[normhead]

OK, I'm not going to play Imageman, but, it does deserve the question, if you can get an Image straight off the Pentax that's as good as what you can do with an unsharp mask, why would you do it in PP? After all, the reason most of us PP is because we can do better in post. In my mind, just as good doesn't cut it. As for your test, #4 is the best image... but, you're splitting hairs, for no good reason. Those images will all print pretty much identically... and if one is done in camera and I don't have to do it in PP, that's an advantage.

[Imageman]

I agree completely Norm you've nailed the important issue.


The reason for using this option as you say, is so you don't have to use post processing. And I wont disagree at all with that.


But there are 2 drivers for my post that I believe justify it.


First, this Pentax diffraction compensation method is only useable in a small number of lenses, Pentax have chosen to limit its use to a few programmed lenses only. So if you happen to use a different lens on your k3 than any of these your out of luck as the option just isn't there.


Second, all responders so far have said that the diffraction compensated image is massively better than an unsharp mask version, but this is after being told which is the Pentax image. Ive merely asked them to prove that the Pentax image is much better than any of the unsharp mask images by spotting it without being given a clue. I think that's reasonable. And given the earlier responses it should be very easy to spot, as its supposed to be very much better.


If nobody can reliably spot the Pentax image as a vastly superior image amongst this image set, then unsharp mask surely becomes a viable option for correcting diffraction in all the other lenses you have in your bag which don't allow diffraction correction.


So it is a worthwhile exercise I think. And its intended to be slightly light hearted. By all means use this diffraction compensation system if your lens allows it, but don't throw out unsharp mask if its a good alternative for when you cant use the Pentax method. So lets find out if it is really a viable alternative.


That's all im saying.


And thanks for stating your preferred image from the set, despite your misgivings over what must seem a trivial test.

[Michaelina2]

I have to say im truly shocked.

Perhaps you just need to calibrate your monitor? Mine is...

Cheers... M

[Imageman]

Perhaps you just need to calibrate your monitor? Mine is...

Cheers... M

Thanks for your contribution, I was wondering however, with your calibrated monitor, which of the 4 do you prefer. Its just a bit of fun, and its personal taste, but it might lead to a deeper understanding.

[Stone G.]

What were talking about here is the suggestion by some seemingly knowledgeable people that a very sophisticated solution is being applied to the image in camera, that somehow deconvolves light ray bundles to end up with the original image before it was degraded by diffraction. Whereas my contention is its simply an application of cheap and cheerful sharpening. Whats your view.

I believe that you are right: It is not some kind of advanced deconvolution.

But was is this "cheap sharpening" then? A Sony user on a domestic forum here in Denmark says the following on Sony's approach: "...it is not in-camera JPEG processing... Rather, it is part of the demosaicing algortithm at the RAW level". However, he also reports that he can get more or less similar results fiddling a bit around with unsharp mask.

Well, I cannot tell whether Sony, Pentax and Olympus all use the same approach, but it would be nice to know, what Pentax is actually applying where in the process.

Anyway, as others have already stated, it is (or can be, at times) a convenient alternative to post-processing - - - - and it is a voluntary option, not an obligation!

[normhead]

I believe that you are right: It is not some kind of advanced deconvolution.

But was is this "cheap sharpening" then? A Sony user on a domestic forum here in Denmark says the following on Sony's approach: "...it is not in-camera JPEG processing... Rather, it is part of the demosaicing algortithm at the RAW level". However, he also reports that he can get more or less similar results fiddling a bit around with unsharp mask.

Well, I cannot tell whether Sony, Pentax and Olympus all use the same approach, but it would be nice to know, what Pentax is actually applying where in the process.

Anyway, as others have already stated, it is (or can be, at times) a convenient alternative to post-processing - - - - and it is a voluntary option, not an obligation!

SO it would seem that whether or not a demosaicing algorithm that is being applied or photoshop type sharpening is at present not distinguishable without help from insiders who know what is used, however, it might be safe to assume that photoshop type sharpening has been used for years and is what it is, where as demosaicing algorithms are just getting started. Although whether or not they'll ever really surpass photoshop sharpening remains to be seen.

For most of us, it won't matter until one is the clear winner.

[kh1234567890]

What's with the crazy +2.3 EV exposure comp and odd focus point on the stamps shots ?

[wombat2go]

Here is a graph of the row of horizontal pixels through the eye.
That is from #1 of imageman's competition.
I would like to do 3 of these to maybe get idea of what the camera does:
1- the uncompensated image, 2- the camera compensated image, and 3- the usm image.
But I don't know where the original uncomp image is, nor has imageman announced the winner of his competition so we know which is the K-3 compensated image.

[rawr]

But I don't know where the original uncomp image is,

Just start from the original image in the Japanese magazine article.

[Stone G.]

Name the Pentax image is it 1 2 3 or 4

the Pentax image is 1 2 3 or 4

(YES, I cheated - and opened the images in PhotoME)

[Imageman]

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.

[wombat2go]

On the understanding that the left side images are the ones with the K_3 compensation, here are the amplitudes of a horizontal row through the eye at F/16
Left is with K_3 compensation, showing a 'peakier" response as expected, compared to right, without.

[Stone G.]

I would like to do 3 of these to maybe get idea of what the camera does:
1- the uncompensated image, 2- the camera compensated image, and 3- the usm image.

While the competition goes on, here is somtehing you might(?) be able to use. For my test, I did not use a flat stamp but a folded curtain - i.e.: with a certain depth in the fabric - to simulate a more real-life situation.

Here samples taken with the DA 18-135 at f/11 (EXIF data should be intact)


First the uncompensated image:



Next the compensated image:



Now, I couldn't get the contrast right with USM when I got the sharpness about right. You may be better at that yourself. But here are some sharpened images using FocusMagic.

Here the uncompensated image sharpened with FocusMagic:



And finally (mostly for the fun of it) the compensated image additionally sharpened with FocusMagic:

[dosdan]

I don't think using FocusMagic to demonstrate compensating diffraction softening is a good idea. I suspect that the type of de-convolution kernel designed for correcting an OOF image is significantly different from what is required to compensate for diffraction softening.

Dan.

[tabl10s]

My view, in response to your request for feed-back, is to simply observe that the Pentax approach produced a better result on my monitor than your application of an un-sharp mask. In other words, the Pentax method seems to do what is promised, that's all. Perhaps, you feel the feature 'ought' to do more, but it is what it 'is'. The two terms have different meanings... eh?

Cheers... M

...and it's free.