[Rupert]

I get the AA filter issue...pretty much..... and don't consider it any likely problem for shooting squirrels.

However, seeing Falc here, I do have a question that I haven't see addressed. Is the new metering system anything radically different? I don't understand how the color sensors work in this and also how this all ties in with the "better" WB...if it does?
A little bit in this area would be of great interest to me. The simpler the better....remember, I work with squirrels, some that are smarter than me!

Best Regards!

[CRPhoto]

I suspect it's aimed at those who want to avoid Moiré fringing in large-scale printing. The higher resolution means finer detail at high magnification, which is paradoxical, as the use of sensor shake will effectively reduce the resolution. I can't imagine that some complex algorithm will restore it, but we'll have to see what the output's like in reality, before passing judgment.

I do print large, 20x30 normally and larger, but my photograpy is not prone to Moire so I guess that's why I love so much the II-s.

[falconeye]

seeing Falc here, I do have a question that I haven't see addressed. Is the new metering system anything radically different? I don't understand how the color sensors work in this and also how this all ties in with the "better" WB...if it does?

Ok, let me try then

Like the 91,000 dot RGB sensor in the Nikon D800 or D4, the 86,000 dot RGB sensor in the K-3 creates a secondary "live view" feed grabbed from the same image you see in the viewfinder, looking at the matte screen.

This secondary live feed is a very coarse 140 x 204 px color image. It is about the quality of many avatar icons or thumbs, i.e., it is just good enough to recognize what is photographed.

Therefore, it is possible to do a content analysis on images this size. I.e., look for sky, faces, snow, sand, lawn, water etc. Or shadow and sun. Allowing for smarter exposure and wb. Of course, there is a lot of magic involved in the firmware here and so, don't expect Nikon-level performance from day zero. Give Pentax a few firmware updates to fully exploit this piece of hardware.

Moreover, the prime task of said content analysis is to track the subject (3D tracking, no idea if the K-3 features it). Frame by frame, the firmware looks for the subject to be tracked (the point the AF.C locked on initially) and reports the image coordinate to the phase AF module which, in turn, determines if any of its AF points covers it. If yes, AF will be adjusted. If no, the AF stays idle (or uses prediction) until the subject moves back into the region covered by AF sensors. This enables the camera to track a subject even if It moves out of the AF area. In practice due to the low resolution, this works well if the tracked subject is well isolated against the background, e.g., by color.

Theoretically, the RGB metering sensor could have HD resolution and make most live view processing available to view finder mode.

Eventually, it will allow the camera to detect everyone's smile and trigger automatically. What a progress

[Col]

Can't wait for the manual... I'm waiting to see if you can include AA emulation strength withing the extended bracketing options.

[Kunzite]

There's also this video:

You Tube

You Tube

I don't know how well it works, but I'd say it has potential.

[FullertonImages]

I'm not sure about this to be honest. Since there seems to be very little moiré in 16MP images from the K-5IIs, I would expect even less at 24MP. So developing such a specific technology to remove such unlikely artifacts would seem excessive?

Definitely not true. The thing with moire, is that it's very very specific to each and every situation and it heavily related to patterns. So if you haven't seen it yet, then you've just been lucky enough to not shoot anything that contains a difficult repeating pattern. Man made things are especially the worst. I have no idea ho the K-5IIs compares to other cameras that don't have AA filters, but I've definitely run into situations with absolutely horrid moire problems.

It's just very specific to what seems to be a very narrow set of parameters that will just throw everything out of whack. I do a fiar amount of studio product shoot with my work for Backpacker Mag, and that's definitely where I encounter it the most. I like the use the K-5IIs as much as I can, because the details are just so sharp; but it's weird. You'll be shooting backpacks or something, and ten in a row will be completely perfect, no problem at all; and then all of the sudden one pack will have such bad moire problem that the image is unusable, and there's absolutely nothing you can do except shoot it with a different camera.

But I've even seen problem, especially using it for video, where even a normal looking t-shirts with an unusual weave will make it go crazy. It's really random, and pops up at the weirdest times, but I would be very hesitant to have a camera with no AA filter as my only camera, because you can't control when something will cause a problem even if it's very rarely, and there's often nothing you can do about it. So I'm glad the K-3 has a means of dealing with it, and I'm interested to see how well it works.

[JimmyDranox]

In the linear movement, mode 1, it don't need to be 1000Hz, to blur a picture with 1/1000 exposure time, because it's only need half of the movement, in one direction, not the full forward and backward move. On the circular mode 2, is different. 100% efficiency on one direction, and 50 % on another.

[Rupert]

Thanks Falc. That is sort of what I thought it might be, how much difference it will make over my old K5 remains to be seen, but all improvements are always welcome.

Best Regards!

[derekkite]

The specs say the AA movement system has two settings, low and high. Would this be simply more amplitude?

The varying effects over the sensor depending on the time and position of the opening of the shutter is interesting. Moire seems to occur only in specific circumstances. with a sharp pattern that has a narrower wave length than the sensor spacing. Even on clothing, the effect happens differently, depending on the angle of view, the sharpness of the focus, etc.

I guess we will see how it works in practice. There will probably be rare circumstances where the vibration creates very strange effects that wouldn't be there otherwise.

Am I mistaken to think that this would be especially valuable in movie mode? The resolutions involved are quite a bit different from the full sample resolution of a still, hence more prone to moire. From what I understand it isn't supported in movie mode.

I'm also curious how this will affect the SR function.

Probably quite a few firmware updates to come to fine tune this specific function as the various corner cases show up.

[cfraz]

Interesting - I do remember someone proposing the SR enabled low-pass filter idea before, must have been your writing, Falk.

I assume the AA simulator works simultaneously with the traditional Pentax SR. So I'm guessing the system superimposes the low amplitude, high frequency (< 1 px, 500 Hz) AA blur signal on top of the accelerometer-calculated SR signal, and that the SR signal is higher amplitude and lower frequency (not positive about that but I think I shake more than 1 px and much slower than 500 Hz). Does that sound right?

[scratchpaddy]

But I've even seen problem, especially using it for video, where even a normal looking t-shirts with an unusual weave will make it go crazy. It's really random, and pops up at the weirdest times, but I would be very hesitant to have a camera with no AA filter as my only camera, because you can't control when something will cause a problem even if it's very rarely, and there's often nothing you can do about it. So I'm glad the K-3 has a means of dealing with it, and I'm interested to see how well it works.

The AA-filter (or lack of it) doesn't have anything to do with your video symptoms, at least not on a camera made for high-resolution stills. The video you're taking is a bunch of two-megapixel images. An AA-filter tuned for 16 megapixels will have so little effect that it might as well not be there at all.

I'm pretty sure this shake-AA system will also be helpless in video. Any AA will have to be done in software.

[FullertonImages]

The AA-filter (or lack of it) doesn't have anything to do with your video symptoms, at least not on a camera made for high-resolution stills. The video you're taking is a bunch of two-megapixel images. An AA-filter tuned for 16 megapixels will have so little effect that it might as well not be there at all.

I'm pretty sure this shake-AA system will also be helpless in video. Any AA will have to be done in software.

I don't claim to be an expert on AA filters, but I'm going to have to disagree with you. Even though the video files are not using the full resolution of the sensor, they are still using the whole sensor. And since the the AA filter covers the whole filter, I would imagine that still has an effect on the video files, since the sensor is still reading all the pixel sites, which are individually affected by the AA filter, and then downsampling it all to HD resolution. But I don't know anything about AA filters, I'm just making all that up.

The point is, I've use the the K-5IIs on multiple shoots, working alongside a pair of normal K-5's; and on more than one occasion, the K-5IIs has had moire problems when the other two K-5's were completely fine. So, for whatever technical reason one would like to postulate, empirically, the K-5IIs shows itself to handle moire noticeably worse in video and photo settings alike, than a standard K-5 does.

[ChrisJ]

I have a Pentax Q, also without an AA filter, I can often see Moire on the rear screen, but when I open them in Photoshop it disappears.

This shot was full of Moire on the chair back, none on the open image.

[
]

Chris

[scratchpaddy]

I don't claim to be an expert on AA filters, but I'm going to have to disagree with you. Even though the video files are not using the full resolution of the sensor, they are still using the whole sensor. And since the the AA filter covers the whole filter, I would imagine that still has an effect on the video files, since the sensor is still reading all the pixel sites, which are individually affected by the AA filter, and then downsampling it all to HD resolution. But I don't know anything about AA filters, I'm just making all that up.

The point is, I've use the the K-5IIs on multiple shoots, working alongside a pair of normal K-5's; and on more than one occasion, the K-5IIs has had moire problems when the other two K-5's were completely fine. So, for whatever technical reason one would like to postulate, empirically, the K-5IIs shows itself to handle moire noticeably worse in video and photo settings alike, than a standard K-5 does.

Okay, the proof is in the pudding as they say, and I can't explain why you're seeing more moire (ha ha) on the K-5 iis in video. BUT, what you're describing is not how video works in a stills camera. There's just not enough bandwidth to handle full-resolution images from the sensor and process them down into a tidy 2MP image 30 times per second. They actually just skip most of the rows (and columns) when reading the sensor, reading every third row or so; that way the camera doesn't have to handle more bandwidth than it needs, but you're left with pixels too far apart for the AA filter to handle.

I have a Pentax Q, also without an AA filter, I can often see Moire on the rear screen, but when I open them in Photoshop it disappears.

This shot was full of Moire on the chair back, none on the open image.

The moire you're describing is caused by the downsampling algorithm the Q uses to display your high-resolution image on the camera's much lower-resolution screen. It only exists in the preview the camera makes for you, and it's not caused by the sensor itself.

[falconeye]

In the linear movement, mode 1, it don't need to be 1000Hz, to blur a picture with 1/1000 exposure time, because it's only need half of the movement, in one direction, not the full forward and backward move.

You have a flaw in your reasoning. I took the complete argument into account when making my post.

In linear movement mode, half the movement only works for 0° or 180° phase. Not the intermediate ones, in particular not 90° where you would move only half distance.

In circular mode, the same applies as it really is just a phase-shifted superposition of two linear movements.

If you follow this through properly, you should end up with the same figures I did. If not, please let me know