[rawr]

Too different is too risky.

A conventional looking camera with great under-the-hood sensor and AF performance, great low-light specs like the K-3, miscellaneous cool enhancements maybe to the viewfinder, modern stuff like WiFi, and a flippy screen is all they need to do to wow people. Think of recent high performing cameras like the 7DII, D750, Sony A77II, A7S , even the recent Samsung NX1. The templates are there. No Frankenstein designs or features.

[Poit]

Too different is too risky.

A conventional looking camera with great under-the-hood sensor and AF performance, great low-light specs like the K-3, miscellaneous cool enhancements maybe to the viewfinder, modern stuff like WiFi, and a flippy screen is all they need to do to wow people. Think of recent high performing cameras like the 7DII, D750, Sony A77II, A7S , even the recent Samsung NX1. The templates are there.

I agree, which is why I'm beginning to think it won't be that 'different' at all, at least for the first release.

As far as future releases go, I can't stop thinking about the design of the K-S1. Particularly the promo video showing the two components coming together, and whether that isn't a teaser for a new mirrorless with modular k-mount adapter...part of their 'innovation' line...

That would certainly be the riskier release, which is why I don't think it'll be first.

[filorp]

Someone can call me crazy but i think Pentax will have hybrid shake reduction - in body stabilization will give about 1.5-2 ev advantage and in lens will give another 2-3 ev: it will be according to pentax policy "where no one was before"

[easyreeder]

What I am saying is that Pentax should start to rethink their position... it doesn't have to be the main priority, though it could be quite profitable to do so, but at least doing something would be nice, and I'd argue that if they don't they'll lose sales.

I've been thinking about what Norm said, that FF is a priority, and there is a great deal of truth to that. On the other hand, if the FF has to have a major mp count to compete with the K3, am I wrong to think it might not need a new processor anyway? Also, even if the FF isn't 4k, won't any future APS-C need to be 4k? That's where that sensor size is going, no? The Samsung NX-1? Sony? A little easier to handle the 4k in APSC (the A7 4k is at a crop). Uh, also, with these new 4k codecs, aren't we looking at a more efficient use of the processor?

Anyhoo, I could be totally wrong, but if Pentax is going to stay in FF and APSC, it would seem to make sense that the next big update in APSC (after the FF) is video, which is maybe only a slightly faster time table than our projections here of two-four years behind everyone else, haha.

[kadajawi]

I've been thinking about what Norm said, that FF is a priority, and there is a great deal of truth to that. On the other hand, if the FF has to have a major mp count to compete with the K3, am I wrong to think it might not need a new processor anyway? Also, even if the FF isn't 4k, won't any future APS-C need to be 4k? That's where that sensor size is going, no? The Samsung NX-1? Sony? A little easier to handle the 4k in APSC (the A7 4k is at a crop). Uh, also, with these new 4k codecs, aren't we looking at a more efficient use of the processor?

Anyhoo, I could be totally wrong, but if Pentax is going to stay in FF and APSC, it would seem to make sense that the next big update in APSC (after the FF) is video, which is maybe only a slightly faster time table than our projections here of two-four years behind everyone else, haha.

4K is coming whenever Fujitsu supports it. Now, maybe Pentax can push them towards 4K, or Nikon does it, but ultimately it's up to Fujitsu, unless they change their image processor supplier. And to be honest I'm not too worried about 4K. Also, image quality per se isn't that big of an issue for Pentax. They aren't awful, just not great either. Nikon has shown that the latest Milbeaut should be capable of good video encoding as far as h264 goes, and that it can downscale the full 24ish MP down to 1080p (2 MP). The next _new_ sensor Pentax will use surely should support it, no matter if it is APS-C or FF.


My main worry comes from a lack of understanding of/experience with video at the side of Pentax. They have good enough hardware, but their software is a let down. I just have to bring up SR... the hardware clearly is capable of it, as live view shows even the software seems able to do it, and without cropping the frame. But it was once, in a review, criticized that one can sometimes hear the SR mechanism, so they turned it off and used what Fujitsu offered them. But had Pentax actually tested the mechanical SR system in real life, they would have noticed that often times you don't hear it. Had they tested Fujitsu's electronic SR system, they should have seen how ugly the results are. Would you rather have (very) occasionally not perfect sound (not that the built in microphone is good anyway, and you can circumvent it by using an external mic, for which there is a port, or not using the audio at all), or a video that is constantly zoomed in and just looks awful almost always? And Pentax has mentioned that the sound is the reason why there is no mechanical SR. People, though not reviews I guess, have asked Pentax again and again, for the past 2 years or so, to at least give the option of mechanical SR, perhaps besides electronic SR. But Pentax isn't listening.


A 1080p FF or APS-C DSLR with pixel binning from a full sensor readout, lots of good MF lenses and a built in SR system that works well with all lenses, a decent video encoder (Nikon has it, so it should be available to Pentax too), plus proper video controls and a flat picture profile, that could very well sell nicely with the video crowds. The SR is the differentiator, and would move cameras, as long as the rest of it is good enough. The GH4 and A7S are smaller and lighter, yes, but the advantage of that would be with handheld shooting... which those cameras won't do well because of the lack of SR. You'll need to use stabilizing gear, which adds a lot of weight and bulk. The Pentax doesn't need it. And Canon and Nikon might offer similar cameras, but not with the SR everywhere. And Olympus has SR, but lacks everywhere else, for the moment (plus the small sensor). Besides, it could be popular with people doing both. Pentax has great ergonomics for stills, always had, and they take some really good photos.



Haven't they stopped stressing "it must be different" in interviews recently? Maybe they have given up and are now content with a me-too product...

[easyreeder]

I just have to bring up SR.

This is a critical problem that I think people are aware of. There are only two lenses that have OS for the k mount in a focal range one would apply in non-tripod video, and these are discontinued. (The Sigma 17-50 and 17-70, or something like that.) Also, the bitrate is a huge problem. But those are last year's disappointments. If we're talking about a 2015 or 2016 camera that wants to have adequate video specs, 4k is there. It is there in Samsung, sort of there in Sony, there in Olympus, there in Panasonic, promised in Fuji (sort of, no?), and not yet in Canon/Nikon but maybe. Two years from now, it's just another deal-breaker.

[noser]

And Pentax has mentioned that the sound is the reason why there is no mechanical SR. People, though not reviews I guess, have asked Pentax again and again, for the past 2 years or so, to at least give the option of mechanical SR, perhaps besides electronic SR. But Pentax isn't listening.

Or it was a contractual thing by Fujitsu... 'use this trashy video SR, or else'. Or it was baked in, and they can't shut it off...

Pentax makes such generally good decisions about key parts of the image making process that I can't help but wonder who was asleep for that one.

[kadajawi]

Or it was a contractual thing by Fujitsu... 'use this trashy video SR, or else'. Or it was baked in, and they can't shut it off...

Pentax makes such generally good decisions about key parts of the image making process that I can't help but wonder who was asleep for that one.

It is a common theme as far as video functionality goes.

Why would Fujitsu force them to do it? And probably the chips for Nikon have it too, and Nikon quickly saw how useless it is, or they even asked Fujitsu to remove it.

Electronic SR can be turned off, only the crop remains. And in live view you get an uncropped video feed that is nicely stabilised. All that is needed exists. They just would have to offer it. There is simply no rational reason behind this, except for Pentax employees being clueless or Sony forcing them (as in without that function no one will want the Pentax anyway...).

[noser]

All that is needed exists.

That's what's so weird about it... the LV feed is ok, but if it's buried in the codec output... then yuck.

There is simply no rational reason behind this, except for Pentax employees being clueless or Sony forcing them (as in without that function no one will want the Pentax anyway...).

I doubt it was the sensor, that's why I'm thinking Fujitsu in the encoder. Maybe Nikon tested it, said no, and could afford to have it shut off or order a batch without it... or it's just a firmware thing and Pentax refuses to put the option to use real SR instead of electronic SR.

I'm leaning towards 'clueless' now; it's true either way. The video team in Pentax seems so out of touch that this is the result. This is one of those cases where if video is pointless or drains resources from a better stills camera they should just forget it.

On the flipside I saw one video review of 645Z output that said due to the lack of believable preview they had to take a few seconds of test footage to check exposure, which made the whole workflow gross. Sounds like they aren't actually draining many resources if they aren't spending any time improving things at all.

[kadajawi]

That's what's so weird about it... the LV feed is ok, but if it's buried in the codec output... then yuck.



I doubt it was the sensor, that's why I'm thinking Fujitsu in the encoder. Maybe Nikon tested it, said no, and could afford to have it shut off or order a batch without it... or it's just a firmware thing and Pentax refuses to put the option to use real SR instead of electronic SR.

I'm leaning towards 'clueless' now; it's true either way. The video team in Pentax seems so out of touch that this is the result. This is one of those cases where if video is pointless or drains resources from a better stills camera they should just forget it.

On the flipside I saw one video review of 645Z output that said due to the lack of believable preview they had to take a few seconds of test footage to check exposure, which made the whole workflow gross. Sounds like they aren't actually draining many resources if they aren't spending any time improving things at all.

Not sure what you are talking about. When you are in still mode, and have LiveView activated, you'll get a wonderful, uncropped image. But when you are in video mode it's different. On the screen you'll get a cropped but unstabilized picture, which is why it doesn't look too bad. The algorithm in the processor will analyse the next few frames and then determine how to stabilize, which is why the screen won't even show the framing you're going to get.


What I'm saying is Sony might have said if you have proper SR, you won't get this sensor, because we don't want you to compete with us in terms of video. That's crazy speculation, but at this point... Again, as SR can be deactivated, it's not something that is forced upon Pentax against their will, it's not hard coded into the processor.


I don't think Pentax has a video team. And I really, really don't understand the existence of the headphone out. There is only ONE reason why the camera has it, and that is better audio. It takes some effort to program the stuff, and it costs money and space to have the plug. For a feature that is only useful for video. Most other brands don't go through this amount of effort for video. But then they manage to f*** up everything else?! That's like fitting a car with 10 airbags, stability control etc. and then making them out of cardboard and with brakes that make a ships captain laugh about their uselessness. The only reason I can see is that reviews criticized bad audio and the automatic gain (and I second that... it's pretty bad in the K-5), so they wanted to fix that, just like they "fixed" the noise the SR mechanism makes. They only listen to reviews like DPReview... (maybe EOSHD should write a Pentax review for DPReview...).

[Fogel70]

One possible reason for SR not used in video is because video put a lot more strain on the SR mechanics.
FI a 1h video is using SR as much as 150 000+ images. So they would need to design SR to withstand up to 1000x more use for video than if it is only designed for images.

They might have had quality issues on SR mechanics on earlier cameras if they where used a lot for video, and solving this might require a re-design on SR that might make the camera more expensive, heavier, larger, or might make it less effective for images. So I believe there might be some logical reasons for not using mechanical SR for video anymore.

[kadajawi]

One possible reason for SR not used in video is because video put a lot more strain on the SR mechanics.
FI a 1h video is using SR as much as 150 000+ images. So they would need to design SR to withstand up to 1000x more use for video than if it is only designed for images.

They might have had quality issues on SR mechanics on earlier cameras if they where used a lot for video, and solving this might require a re-design on SR that might make the camera more expensive, heavier, larger, or might make it less effective for images. So I believe there might be some logical reasons for not using mechanical SR for video anymore.

That sounds sensible, though I haven't heard about SR failure... not that I've been looking for it either. And I have recorded many hours of video... more than 82 hours of it, and always with SR on, including lots of walking etc (i.e. extreme movements with the sensor hitting the boundaries of the system). That would mean 12.3 million photos...


Btw., I previously suggested that Pentax could team up with Magic Lantern. Hire some of them, or in other ways work with them. Well... they are working with AXIOM, which is a 6000 Euro APS-C camera, and much of it's value comes from Magic Lantern.

[easyreeder]

Btw., I previously suggested that Pentax could team up with Magic Lantern. Hire some of them, or in other ways work with them. Well... they are working with AXIOM, which is a 6000 Euro APS-C camera, and much of it's value comes from Magic Lantern.

Ha, well, that would be amazing.

One possible reason for SR not used in video is because video put a lot more strain on the SR mechanics.

People do use the mechanical SR in video on the K5 and other cameras previous to the K-01 and K3. I've never heard of an SR failure either.

Hmm, something else entirely occurred to me, which is wildly optimistic perhaps. The in-lens stabilization for video is choppy (from what I've seen in the Sigmas). Samsung maybe has in-lens video stabilization that looks good, but there aren't many tests to look at. If it is the codec, for example, maybe Pentax is wise to pursue a software angle. I mean, a lot of ifs, but let's say in five years, every camera has a video codec three generations more advanced than what we have, and let's say all video is global shutter, and let's say pentax develops a way to turn off the face tracking if you want to turn it off, isn't that a much, much stronger direction than in-lens?

[kadajawi]

Ha, well, that would be amazing.

People do use the mechanical SR in video on the K5 and other cameras previous to the K-01 and K3. I've never heard of an SR failure either.

Hmm, something else entirely occurred to me, which is wildly optimistic perhaps. The in-lens stabilization for video is choppy (from what I've seen in the Sigmas). Samsung maybe has in-lens video stabilization that looks good, but there aren't many tests to look at. If it is the codec, for example, maybe Pentax is wise to pursue a software angle. I mean, a lot of ifs, but let's say in five years, every camera has a video codec three generations more advanced than what we have, and let's say all video is global shutter, and let's say pentax develops a way to turn off the face tracking if you want to turn it off, isn't that a much, much stronger direction than in-lens?

Maybe there are failures, but they are a bit rare? Or Pentax used a better, more robust SR system in the K-5 etc.?


I've seen great in lens stabilization from Panasonic camcorders, so I think it's not necessarily choppy.


There are sensors with global shutters, and AFAIK they are significantly less sensitive to light. The next generation video codec is h265 aka HEVC, and it won't be superceeded for many years. But yes, I think it is sensible to expect a Pentax in 5 years to have it.


The "face tracking" (it's not really that) is a result of Pentax not using any of the sensors that track the movement of the camera, thus the camera has to guess if the camera has moved. If too much of the motive is moving, voila, "face tracking". Fujitsu would have to program a way to use the data for the electronic stabilization.


However even when you have eliminated that, and you have a fast enough sensor readout, a global shutter or software that corrects the wobble (Sony seems to have that, my Qualcomm Snapdragon 801 based smart phone too) so that the wobbling isn't an issue anymore, you'll still end up with inferior video quality. First of all it is zoomed in, you can't use the whole sensor. Secondly say you are shooting 1/50th at 25 fps. Any shaking that is faster than 1/50th will be recorded (and imagine shooting at 1/25th), so what you'll end up is a video that is occasionally blurry/has motion blur, in different directions, seemingly randomly. Watch even a well stabilized video (no wobbling etc.), and you'll see this. It's unavoidable.


--> mechanical SR is always superior, electronic SR can only ever compete at fast shutter speeds.

[easyreeder]

I've seen great in lens stabilization from Panasonic camcorders, so I think it's not necessarily choppy.


There are sensors with global shutters, and AFAIK they are significantly less sensitive to light. The next generation video codec is h265 aka HEVC, and it won't be superceeded for many years. But yes, I think it is sensible to expect a Pentax in 5 years to have it.


The "face tracking" (it's not really that) is a result of Pentax not using any of the sensors that track the movement of the camera, thus the camera has to guess if the camera has moved. If too much of the motive is moving, voila, "face tracking". Fujitsu would have to program a way to use the data for the electronic stabilization.


However even when you have eliminated that, and you have a fast enough sensor readout, a global shutter or software that corrects the wobble (Sony seems to have that, my Qualcomm Snapdragon 801 based smart phone too) so that the wobbling isn't an issue anymore, you'll still end up with inferior video quality. First of all it is zoomed in, you can't use the whole sensor. Secondly say you are shooting 1/50th at 25 fps. Any shaking that is faster than 1/50th will be recorded (and imagine shooting at 1/25th), so what you'll end up is a video that is occasionally blurry/has motion blur, in different directions, seemingly randomly. Watch even a well stabilized video (no wobbling etc.), and you'll see this. It's unavoidable.

This is all great information, thanks! Yes, well I was calling it "face tracking," because I can't duplicate the problem without a hand or face. I do think there's some face recognition coding in there. But it does seem fixable, because the problem goes away at about 12 feet. (I suspected this, so I did some tests with the kids, ha.)

I didn't say I'm shooting 1/50 at 25 p. I shoot 30p. But that's rather interesting. I don't mind a little blur so much, and the low light benefit of shooting 1/40 or even 1/30 for a static situation, like an interview, is something I've been experimenting with. The "blur" also looks more filmic (film has blur), and seems to me a better approximation of film than 25p, and it covers some of the shake reduction problems. I'm not a sports photographer, ha.

As for the crop, yes, there's more crop, but if you try to apply crop reduction in premiere or fcx you get an even bigger crop, and you don't get the resolution, so it seems to me like it'd be better to do it in camera. The lesser evil.

I'm not against mechanical SR by any means, haha, but I'm just trying to understand their decision making process. When the software SR works in video mode, you can get good results. Unfortunately, that's an extremely unique situation: shallow depth of field with a static object that you're focused on in the foreground.