I did wonder a bit, as it came right in the middle of a few of my posts about the tests. I like a good dabate, so felt left out I guess But thanks for the explanation... i'll try and be less sensitive next time

But test shots with completely controlled conditions will be the only way to ensure that all variables are eliminated, at least for the purposes of such detailed comparisons as we're talking about here. Assuming that the specific detail and comparisons are considered a valid and meaningful test, for a purpose, then you have no choice but to set up such contrived conditions.

OK, here's a bit of fun for everyone I'm giving you now two comparison images, two of about 30 versions in total I processed for my original testing. They come from the earlier stages, what I called 'Capture Sharpening Output'. They were both processed in Pentax Digital Camera Utility from the DNG files directly, and were saved out as 16bit TIFFs. For the purpose of this forum I have used Photoshop CS2 to produce web-sized version JPEGs, of 60% original size and about 1.2mb files, bicubic sharper - but effectively identical output processes right from the DNGs (via the PDCU TIFFs).

The only difference was that I achieved 'equalisation' I believe in PDCU by using a sharpness setting of +3 for the K5IIs shot, and +4 for the K5II shot. This is a fairly minor adjustment in this RAW converter. Bear in mind though that most likely the built-in file profiles will have already kicked in by default to go towards evening out the images. The exposure & contrast editing was my own (but identical on both) with no colour adjustments. The images were a little underexposed straight, unsurprising I guess given the bright white and with no exposure compensation applied on-camera.

But anyway, the two shots you are now going to see have only that one processing difference ... an extra +1 sharpness setting on the K5II image. Remember , these were shot in identical conditions, same lens, same camera settings - just two different bodies.


Now its time to have a look and tell me which is which

IMAGE A






IMAGE B

I think we're saying the same things : It is exactly for the reason (as you say) that 'inherant sharpness' is not the same as PP sharpening, that we can only assess the effectiveness of each image, or indeed our own need for the filter or not, by running the images through a processing chain and seeing at what point the image 'breaks down'. If this point is beyond the limits of where we usually need to go, then my conclusions were its not worth it.

Because the processes are different, and we can't make any use of straight 'capture data' without processing, we can only understand the significance of the extra 'inherant sharpness' by seeing how the images respond through the PP chain.

The fact is you can only apply extra sharpening to the K5II image, and achieve 'equalisation' (I've defined my meaning of that in detail above) up to a point. From then on, you can minimise the most obvious negative effects (like surface break-down on smooth-toned areas) by use of selective and radius-targeted approaches - and still achieve an effective, but not true, equalisation. If you can do this, then it still may not be worth it. If not, then the recomendation from the testing was for the 's'.

Like your lens analogy - yes, the inherant extra quality of the better one is always there ... but that doesn't mean that any one photographer will be able to actually gain fully from it - it would depend on what camera it was on, and whether they could shoot steady etc etc. The extra sharpness is there on the K5IIs image at capture - but nobody can directly access and use that data ... it is only through processing (either default, auto actions like camera JPEGs, viewing software, deafult RAW profiles, or direct user software controls) that we actually experience the image, and for many it would take almost no or minimal effort to 'equalise' a K5II image to the exact same image shot by an 's'. If those efforts represent the extent to which they'll ever need to go, then I'm saying you don't need to spend the extra.

Its certainly a better way to remove noise than plain old DeNoise software (and all the blurring that goes with that).

Yet another superb technical feature on our amazing K7!

And I appreciate the appreciation! I sometimes find its in short supply. I am sure there will be certain images, based on your described software processes, that will gain from the lack of blurring at capture - there will be other images that the K5II could deliver equally, but you've got the extra there for when its needed, and that is always your own personal buying choice.

Cheers !

+1 to that - thanks for sharing your decision Simon, and I too am sure you won't be disapointed. As David said earlier there are only satisfied K5II & s owners!

There is no justification for me being labelled a naysayer (I'm assuming it was directed at me). I have been here clarifying and responding to questions about my image testing .... and surely everyone will remember that there was a conclusion to all that which recommended both versions of the camera - to different people. So I have never said to anyone do not get a K5IIs. I have been suggesting the things that I think should be considered, in terms of sharpening approaches and your final output, in order to make an informed personal choice.

It is the counter 'argument' - the one that effectively says 'just get a K5IIs - it is worth the extra money, it doesn't matter who you are, what you shoot or how you process ..' - that should be questioned seriously.

But isn't it comforting to know there's always things you can rely upon?

Why not say it to us all? The OP asked the question on an open forum - he didn't send PM's out to people asking for their views - well, he didn't send me one anyway If its something about the discussion here on this thread why not contribute to it by sharing? Is it something you shouldn't say in public?

Glad to have enticed you back When I say equal or equalised I am talking about visible differences to the two images, side by side at high magnification. My testing showed such 'equalisation' at certain stages in the processing chains (as I've detailed a bit above). I also mean that the images achieved such visible equalisation without IQ degradations.

I am not referring to how the images arrived at being 'equal'. In some cases this will be without much active user input depending on your software (eg if comparing embedded camera JPEGs representing the PRIME processor working, or if using later versions of PDCU that read and act on the metadata.) But in any case, for that output stage, they are effectively 'equal'.

I have never said the filter-less system is not 'inherently sharper'. Obviously by design it is, at capture stage. But we do not use images, for viewing or anything, at capture stage .... they are all processed - that is digital capture - either by the camera processor or by us on the computer - they are even processed to be viewed - they are rendered into pixel-based files and each viewer software applies various default parameters for its display. So the extra 'inherent sharpness' of an 's' file can only be of any significance when considered through an output and processing chain, and looking at what happens to it compared to the 'non-s' image.

Regarding the amount of PP, for those who don't want to apply sharpening or don't want to do any sort of controlled detailing of their photos, it may be that the 's' version may offer more - see my conclusions in the report for more detail on that.

But I'd recommend to anyone to think about selective application of sharpening - I mean, why would you want to sharpen skys, or peoples skins for example? If you already take some control over you image sharpening then you may find that the 's' will not offer any particular advantages to you once you get to your final output ... that was my own conclusion for myself.