camera all in one body. It could very well be a game changer.