[JPT]

Regarding imageman's challenge, I can't see much difference between them. But just because the end result looks the same, that doesn't mean it is produced by the same means. It seems to me that both techniques are aiming to make the image sharper, so I'm not surprised the results are so similar.

Nothing here demonstrates that lens-specific information is not used in whatever method Ricoh is using here. I think the accusation is that Ricoh is just making stuff up to limit the use of diffraction correction and make the newer lenses look more attractive in comparison to used lenses. I don't think you're any closer to proving that. And I don't think it would be in character for Ricoh to do that. What other features are artificially disabled for old lenses? I would tend to take Ricoh's word for it.

So I see this as being a very useful feature, which means better out-of-camera results for many users.

[awaldram]

Ive not closed the comp because so far ive only had a couple of entries, ill let it run a little to get a consensus.


I don't need many but 3 or 4 entries aren't enough to see if you guys are sharp eyed enough to see a difference.


The real question of course is are the unsharp mask versions close to the pentax version in quality or is the pentax version miles better.


And I know this is being picky, but id prefer you to judge with your eyes rather than use a sophisticated image analysis tool to make the judgement.

I am lost here

Pentax introduce a procedure that is supposed to increase resolution for diffrection limited images.

you in contention produce 4 images from a low resoltuion target and ask which is from the Pentax procedure.

Given no part of your presented image would have been blurred or obscured by difraction at any aperture just what are you trying to show ??

If what you showing is that unsharp mask sharpens images just as Pentax procedure does then yes but Dah pretty obvious, If you trying to show unsharpmask will recover as much lost detail the your 'test' images are a fail as there is no lost detail to recover anyway.

---------- Post added 18-07-14 at 11:33 AM ----------

Regarding imageman's challenge, I can't see much difference between them. But just because the end result looks the same, that doesn't mean it is produced by the same means. It seems to me that both techniques are aiming to make the image sharper, so I'm not surprised the results are so similar.

Nothing here demonstrates that lens-specific information is not used in whatever method Ricoh is using here. I think the accusation is that Ricoh is just making stuff up to limit the use of diffraction correction and make the newer lenses look more attractive in comparison to used lenses. I don't think you're any closer to proving that. And I don't think it would be in character for Ricoh to do that. What other features are artificially disabled for old lenses? I would tend to take Ricoh's word for it.

So I see this as being a very useful feature, which means better out-of-camera results for many users.

I think you give more credit than is due

Pentax are a commercial organisation they will do/say anything legal to increase profits within their own moral compass.

give ca correction is enabled in the da18-250 but not in the identical tamron 18-250 Pentax has clearly shown their view on in camera correction and lens compatability.

so I'd say it is in chracter for Ricoh and any other manafacture to promote sales of their own lens by camera additions , I'd also say I see nothing wrong with this it fair pay for the R&D used to develope the feature.

Besides the very fact the camera identifies the lens prior to 'correction' means in all probablilty MTF from the lens is used in algorythmns so Pentax claims are undoutedly true. (maybe not required but true none the less).

[Na Horuk]

Well hey, this discussing is going into absurd places. This is for jpegs, it is meant to be used like CA and distortion correction. It will improve photos for many people who shoot jpeg. But did you honestly expect a magic bullet in a firmware update? No, its like those other functions - great to have! But if you shoot raw and do your own processing, you can remove CA, distortion, correct shadows, vignetting, etc. on your own, with whatever software and method you want.
But I do believe this diffraction correction uses lens-specific profiles, taking into account the settings and scene. As always, it is possible to use other methods to get similar results. That doesn't make the feature worthless, though

[Imageman]

Hi awaldram,


Lots of people are misinterpreting what I have said and id like to put the record straight.


This discussion started with the contention that Pentax were doing something magical that is not anything to do with sharpening. It doesn't really matter if those claims are correct.


The Pentax solution is being put forward as something beyond sharpening and that it somehow achieves the look of the original image not possible with sharpening.


What im exploring is the other side of the coin, which many people have stated, and that is that simple sharpening can yield similar results, and this simple test will I hope show that the images whether the Pentax system is used or unsharp mask is used, are all pretty similar in results.


The reason for doing this is that the Pentax system is only available in the k3 and for a handful of lenses, whereas sharpening in post is available to all lenses on any camera.


The reason for trying to show this is to hopefully prove to a casual shooter that rather than abandon any hopes of compensating for diffraction with his k3, because he isn't using an approved lens, he can use unsharp mask and do the same thing or pretty close.


And any shooter with a k5 or a k10 or whatever doesn't have to wait for Pentax to ship a diffraction correction firmware, or buy a k3, he or she can just sharpen to get rid of the diffraction satisfactorily.


I think it makes sense if you are bothered about diffraction, to use unsharp mask and solve the issue rather than ponying up for a k3 simply to get this feature.


Hopefully showing that unsharp mask can help reduce diffraction, might encourage casual shooters to start post processing, rather than waiting for a firmware upgrade to offer what they can already do with photoshop and a bit of practice.


Of course its a reasonable feature to have and I am not in any way suggesting it isn't and im not suggesting its in any way bogus.


As for your suggestion that these images didn't suffer diffraction in the first place, that's incorrect The images are taken from the original post that showed diffraction effects at a range of apertures and the successful correction of it, using the Pentax system.


One of these 4 images is the image captured with the feature tuned on, and the other 3 are the soft diffraction affected image sharpened in 3 intensities in an attempt to correct for diffraction in post.


All of the images therefore are corrected for diffraction. I was hoping for more than 2 responses by now however.


So far if I were to end the competition now after 3 days, with only 2 responses, I would have to conclude that the result is, unsharp mask is able to deliver usable diffraction correction indistinguishable from the Pentax method.


Considering that when I told people which image was the Pentax one in an earlier post they all said how much superior the Pentax image looks compared to the usm version, im disappointed that now when I don't tell them which one the Pentax image is, they suddenly have great difficulty spotting it. If it was that good and obvious in the first place, how come only 2 people have a view now and the opinions now differ.


Im not suggesting that Pentax is not using a sophisticated alternate method, and im not saying simple sharpening is superior, im simply saying that I think unsharp mask is adequate for the task of controlling diffraction. And I think this little test so far confirms that.


And yes the original image did suffer diffraction, its in the original post alongside the Pentax corrected image and is the f22 image in the image set.

[JPT]

One of these 4 images is the image captured with the feature tuned on, and the other 3 are the soft diffraction affected image sharpened in 3 intensities in an attempt to correct for diffraction in post.

This is the point, isn't it? You had to make a judgement as to what degree of correction to apply based on the image. This will vary based on the lens and aperture value, so the camera cannot make this judgement without human input. That's why Pentax has tested all current lenses at different apertures, made a judgement about the optimal amount of diffraction correction to apply at each setting and included this in the lens profile in the camera firmware. That's why it won't work with lenses that are not profiled or lenses that don't report aperture value to the camera.

So I don't think the limitation to FA ltd, DFA and DA lenses is artificial. You can argue that they should profile older lenses and third-party lenses, but they have to draw the line somewhere and digital era lenses seems a sensible rationale to me. If they had to test every A, F and FA lens, the feature would take that much longer to implement.

I know there is often scepticism when camera makers offer this kind of automation in the camera, but there are a lot of users, myself included, who just don't want to spend a lot of time in front of the computer. If my next Pentax has diffraction correction, I'll be very happy (not planning to get a K-3). From the camera makers' point of view, they don't control or have responsibility for what you do in PP, but they probably do see it as their responsibility to make sure that the out-of-camera results are as good as possible. I think that is their motivation for adding this kind of feature.

[Rondec]

Hi awaldram,


Lots of people are misinterpreting what I have said and id like to put the record straight.


This discussion started with the contention that Pentax were doing something magical that is not anything to do with sharpening. It doesn't really matter if those claims are correct.


The Pentax solution is being put forward as something beyond sharpening and that it somehow achieves the look of the original image not possible with sharpening.


What im exploring is the other side of the coin, which many people have stated, and that is that simple sharpening can yield similar results, and this simple test will I hope show that the images whether the Pentax system is used or unsharp mask is used, are all pretty similar in results.


The reason for doing this is that the Pentax system is only available in the k3 and for a handful of lenses, whereas sharpening in post is available to all lenses on any camera.


The reason for trying to show this is to hopefully prove to a casual shooter that rather than abandon any hopes of compensating for diffraction with his k3, because he isn't using an approved lens, he can use unsharp mask and do the same thing or pretty close.


And any shooter with a k5 or a k10 or whatever doesn't have to wait for Pentax to ship a diffraction correction firmware, or buy a k3, he or she can just sharpen to get rid of the diffraction satisfactorily.


I think it makes sense if you are bothered about diffraction, to use unsharp mask and solve the issue rather than ponying up for a k3 simply to get this feature.


Hopefully showing that unsharp mask can help reduce diffraction, might encourage casual shooters to start post processing, rather than waiting for a firmware upgrade to offer what they can already do with photoshop and a bit of practice.


Of course its a reasonable feature to have and I am not in any way suggesting it isn't and im not suggesting its in any way bogus.


As for your suggestion that these images didn't suffer diffraction in the first place, that's incorrect The images are taken from the original post that showed diffraction effects at a range of apertures and the successful correction of it, using the Pentax system.


One of these 4 images is the image captured with the feature tuned on, and the other 3 are the soft diffraction affected image sharpened in 3 intensities in an attempt to correct for diffraction in post.


All of the images therefore are corrected for diffraction. I was hoping for more than 2 responses by now however.


So far if I were to end the competition now after 3 days, with only 2 responses, I would have to conclude that the result is, unsharp mask is able to deliver usable diffraction correction indistinguishable from the Pentax method.


Considering that when I told people which image was the Pentax one in an earlier post they all said how much superior the Pentax image looks compared to the usm version, im disappointed that now when I don't tell them which one the Pentax image is, they suddenly have great difficulty spotting it. If it was that good and obvious in the first place, how come only 2 people have a view now and the opinions now differ.


Im not suggesting that Pentax is not using a sophisticated alternate method, and im not saying simple sharpening is superior, im simply saying that I think unsharp mask is adequate for the task of controlling diffraction. And I think this little test so far confirms that.


And yes the original image did suffer diffraction, its in the original post alongside the Pentax corrected image and is the f22 image in the image set.

I don't think it is a big deal. I think you are right. Pentax did not introduce some kind of focus magic like deconvolution algorithm, but that said, a lot of people don't like to post process or, don't own a program to do so. Being able to do this in-camera is handy for those who happen to sail in that boat. If you are mad-good at photo shop, you can just keep all the these sorts of things turned off on your camera. Not a big deal either way.

[Stone G.]

The Pentax solution is being put forward as something beyond sharpening and that it somehow achieves the look of the original image not possible with sharpening.

Did/do anyone actually insist on that?

I recall that prior to the release there were guesses what this might be and what could possibly be done (now and in the future, given adequate processing power). But I see nowhere claims now that this IS in fact something very advanced and/or magical?

PENTAX/RICOH themselves certainly didn't make any big marketing "roar" upon the release of this feature. It is just a single line-item in the 645z specs. and just a note upon the K-3 firmware 1.10 release.

I would still very much like to know, what exactly is done where in the image developing process. But until we know for sure, we may yet discuss the results: Is it any good and is it convenient as compared to 'conventional' post-processing techniques?

[Imageman]

The linked Japanese article has some more information about the diffraction correction added in firmware 1.10.

Some points of interest:
- the correction applied is tailored to the lens used, not just a generic unsharp mask
- supported lenses are DA, DFA, FA limited
- diffraction correction is only possible because of the power of the latest generation of processors
- it makes a noticeable difference with apertures of f8+
- the correction is very effective up to f16 and makes f22 more usable

http://m.dc.watch.impress.co.jp/docs/review/minirepo/20140717_658292.html
(I don't know why, but I have trouble posting links to Google Translate on my phone - sorry)

The important statement is -"diffraction correction is only possible because of the power of the latest generation of processors".


In camera sharpening has always been offered so this statement is saying as plain as can be, that diffraction correction is very advanced and demands high processing power only now available in the latest cameras. This is not saying it might be or it could possibly be its saying it is.


I realise that this is not JPT declaring this, the published article that he quotes is. but the statement exists and can only mean as I said - "The Pentax solution is being put forward as something beyond sharpening and that it somehow achieves the look of the original image not possible with sharpening".


I have no issue with opinions that it may be an advanced algorithm beyond normal sharpening, I have my own opinion too I suspect it to be a sharpening algorithm applied in varying strengths depending on the lens and aperture. Simple quick and effective.


This is supported by the fact that a simple zero cost firmware upgrade applied it retrospectively. A zero cost software change cannot install new hardware so it was more than likely a rework of the way the camera sharpens rather than something very advanced. But thats just my opinion and probably not worth much given the lack of information.


I just think statements that only the newest breed of camera can apply diffraction correction to be rash and misleading until we know what is really happening.

[kh1234567890]

How much extra processing time does this correction take ?

Would anyone with a K-3 like to time it ?

[rawr]

A zero cost software change cannot install new hardware so it was more than likely a rework of the way the camera sharpens rather than something very advanced. ... I just think statements that only the newest breed of camera can apply diffraction correction to be rash and misleading until we know what is really happening.

As I mentioned in a previous post, diffraction correction or 'lens resolution correction' as they call it is now built into the latest '7th generation' Fujitsu Milbeaut image processor. That chip only came out in September 2013 and (so far) has only found it's way into the K-3 and 645Z.

So the statement '"diffraction correction is only possible because of the power of the latest generation of processors" is simply telling it like it is. This is the culprit:

[aglet]

the 2nd one is the diffraction corrected one

FWIW, the 4th one is just as appealing

I don't think anyone was really arguing that either method relegates the other to the dustbin of time, it's just another option.
A proper diffraction correction algorithm may be able to do a more accurate job, with fewer data-damaging artifacts, than USM.