[cbaytan]

1.Don't get the K-5iis if you are going to shoot people, wildlife

3.not sure

I decided not to go for the K-5 IIs given that I do shoot people and wildlife

How in the earth a wildlife shot might produce moire? The only wildlife scene I can imagine is a huge zebra flock taken from a great distance. Any more ideas or samples?

[JohnBee]

How in the earth a wildlife shot might produce moire? The only wildlife scene I can imagine is a huge zebra flock taken from a great distance. Any more ideas or samples?

Think feathers, wings, compound eyes as well as texture or patterns etc.

[RobG]

Why would you deblurr the K-5 IIs?

Because any digital sensor will produce a degree of blur?

The same deblur is available in Raw Therapee for free.

Surely it's not the same or nobody would pay for it?

The thing we can rely on here is that detail doesn't come out of thin air. Therefore, when we deblur the K-5 II and find similar detail, we can be assured that both systems are effectively capturing similar amounts of information. However, the K-5 II comes out much softer due to the AA filter and so it becomes necessary to counter the effects of this in software. And so were left with a trade-off I suppose.

When you try to reconstruct an image blurred by an AA filter, will you reconstruct it accurately? You could if you had an accurate enough 2D fourier transform, but the processing required would be horrendous. You're welcome to your opinion of course, but it seems to me that blurring an image and then trying to reduce the blur after the fact will result in a less accurate reproduction than simply not blurring it in the first place. Yes, I know, what about the dreaded moire? As I mentioned previously, I'm not worried about it, and so far I haven't seen it. If photos are rarely ruined by it, I'd rather have the vast majority of my images sharper and more accurate with less post processing. If others are worried about moire, then buy the K-5ii. Simple.

[JohnBee]

Because any digital sensor will produce a degree of blur?

Using these particular samples as reference, I found that applying a deblur adjustement on the K-5 IIs resulted in an overcooked image. The same can be said for sharpening as the IIs which proved to have far less latitude than the K-5 II.

Surely it's not the same or nobody would pay for it?

Topaz Detail contains numerous spacial detail enhancement tools which Raw Therapee lacks. Though the deconvoluted sharpening tool is based on the same technology. When I began this case study, I was using Raw Therapee exclusively. However, a forum member asked for a none RAW alternative and so I switched to Topaz Detail as as result.

When you try to reconstruct an image blurred by an AA filter, will you reconstruct it accurately? You could if you had an accurate enough 2D fourier transform, but the processing required would be horrendous. You're welcome to your opinion of course, but it seems to me that blurring an image and then trying to reduce the blur after the fact will result in a less accurate reproduction than simply not blurring it in the first place. Yes, I know, what about the dreaded moire? As I mentioned previously, I'm not worried about it, and so far I haven't seen it. If photos are rarely ruined by it, I'd rather have the vast majority of my images sharper and more accurate with less post processing. If others are worried about moire, then buy the K-5ii. Simple.

The topic of a script of tool to counter the effects of the AA filter has come up numerous times and I'm thinking it won't be long before we begin seeing such things appear in the near future. That being said, I wouldn't obsess over theoretical output at this stage since we have samples to look consider. To which I'd add, the proof is in the pudding(so to speak), as the K-5 II image seems very well positioned to keep-up and in many areas even outperform the K-5 IIs in terms of final IQ which you yourself pointed out, as the effects of aliasing can extend beyond that of the expected color patterns we often see.

At any rate, I agree with the notion that such decisions are highly person dependent in the end. Not to mention that the K-5 IIs has the advantage of producing much sharper images OOC than the K-5 II. However, I also think it's important to put out there that the K-5 II can be processed to match the K-5 IIs in cases where shooters are willing to process their images. And so, it really comes down to what it most important to people in the end I suppose.

[cbaytan]

When you try to reconstruct an image blurred by an AA filter, will you reconstruct it accurately? You could if you had an accurate enough 2D fourier transform, but the processing required would be horrendous. You're welcome to your opinion .

I guess I know the solution, Just attach a small part of it in front of your lens. If you want more techno solution use a blur filter or a crappy lens, seriously wouldn't it work? Wish someone with K-IIs have it tested this, or I will do it in the future anyway.

[RobG]

I guess I know the solution, Just attach a small part of it in front of your lens. If you want more techno solution use a blur filter or a crappy lens, seriously wouldn't it work? Wish someone with K-IIs have it tested this, or I will do it in the future anyway.

Sorry, I don't have any access to suitable materials.

[Alizarine]

Wish someone with K-IIs have it tested this, or I will do it in the future anyway.

lol but seriously I wonder if anyone with a k-5 IIs has actually tested it on such. Like, really provoking the moire to come out of fabric like tulle or most of those in wedding gowns, suits, and patterned shirts.

[RobG]

lol but seriously I wonder if anyone with a k-5 IIs has actually tested it on such. Like, really provoking the moire to come out of fabric like tulle or most of those in wedding gowns, suits, and patterned shirts.

I don't have any materials like that but I did try a shirt with a fine pattern in the weave (not a colour variation). I got some vague colour variations which may have been moire. The moire was most obvious in the live view of the camera, not in the resulting image.

Here's the pixels:


And here's the full frame. The colour is odd because the lighting was bright tungsten and the shirt is actually pale blue.


The thumbnails in Windows Explorer show moire, the camera live view showd moire, there's minor colour shifts at the pixel level, but the picture viewed as the whole frame looks fine. Most of the moire I've mentioned has been in quick rescaling. Anyway, I don't think I've succeeded in demonstrating a problem. For comparison I really ought to try the same shirt with the K-7.

[rawr]

Remember, as Falconeye hinted in another thread, colour moire is but one type of moire, albeit perhaps the most common in digital cameras.

Maybe these don't all afflict Bayer array sensors, but based on a quick google the list includes 'maze artifacts', line moire, shape moire, shadow moire etc etc etc...

It's enough to give a K-IIs or D800E user nightmares.

[Alizarine]

I don't have any materials like that but I did try a shirt with a fine pattern in the weave (not a colour variation). I got some vague colour variations which may have been moire. The moire was most obvious in the live view of the camera, not in the resulting image.

Thank you very much for taking the time to get a sample. I still see the moire on-screen at 100% but it's far from being obvious. At least that rules out scaling artifacts from zooming in and out... what lens did you use here?

[RobG]

Thank you very much for taking the time to get a sample. I still see the moire on-screen at 100% but it's far from being obvious. At least that rules out scaling artifacts from zooming in and out... what lens did you use here?

I used the DA35 limited macro, since it's the sharpest lens I own. The test would be better repeated using daylight and a tripod.

[RobG]

Remember, as Falconeye hinted in another thread, colour moire is but one type of moire, albeit perhaps the most common in digital cameras.
Maybe these don't all afflict Bayer array sensors, but based on a quick google the list includes 'maze artifacts', line moire, shape moire, shadow moire etc etc etc...
It's enough to give a K-IIs or D800E user nightmares.

Doesn't give me nightmares. It only seems to be giving people who don't own a K-5iis nightmares.

[Alizarine]

I used the DA35 limited macro, since it's the sharpest lens I own. The test would be better repeated using daylight and a tripod.

Thanks for that info. Most moire I get from my K-x is when I shoot with a flash and it's quite different from daylight so I suppose it's going to be less of a problem under the sun. I just hope the algorithm's been adjusted for it as (per falconeye's explanation) luminance moire is more frequent than color moire when I get them.

[RobG]

Thanks for that info. Most moire I get from my K-x is when I shoot with a flash and it's quite different from daylight so I suppose it's going to be less of a problem under the sun. I just hope the algorithm's been adjusted for it as (per falconeye's explanation) luminance moire is more frequent than color moire when I get them.

Interesting! By the way I note that the Pentax Digital Camera Utility version 4.4 has a "spurious colour signal reduction" feature which I don't remember seeing on the older versions. It defaults to a setting of "80". There's also "random noise reduction", "multi-level noise reduction", "spotting" and "red-eye correction".

The "spurious" noise filter seems to be specifically for colour moire. The "spotting" feature seems to be to remove dust spots? The help file doesn't even mention the multi-level noise reduction!

[gooseta]

Doesn't give me nightmares. It only seems to be giving people who don't own a K-5iis nightmares.

lol, yes while k5iis owners like you and I state many times that moire is extremely rare, people who don't own one say there is a huge moire issue