[Ash]

Slightly blurry images will be slightly blurry no matter what sensor is used to capture them. The theory that moire is a major problem overruling the virtue of added sharpness from AA filterless sensors is difficult to argue against if it is true. Real world shooting just doesn't exemplify this phenomenon, and yet the added sharpness is not immeasurable between the filtered and filterless versions of the Sony 16Mp sensor.


Ash.

[gaweidert]

Probably a stupid question, but does moire show up in prints as bad as it does on a computer monitor? I took some shots at a local living history museum with my K5-IIs and when I opened them with the software package that came with the camera I thought I had made a big mistake in purchasing the K5-IIs. The amount of moire was amazing. However, once hit the "+" button the enlarge the print a little the moire all but disappeared. At least some of the moire seemed to be related to the monitor as it too uses a grid pattern. I was doing this on my HP laptop.

From my job as a field repair tech working on platesetters, I know that there is a technology in the printing industry called stochastic printing. The image goes through a RIP that generated a psuedo random pattern of dots. Unlike line printing where the dot size varies to control the amount of ink placed on a page, in stochastic printing the dot size remains the same and you put more dots where you want more ink. In fact it is generations better than line printing at it's highest resolution. I have seen and image made on high quality paper stock a printing press that was about 6" by 8". It looked like it was contact printed from the original negative. Blew me away with the quality.

Not everything is printed using stochastic printing because it really pushed printing press technology and older presses and those not maintained to the highest levels can have issues with it. Especially when used at it's highest resolution. Anybody out there who works in pre-press and in familiar with a product called Staccato is familiar with this process.

Using this technology on the better "photo quality printers" out there would produce image quality that may rival conventional photographic prints. I may also make moire a thing of the past.

[Anvh]

Using this technology on the better "photo quality printers" out there would produce image quality that may rival conventional photographic prints. I may also make moire a thing of the past.

With printing yes, but if moire is in the photo itself then that is another story.
Some are experiementing with different bayer types to minimize it's effect but most likely the "pixel race" will end this problem sooner.

[Class A]

Slightly blurry images will be slightly blurry no matter what sensor is used to capture them.

Slightly blurry images captured with a filterless sensor are less blurry compared to those that receive additional blur through a Bayer-AA-filter. That's why you are seeing a sharpness gain with a K-5 IIs.

Real world shooting just doesn't exemplify this phenomenon, and yet the added sharpness is not immeasurable between the filtered and filterless versions of the Sony 16Mp sensor.

Real world shooting involves some level of blur from one source of another.

As I wrote, if your image taking is less than painstakingly optimised for sharpness, you'll typically generate enough blur for the Bayer-AA-filter to become superfluous. But if you use a tripod, flash, optimal focus, etc. and have high spatial frequency in the scene, you'll see moiré and colour artefacts. Both will appear sooner than with a regular camera whose Bayer-AA-filter is designed as a compromise between avoiding moiré and destroying image detail.

I believe there is a misconception about the potential market for cameras like the K-5 IIs. The majority of shooters will in most of the cases benefit from it (as they generate enough blur to retire the Bayer-AA-filter). It is a smaller population of shooters that is after the best image detail and knows how to obtain it that has a problem with the K-5 IIs.

A 16MP camera without a Bayer-AA-filter is an incomplete imaging system. This remains a fact, even though many people manage to create enough blur that there is no need for a Bayer-AA-filter.

Probably a stupid question, but does moire show up in prints as bad as it does on a computer monitor?

Moiré has the property that it makes a problem at the microscopic scale (subsampling of spatial frequencies) visible at the macroscopic scale (mesh like patterns). So if you are unlucky, large areas could be affected by moiré and these will be visible in prints.

Color artifacts that do exist in details (e.g., colour were there should be just a sharp B&W transition) that do not combine to create an emerging moiré pattern, will go unnoticed in most prints because they are rarely big enough to allow such details to be discriminated.

However, once hit the "+" button the enlarge the print a little the moire all but disappeared.

It is most likely that the moiré you've seen was generated by a crude scaling algorithm.

In other words, the displaying of the image is to blame, not the image (or camera).

Using this technology on the better "photo quality printers" out there would produce image quality that may rival conventional photographic prints. I may also make moire a thing of the past.

Yes, stochastic printing rocks.

But it cannot do anything against moiré in the source.

[gaweidert]

Class A,

Thanks for the answers. Been taking photographs for over 40 years but pretty new to serious digital photography. I am still not sure I like it batter than film. The random distribution of silver halide crystals throughout the various layers of the film emulsion makes it, in my opinion, a far superior method of image capture. Much better resolution and far fewer imaging artifacts.

[krebsy75]

Class A, superb explanation.

[Class A]

The random distribution of silver halide crystals throughout the various layers of the film emulsion makes it, in my opinion, a far superior method of image capture.

It is probably fair to say that it is more "organic".

Did you try to take high ISO shots? The resulting noise has a grain-like effect, in particular when you convert to B&W. If you shoot in colour then chroma noise reduction becomes pretty much a must.

For B&W film fans, I also highly recommend NIK's Silver FX software. I've tried to emulate its results using Lightroom's channel mixer but with mixed results. I really like their conversion technology. The software also offers the emulation of various film types, but I haven't played with that much.

Class A, superb explanation.

I'm glad you found it useful.

[ak_kiwi]

I believe there is a misconception about the potential market for cameras like the K-5 IIs. The majority of shooters will in most of the cases benefit from it (as they generate enough blur to retire the Bayer-AA-filter). It is a smaller population of shooters that is after the best image detail and knows how to obtain it that has a problem with the K-5 IIs.

A 16MP camera without a Bayer-AA-filter is an incomplete imaging system. This remains a fact, even though many people manage to create enough blur that there is no need for a Bayer-AA-filter.

Well put Class A - this and the 5 points you previously mentioned go some way to explaining why my shooting has improved with the K-5 IIs.
When I start getting moiré I know I will have perfected my technique (yeah right.....)

[atarget]

Probably a stupid question, but does moire show up in prints as bad as it does on a computer monitor? I took some shots at a local living history museum with my K5-IIs and when I opened them with the software package that came with the camera I thought I had made a big mistake in purchasing the K5-IIs. The amount of moire was amazing. However, once hit the "+" button the enlarge the print a little the moire all but disappeared. At least some of the moire seemed to be related to the monitor as it too uses a grid pattern. I was doing this on my HP laptop.

From my job as a field repair tech working on platesetters, I know that there is a technology in the printing industry called stochastic printing. The image goes through a RIP that generated a psuedo random pattern of dots. Unlike line printing where the dot size varies to control the amount of ink placed on a page, in stochastic printing the dot size remains the same and you put more dots where you want more ink. In fact it is generations better than line printing at it's highest resolution. I have seen and image made on high quality paper stock a printing press that was about 6" by 8". It looked like it was contact printed from the original negative. Blew me away with the quality.

Not everything is printed using stochastic printing because it really pushed printing press technology and older presses and those not maintained to the highest levels can have issues with it. Especially when used at it's highest resolution. Anybody out there who works in pre-press and in familiar with a product called Staccato is familiar with this process.

Using this technology on the better "photo quality printers" out there would produce image quality that may rival conventional photographic prints. I may also make moire a thing of the past.

Just FYI moire pattern in printing process is not related to differences between linear and stochastic method. It's about the way the relative angling of screens laid to build the pantone color array. And that didn't change with introduction of stochastic methodology.
p.s. I used to work as a 4/c stripper in various prepress locations for several years

[alamo5000]

Slightly blurry images captured with a filterless sensor are less blurry compared to those that receive additional blur through a Bayer-AA-filter. That's why you are seeing a sharpness gain with a K-5 IIs.



Real world shooting involves some level of blur from one source of another.

As I wrote, if your image taking is less than painstakingly optimised for sharpness, you'll typically generate enough blur for the Bayer-AA-filter to become superfluous. But if you use a tripod, flash, optimal focus, etc. and have high spatial frequency in the scene, you'll see moiré and colour artefacts. Both will appear sooner than with a regular camera whose Bayer-AA-filter is designed as a compromise between avoiding moiré and destroying image detail.

I believe there is a misconception about the potential market for cameras like the K-5 IIs. The majority of shooters will in most of the cases benefit from it (as they generate enough blur to retire the Bayer-AA-filter). It is a smaller population of shooters that is after the best image detail and knows how to obtain it that has a problem with the K-5 IIs.

A 16MP camera without a Bayer-AA-filter is an incomplete imaging system. This remains a fact, even though many people manage to create enough blur that there is no need for a Bayer-AA-filter.

I have been seriously considering upgrading to the IIs (in fact to get two bodies)... but here is another spin on that comment you made...

How much does lens quality play into that scheme of things? IE I have some lenses that are in and of themselves not 'as sharp' as some others just by the inherent nature of the lens itself. On the flip side would lenses that are known to be super sharp, or at least a lot sharper, be easier to screw up shots with the IIs?

[Class A]

On the flip side would lenses that are known to be super sharp, or at least a lot sharper, be easier to screw up shots with the IIs?

Yes.

Of course, if you are using high f-ratios such as f/11 or higher then you are save due to diffraction. Wide open most lenses introduce enough aberration to avoid issues, but great lenses peak at f/4 already so there is quite of range of f-ratios that may cause problems.