Would you consider a mirrorless FF ?

Can't agree more. I'd go one step further... interchangeable optical viewfinder like the old Nikon F-1 series so you could have an APS-C viewfinder for those APS-C only lenses, FF viewfinder, Square VF, 4/3 viewfinder, whatever suits your aspect-ratio and shooting style (studio etc.) fancy. If Pentax did this, it would compete with (and exceed) the best of FF DSLRs offerings out there, and wouldn't have to upgrade it's lens lineup at once.

sjgcit said:

No.

If I ever see a Pentax FF I want interchangeable focus screens and a full sized optical viewfinder ( 100% baby ! )

A mirror-less EVF FF would be like that other crime - sports cars with automatic gearboxes.

--
StephenG

Pentax K100D
Fuji S3 Pro
Fuji S9600

Click to expand...

--
Roger

A large sensor = a superior optical viewfinder, so the bigger you go, the less reason there is to use EVIL.

--

K10D, Sig 17-70, DA 55-300, FA 50/1.4 "Billy Bass", M 400/5.6 "the Great Truncheon"

zepzr said:

Aside from size argument, contrast detect AF requires differently designed lenses. The old lesnes that work fine with phase detect AF would take 2 to 3 seconds to AF on mirrorless cameras.

Click to expand...

Wait, what is the basis for this? Is it really a problem with the lens, or is it a problem with the current state of contrast detection firmware?

My eyeball focuses old lenses just fine by contrast detection.

zepzr said:

I am not sure about the technical detail, but the lenses must be redesigned for them to work effectively with contrast detect AF.

Click to expand...

If you're not sure about the technical detail, then maybe you should either go learn the technical detail, or leave off posting an unevidenced assertion as if you knew whether it was true or not.

The digital revolution of many years ago meant more than just hanging a digital sensor on a Rube Goldberg mirrorbox. No mirror, short flange distance to eliminate retrofocus wide angle designs, and a good EVF like the Panasonic G1 and I would be thrilled!

zepzr said:

I can't think of a single plus for using K-mount on a mirrorless camera.

Click to expand...

The one advantage it has is that it allows a smaller company like Pentax to get into the market in a competitive way with no lag time. Any other option and they are forced to design a good sized set of new lenses. With this approach they can throw a body out there and improve the lenses later. They could sell a new body with just one example of a new-style wide angle lens.

But this argument works much better for an APS sized sensor than for an FF one. Besides the fact that Pentax's current lineup hardly supports FF at all, FF has less to gain from a short glass-to-sensor distance (which in the case of the K-mount would mean the back of the lens protruding inside the mount flange) and also less to gain from an electronic viewfinder.

I do think, though, that as others have said, in the long run electronic viewfinders will climb up from small formats to large ones, eventually being adopted even for medium format.

paulkienitz said:

A large sensor = a superior optical viewfinder, so the bigger you go, the less reason there is to use EVIL.

Click to expand...

Except for the fact that there are a lot of people out there WANTING AN EVF.

So yes, there is a very good reason to put out an EVF camera, even if you personally don't like them.

--
J. D.
Colorado

paulkienitz said:

zepzr said:

I am not sure about the technical detail, but the lenses must be redesigned for them to work effectively with contrast detect AF.

Click to expand...

If you're not sure about the technical detail, then maybe you should either go learn the technical detail, or leave off posting an unevidenced assertion as if you knew whether it was true or not.

Click to expand...

I am pretty sure it's true. The lenses must be redesigned specifically for contrast detect AF. Want proof? Stick 4/3 lens on EP2 and the AF will be just as slow as Pentax Kx in live view. Now stick a Panasonic m 4/3 lens on the same camera and the AF would be lightning fast. For contrast detect AF, most of the trick is in the lens, not in the camera.

A mirrorless camera with K-mount is utterly stupid idea. Use the live view on Kx instead! There you have the mirrorless camera.

All the advantages to mirrorless camera come from a new mount and new lenses (fast contrast detect AF, small size, a lot of MF lens options). Otherwise, use live view on KX (that's mirrorless too since mirror is down). Why do you need a new mirrorless camera with K-mount when you have Kx/K7 with live view?

they do not made it like the Leica do .. the principle idea here is compact as OP stated. In past eon of film days, there is the Contax G and Leica M. But there is also the like of Ricoh GR series and Contax T3 ..

They all tell us that SLR is not the only form factor that quality are seek from a tool ( and imaging capability as well as image quality ). That philosophy need to be carried over to the new product if and Pentax decide to do one.

Though, with Samsung's suposed to be around FF sensor development still very much in development. I wonder if Pentz would had the source and resource to do such a product.

--

• Franka -

paulkienitz said:

zepzr said:

I can't think of a single plus for using K-mount on a mirrorless camera.

Click to expand...

The one advantage it has is that it allows a smaller company like Pentax to get into the market in a competitive way with no lag time. Any other option and they are forced to design a good sized set of new lenses. With this approach they can throw a body out there and improve the lenses later. They could sell a new body with just one example of a new-style wide angle lens.

Click to expand...

You guys wanting a mirrorless camera with K-mount are utterly clueless.

You already have two cameras with K-mount that are mirrorless. It's called liveview! The mirror is down (and plays no part) when Kx & K7 are in live view.

Why do you want Pentax to release a Kx with no mirror? Why not just use the live view on Kx instead?

All the real advantages to mirrorless camera require a new mount and new lenses (fast contrast detect AF, small size, a lot of manual focus lens options - all these come from a new mount and new lenses. k-mount mirrorless is K7/Kx in liveview, and you already have it.

Hi Dominique,

I also second you. No EVF for me. No way for now.

no, never

Hi C,

Yes you're right. It would be better that way. A shorter flange sensor distance would allow rangefinder style lenses - ie true wide angles instead of the retrofocus designs used for SLRs. We might finally see an end to rectilinear distortion!

Cheers, Rod

zepzr said:

paulkienitz said:

zepzr said:

I am not sure about the technical detail, but the lenses must be redesigned for them to work effectively with contrast detect AF.

Click to expand...

If you're not sure about the technical detail, then maybe you should either go learn the technical detail, or leave off posting an unevidenced assertion as if you knew whether it was true or not.

Click to expand...

I am pretty sure it's true. The lenses must be redesigned specifically for contrast detect AF.

Click to expand...

What on Earth is the difference? What makes one optical design work and another not? What even could make one work and another not? The only optical property that might affect it is the bokeh, and people aren't going to change their bokeh preferences for the sake of a focusing implementation that's going to be superceded in a couple more years.

zepzr said:

Want proof? Stick 4/3 lens on EP2 and the AF will be just as slow as Pentax Kx in live view. Now stick a Panasonic m 4/3 lens on the same camera and the AF would be lightning fast.

Click to expand...

That's not proof, that's just a data point. One kind of camera has a problem when it uses the wrong lens. That's no evidence at all that making a version that works properly requires any specific kind of lens.

zepzr said:

For contrast detect AF, most of the trick is in the lens, not in the camera.

Click to expand...

Now this is just factually false. A blur circle is a blur circle, no matter how you build the lens, and the principle of adjusting focus to minimize it is simple and straightforward, widespread in naturally evolved eyes, and the way all camera lenses were focused before the eighties. And the information that is needed to do this comes 99% from the sensor and 1% from the lens. Just because doing this digitally might currently require an uncomfortably large magnitude of parallelized computation doesn't mean there are whole classes of optical design that don't work with it. It just means that somebody tried to take a shortcut and it choked.

While I think Mr Z overexaggerates the problems (for example, even the fastest contrast-detect AF lenscurrently available for EVils is not "lighting fast", as far as I am concerned), there is more to a lens than optical design - what you are focusing on.

paulkienitz said:

paulkienitz said:

I am pretty sure it's true. The lenses must be redesigned specifically for contrast detect AF.

Click to expand...

What on Earth is the difference? What makes one optical design work and another not?

Click to expand...

It's not about optical design, it is about the electronics.

Current phase detect AF lenses tell the body who they are and how they need to be focused. Remember all those backfocus/frontfocus threads and focus adjustments?

That's because the existing implementation of phase detect works by a feedforward control, using data sent from the lens to the camera. This seemed great when it was designed: Phase detect AF can not only say how in-focus a shot is, it can even tell the direction. So, if you know how you need to focus the lens, the camera basically does this focus movement and once focused there it will think it's done.

Remember the complaint that certain brands do not verify focus? It's usually possible because of this trick.

With contrast-detect AF, you do not know how in-focus or not a shot is. You need a close-loop control to make sure it is in-focus. Part of the info that the lens told the phase-detect camera is now useless and other info that the lens does not tell would be needed for a better contrast-detect algorithm.

However, it is my rather firm believe that the current implementation of CDAF cameras with PDAF lenses is poorly done, for whatever reasons.
Here, I agree with your point:

paulkienitz said:

A blur circle is a blur circle, no matter how you build the lens, and the principle of adjusting focus to minimize it is simple and straightforward, widespread in naturally evolved eyes, and the way all camera lenses were focused before the eighties. And the information that is needed to do this comes 99% from the sensor and 1% from the lens. Just because doing this digitally might currently require an uncomfortably large magnitude of parallelized computation doesn't mean there are whole classes of optical design that don't work with it. It just means that somebody tried to take a shortcut and it choked.

Click to expand...

--

'Well, 'Zooming with your feet' is usually a stupid thing as zoom rings are designed for hands.' (Me, 2006)
'I don't own lenses. I pwn lenses.' (2009)
My Homepage: http://www.JensRoesner.de

paulkienitz said:

paulkienitz said:

I am pretty sure it's true. The lenses must be redesigned specifically for contrast detect AF.

Click to expand...

What on Earth is the difference? What makes one optical design work and another not?

Click to expand...

There is a difference. If there was no difference, then liveview on Canon, Pentax, and Nikon DSLRs would focus just as fast as they do on m 4/3, Samsung NX, and Sony NEX, but they don't. They are slower -- much slower.

There is a difference between lenses designed for contrast detect AF and phase detect AF. Just releasing a mirrorless camera with K-mount won't make the AF any faster than in liiveview on DSLRs. That makes the idea of K-mount mirrorless a pretty dumb one. You will need new lenses for contrast detect AF anyway.

Ah, but Don,

the current LV cameras have CDAF more as an afterthought than as a fully operational AF - it's poorly executed.
It sure can be improved and with appropriate R&D even drastically so.

With screw-drive AF, things are even more easy - if the manufacturere wanted. All you'd need is a dedicated control algorithm and maybe a new motor driver. All that is camera-related and would work with "old" lenses perfectly fine.

Now, SDM lenses, the story might be different. Some black magic going on there. And maybe that's their reason for putting up that "special lenses needed" FUD: How'd they argue that the top-of-the-line SDM lenses suck on CDAF, when the screwdrivers snap into focus...

Are we again buying into the "Oh no, you need to upgrade to this and that" lure/trap?

Cheers
Jens

--

'Well, 'Zooming with your feet' is usually a stupid thing as zoom rings are designed for hands.' (Me, 2006)
'I don't own lenses. I pwn lenses.' (2009)
My Homepage: http://www.JensRoesner.de

JensR said:

Ah, but Don,

the current LV cameras have CDAF more as an afterthought than as a fully operational AF - it's poorly executed.
It sure can be improved and with appropriate R&D even drastically so.
With screw-drive AF, things are even more easy - if the manufacturere w

Click to expand...

Liveview on DSLRS is not just an afterthought. Sony installed a second sensor on cameras like A350 and A550 just to get faster AF liveview. If the solution was so simple, wouldn't they have implemented that instead of using a secondary sensor for live view?

In September they are releasing movable pellicule mirror cameras. What problem are they trying to solve? Again, AF in live view and video. If the solution was this simple, wouldn't they have just implemented that instead of spending millions on R&D for alternative?

Why would companies consider it just an afterthought when that's a major obstacle for p&s users who are used to using lcd?

I think it's a much more serious problem then a problem of just an "afterthought" for companies.

JensR said:

Now, SDM lenses, the story might be different. Some black magic going on there. And maybe that's their reason for putting up that "special lenses needed" FUD: How'd they argue that the top-of-the-line SDM lenses suck on CDAF, when the screwdrivers snap into focus...

Click to expand...

According to DPR staff (Andy Westlake), SSM, SDM, USM are terrible for conrast detect AF

http://forums.dpreview.com/forums/read.asp?forum=1037&message=35346582

JensR said:

Ironically ring-type ultrasonic motors, which are the best for PDAF, turn out to be terrible at CDAF, as they just can't manage the smooth, continuous motions required.

Click to expand...

JensR said:

If you get the chance, you can appreciate the problems Sony faces by playing with a Canon/Nikon/Olympus/Pentax DSLR that can CDAF in live view with an ultrasonic lens. Focusing is jerky and painfully slow, taking several seconds, and not necessarily even critically accurate. Manual focus ends up being faster and more accurate, especially with well-implemented live view magnification (which Sony has got right on the NEX interface).

Click to expand...

It's a far serious issue than just an afterthought or companies just trying to make "money" with new lenses.

KingDon said:

It's a far serious issue than just an afterthought or companies just trying to make "money" with new lenses.

Click to expand...

It's true that today's CDAF blows. But this does not mean that the cure for it is a different lens. I believe the core problem is that this job is straining the capacity of today's sensors and image processors, which are called upon to evaluate the entire image in a complex way in a very brief interval of time. Future generations of chips are going to have to have faster clocks, wider buses, more processing pipelines, and so forth, in order to be able to "watch" the focus in close enough to real time to operate the focus motor efficiently. A specially tuned lens might allow one to fake it sometimes with today's chips, but this isn't a long term solution... only when the muscle is there to handle the job in a general way is the technology really going to work properly.