[mee]

With a translucent mirror that should be possible. If the mirror can be flipped up it wouldn't have the light loss as Sony SLT did.

That said, On sensor PDAF is not written in silicon. Its an add on feature like the color filter array or anti aliasing filter. Ricoh can easily negotiate it away if they dont want it in their order. The expensive part, the silicon, can still be the exact same.

Both options seem rather tasty!

[Simen1]

If the new FF body has access to the BSI 6th generation sensor - as has been rumored, and the new ASP-C also gets a BSI based sensor, then there will be an open question of how much of the FF advances over the K3/K3II is sensor size (FF) alone, as opposed to the IQ enhancements of BSI. With my recent lens purchases (ASP-C limited 18-35, 60-250), I have somewhat limited (by choice) to ASP-c. So, the new K3 replacement with a 6th gen BSI sensor is my obvious choice. In addition the FF body will act as a price ceiling on the new ASP-c body - making it more affordable. It has been said that the FF segment of the market is about 10%. So, most of us will stay with the ASP-C form factor.

I will use my Sigma 18-35 f/1,8 on Pentax FF and i don't feel bad about it. Why should lens image circle size always ble oversized compared to the sensor? It doesent make any sense. Its OK that the sensor is oversized in some cases too. Like in the wide end of 18-35. The same story goes for Pentax DA* 60-250mm f/4 that is even closer to full frame coverage.

[interested_observer]

I will use my Sigma 18-35 f/1,8 on Pentax FF and i don't feel bad about it. Why should lens image circle size always ble oversized compared to the sensor? It doesent make any sense. Its OK that the sensor is oversized in some cases too. Like in the wide end of 18-35. The same story goes for Pentax DA* 60-250mm f/4 that is even closer to full frame coverage.

I don't feel bad at all with my acquisitions. The point I was making is that some folks - and especially since the FF body was announced, have limited their lens purchases to only lenses that will support the FF sensor size with their image circle. I am happy taking pictures with my K5IIs. It is a good size for me, along with the lenses. Any larger in terms of body size and lens weight, it would start to become more of a nuisance. I was recently in Boston for 2 weeks on business, going out each evening for several hours to shoot. I was toting 3 lenses, the body and a tripod all in a backpack. The 18-35 is a heavy lens. Substantially heavier than the 8-16. The 60-250 is no lightweight either.

Comparing these three that I took along to the 2 recent lenses announced by Pentax - the 70-200/f2.8 and the 150 - 450/f2.8, each weighing in at around 4 pounds. They are reportedly wonderful lenses, but I am still happy with the ASP-C sensor size, so just going with what works for me currently - without worrying about what the future may bring.

[Mistral75]

(...)

That said, On sensor PDAF is not written in silicon. Its an add on feature like the color filter array or anti aliasing filter. Ricoh can easily negotiate it away if they dont want it in their order. The expensive part, the silicon, can still be the exact same.

Er.... no. On-sensor PDAF photodiods are neither some kind of filter nor an add-on in front of the sensor. They are part of the sensor, engraved in silicon.

[Zygonyx]

A quite useful feature for live view AF...

[RonHendriks1966]

I will use my Sigma 18-35 f/1,8 on Pentax FF and i don't feel bad about it. Why should lens image circle size always ble oversized compared to the sensor? It doesent make any sense. Its OK that the sensor is oversized in some cases too. Like in the wide end of 18-35. The same story goes for Pentax DA* 60-250mm f/4 that is even closer to full frame coverage.

You only will get the aps-c size cut from the sensor, so you might as well just stick with an aps-c sensor camera.

[jimr-pdx]

I don't feel bad at all with my acquisitions. The point I was making is that some folks - and especially since the FF body was announced, have limited their lens purchases to only lenses that will support the FF sensor size with their image circle. I am happy taking pictures with my K5IIs. It is a good size for me, along with the lenses. Any larger in terms of body size and lens weight, it would start to become more of a nuisance. I was recently in Boston for 2 weeks on business, going out each evening for several hours to shoot. I was toting 3 lenses, the body and a tripod all in a backpack. The 18-35 is a heavy lens. Substantially heavier than the 8-16. The 60-250 is no lightweight either.

Comparing these three that I took along to the 2 recent lenses announced by Pentax - the 70-200/f2.8 and the 150 - 450/f2.8, each weighing in at around 4 pounds. They are reportedly wonderful lenses, but I am still happy with the ASP-C sensor size, so just going with what works for me currently - without worrying about what the future may bring.

I agree - what works now beats what might work later. Besides all I do is swap gear anyway
My zooms are pretty much all modern APSc, my primes older film types (except the DA15). As to weight I'm using the K-r right now and I like that size a lot; it's a smaller K-5 and has SR for video unlike recent bodies (until the K-3 II?). It still hurts that some basic features (image rotate, interval stills/movies) are missing from the K-s1 as I love its size especially for primes. One thin membrane over the top would nearly make it WR thanks to having no mode dial up there. A sandwich bag probably fits over the whole cam!

[Parallax]

This thread is neither Pentax news nor a rumor. It's pure speculation based off of a rumor

Right. Moved to proper forum.

[fredralphfred]

. . . some basic features (image rotate, interval stills/movies) are missing from the K-s1 . . .

The k-s1 doesn't support rotating the image? My K200D can rotate the image! That screen is small enough as it is. Wow. I'm glad I got the K-50 instead.

[Doundounba]

The k-s1 doesn't support rotating the image?

Jim is probably referring to "composition adjustment", which, incidentaly, is also apparently missing on the K-S2. Personaly, I think it's pretty disappointing that Ricoh is disabling in software features that the hardware is perfectly capable of supporting, and which were included in previous incarnations of the lower and middle tier models.

[kadajawi]

OK, I'm being a bit facetious in the last bit, but really -- video cameras and still cameras are two completely different beasts and always have been. They may converge a little more now that we've had a digital revolution, but getting a still camera to be a good video camera (and no, the Canon 5d mk. ii NEVER qualified) is really about as easy as modifying an old Pentax film camera to take motion pictures. People are complaining about the battery life on the K-3ii. What will they say when it gets a 4K capable processor? HD video takes plenty enough juice, and that's probably the reason Pentax isn't supporting it better right now -- if the processor can handle still pictures with judicious speed, upgrading it to handle video just eats battery life. And it's all software (firmware) nowadays -- there are no more dedicated ASIC processors, especially low power ones, for encoding. It's just too complicated, and the standards change too quickly. Besides, about the only camera company that makes lenses good enough for video (they have different requirements) is Canon. (Although apparently you can get an adapter to fit old TV studio camera lenses on the Pentax Q [and I have to specify now because there are about 3 or 4 different "q" cameras on the market at the moment]).

The built-in video mode is great for if you're somewhere with your still camera and want a casual video of something that's happening. It would be irritating to carry a camcorder for a "what if" like that, and yes, the gear you have is probably better than something small and dedicated (not to mention expensive). But everyone seems to want production video features on their still camera. Considering what they won't be able to do, regardless of Pentax's wishes (or more probably its users'), I have to wonder why.

TV shows are produced on the NX1. Philip Bloom's gorgeous documentary series for CNN was shot in parts on an A7S.

Ergonomics are a bit of an issue, I agree, but dedicated cameras are quite expensive, and there are enough people who want to/have to shoot stills AND video... and who may be happy to give up a bit of ergonomics for the convenience of only having to own and carry one camera.

The K-5 was already a good video camera, all it really lacked was manual controls. The quality was good IMHO, and I'll take it over many other cameras for video. It's not too difficult.

As for 4K capable processor... such a processor should generally be faster, meaning that even if it uses more power, it will use it for a shorter time. The power consumption per task (i.e. saving a raw photo) may not be so bad. And if you don't record 4K video, the power consumption for doing so shouldn't matter I believe.

The video encoder has to be in hardware. Encoding 1080p video in h264, even in baseline settings, in real time, requires some serious processing power. Think i5, i7. There's no way they can be doing that part in software, there has to be some hardware support for certain tasks at least, though from what I heard the K-3 (?) did fix some encoding errors through a firmware update. The rest I do think is software, and Pentax could improve it for video.

There are video users shooting their Canon etc. with old Pentax glass. It doesn't have to be high end, IMHO. There's an area between Hollywood and home videos that may not be able to afford all the expensive Hollywood gear. Working on a small budget. Or even ambitious home video people (probably more where I am).

Back to BSI sensor... they mostly help with high pixel densities, so for example, 28 MP sensors that otherwise are more like 20-24 MP sensors. If you do it with a 20 MP sensor, the benefits will be smaller.

Separate PDAF sensors (DSLRs) should be superior to on sensor PDAF sensors in terms of low light performance, because they are bigger, have optics to bundle more light, ... the on sensor sensors have to fight with all the pixels for space. On the other hand you can spread the on site sensors much more, have more in general, and they are on the same plane as the image sensor. No need to adjust...

[Na Horuk]

I have yet to see anything that says BSI improves on sensor AF or 4K video. Do you have link to this?

No, the BSI itself might not improve AF, but the AF technology itself has been improved. Now the sensor features special sites for AF, with better algorithms, than what was available for the K-01, which is basically using the same sensor as K-5 or K-50. A sensor that is more specialized for mirrorless would.. by definition be better suited for a mirrorless camera. Will it be significantly better than the K-3II (current flagship)? Maybe not by a big leap, but it would be a good next step for the K-01. Even better noise performance, even sharper, and snappy AF (AF being the main weakness of the original K-01).

And LCD screen technology improved since the K-01 as well. Air gapless thingy, as well as OLED and other technologies. Make it tiltable, improve focus peaking (And allow user chosen colours), add a removable EVF (which would also work with the Q).. and use faux leather instead of rubber this time, so the camera has more mainstream appeal. Would be amazing.

[reduno]

One possible issue that I have not seen addressed is the Pentax in body SR. This effectively means the sensor has to be mounted separately from components on the main board, with the connection made by multiple flex circuits (in the models I have seen inside anyway). This can provide a bottle-neck for reading large amounts of data rapidly out of the chip (as needed for 4K video). Even with the same number of connections, it is much harder to achieve the bandwidth with a longer flex trace, compared to having the chips sitting right next to each other on the same board. The optically stabilized lens systems may have a subtle benefit here. Just a thought?

[kadajawi]

One possible issue that I have not seen addressed is the Pentax in body SR. This effectively means the sensor has to be mounted separately from components on the main board, with the connection made by multiple flex circuits (in the models I have seen inside anyway). This can provide a bottle-neck for reading large amounts of data rapidly out of the chip (as needed for 4K video). Even with the same number of connections, it is much harder to achieve the bandwidth with a longer flex trace, compared to having the chips sitting right next to each other on the same board. The optically stabilized lens systems may have a subtle benefit here. Just a thought?

Doesn't Sony now have FF 4K cameras with SR? That, btw., works during video.

[reduno]

Doesn't Sony now have FF 4K cameras with SR? That, btw., works during video.

Yes, you are correct. Sony does have the capability but the camera is not cheap. It is certainly doable but not as simple as placing two chips next to each other. The killer to high bandwidth is usually the capacitance of the connecting traces, which is why modern logic chips are running with lower and lower supply voltages (so they don't need to switch as much charge).

To achieve a longer connection requires the use of a controlled impedance line, with matched driver and receiver terminations (similar to the distribution of cable TV signals on a smaller scale). These could be built into the chips and the flex circuit designed to match the impedance. Every connection (solder joint) then becomes a potential miss-match in the transmission line that can degrade the signal.

It is certainly not insurmountable but requires careful design and potentially special chip designs.