K-5IIs - more discussion

+1 for this. I thank David for his efforts.

I have a K-r and any move to any type of K5 would be a major step up. This thread, amongst others, is helping me decide where my hard-earned (well, not really) dosh is going to go when I've saved enough. The difference between a used K5 and a new or used K5IIs is quite considerable. And I don't even know what the next generation camera will be and how much it will cost, though I probably will be the time I've saved any money.

The more "evidence" of quality differences the better, in my view!

Any comments on the last two shots I posted (the ones of the cotton material)? There's no focus error on these.

Regards
David

Hi David, out of curiosity I've just tried something with those last two images, purely out of interest and in no way intended to disprove your assertion that the K5IIs is giving you noticeably more detailed images than the K5. I have also been able to see differences in the various K5, K5II and K5IIs images posted albeit I believe most of the obvious differences are the result of factors other than the undoubted technical differences between the models.

What I wanted to see was how close I could make the K5 image to the slightly more detailed looking K5IIs image. So, I downloded your image and imported into Aperture. I then selected the 'Highlights and Shadows' adjustment set to pull the shadows up about 1/4 f the full range, and applied it just to the K5 image with a large brush (about half the height of the bottom K5 image) so as to apply it to the K5 image only. I then did the same with the sharpening tool, again applying about 1/4 of the available adjustment. If this had been a single image I would have just nudged the two sliders to apply the adjustment to the whole image rather than used the brushes.

Now, this was a very rough and ready job (literally a couple of minutes) and based on web jpegs, but I think it shows that some of the detail in the K5IIs image is there in the shadows of the K5 image and differences in processing are playing a part too. As I said, I'm not trying to disprove anything or dismiss your efforts and don't doubt there K5IIs will have the ultimately better potential to capture detail, just trying to add constructively to the discussion. And what do I know anyway, I'm happy to stick with my K20D

Results of pulling up the shadows and adding sharpening to the K5 Image with the K5IIs image left unprocessed:
[IMG]http://i214.photobucket.com/albums/cc276/Crossed-up/PUF/13596_1362221769_zpsfdacb662.jpg[/IMG]

Looking at the wisps of fibres in the centre of each image,and indeed the fibres you can see in the threads in the material, there's no difference in detail that's going make difference in real world photography. And yes I magnified the images 2 fold = 200% from your 50% to avoid any doubt that my monitors or eye sight were affecting anything.

It sometimes seems to me that my contributions to this forum are just as invisible as the differences between the comparison images posted here. I'll try again by quoting from the dpreview page that I linked to in my last post:

"In theory, a sensor without an AA filter, when coupled with a 'perfect' lens, will deliver resolution equal to its Nyquist limit..."

"... vertical and horizontal detail is accurately described (by the original K-5) up to around 2800LPH, with some line detail surviving up to, and even slightly beyond Nyquist."

Refer back to that linked page for a clear description of the original K-5's resolution performance. The implication there is that the K-5, with its weak AA filter, is capable of resolving up to the theoretical maximum of its sensor's potential. It therefore seems reasonable to suggest here that removing its weak AA filter (to give a K-5IIs) results in imperceptible differences in real world photography.

Philip

Thanks to everyone for your contributions here - a lot of technical knowledge and differing opinions have been imparted, I'm sure we'll never gain agreement on the relative performance of different cameras and sensors, the main thing is that we enjoy our photography

Regards
David

Thanks for taking the time to post the images David. At least this latest K-5 v K-IIs thread is civilised

Still wouldn't mind checking out the RAWs, but no rush if you're busy

If they think you can go beyond the Nyquist limit of a sampling system then they don't know what they are talking about! It's called a limit for a reason. Anything beyond the limit will alias - I.e. appear as an incorrect lower frequency artefact, which in images we call moire. That's why you have an anti-aliasing filter: to remove frequencies higher than what can be accurately sampled before you sample them.

The point that the K5 has a weak AA filter and performs close to the theoretical limit may well be true (and backed up by it scoring higher for sharpness than other cameras with the same sensor on some sites) but to say it can go beyond that limit is just cobblers and would make me add a pinch of salt to anything else I read there.

I suspect that now that sensor resolutions are getting so high, any real lens is going to act as an adequate AA filter and we are going to see more and more filterless cameras being released (e.g. The Nikon D7100 which is only available without a filter on a 24mp APS sized sensor).

Well here is a good comparisn between mkii and mkiis sharpness and the conclusion may surprise...

http://www.digitalcamerainfo.com/Features/The-S-Is-for-Subtle-The-Pentax-K-5-II-...

I've already posted that link on page 4

-

Just to second Dr O and Darkskies... Thanks David for taking the time to post these examples and crops. Extremely helpful comparisons

Long thread lol

I've joined in late here I know - just a thought about these type of comparisons from Pentax DSLRS and files types : unless we can use two plain RAW shots and put them through identical processing parameters in a single program, I am concerned that we may be being misled by things out of our control.

I'm not sure by what process david output his JPEGs, but if the shots are produced in camera, then there is no way to avoid a high amount of automated editing being applied without the users input at all. And we have no way to be sure that apparently equivalent settings of camera parameters between two models are in fact the same at all

I don't think you've read the thread properly Nigel, the shots were RAW and have not been manipulated, except to resize them for upload here so there is no 'high amount of automated editing' happening on these images. Take a look at the last two sets I posted in the thread and I do explain this very clearly.

Regards
David