[MJKoski]

Noise does not ruin anything. If you print big it adds some raw grit to the image which is good as long as it is luminance noise. This why large format and some edgy processing with high contrast developer like Rodinal looks delicious. Same principle happens with analog synthesizers, they usually have white noise generators and dude they pack some punch if used right.

Blurring luminance noise makes the image appear cell-phone like as happened with the rabbit scene. Combine blurred texture with edge enhancement or clarity filter and the outcome is more or less a mess.

[BigMackCam]

Noise does not ruin anything. If you print big it adds some raw grit to the image which is good as long as it is luminance noise. This why large format and some edgy processing with high contrast developer like Rodinal looks delicious. Same principle happens with analog synthesizers, they usually have white noise generators and dude they pack some punch if used right.

Blurring luminance noise makes the image appear cell-phone like as happened with the rabbit scene. Combine blurred texture with edge enhancement or clarity filter and the outcome is more or less a mess.

We're kind of back where we started, with you comparing the K-1II's images to a cell-phone, and me disagreeing wholeheartedly for what I consider to be a gross exaggeration. Similarly, I see no "mess" in the example shots I've reviewed - though I do see some very slight evidence of the effect you describe at 100% reproduction.

When all's said and done, it's going to be a subjective thing for the individual photographer. Some will prefer the K-1 files, others will prefer the K-1II. You're in the K-1 camp, and that's fine. Different strokes for different folks, right?

I do think it would be great if somehow, given the image pipeline of the new camera, the noise reduction in the image accelerator could be disabled or adjusted. That would give us the best of both worlds... maximum detail for those who require it, and better noise handling for those who prefer that.

Hopefully we'll find out soon enough whether that's possible or not...

[maw]

Noise does not ruin anything. If you print big it adds some raw grit to the image which is good as long as it is luminance noise. This why large format and some edgy processing with high contrast developer like Rodinal looks delicious. Same principle happens with analog synthesizers, they usually have white noise generators and dude they pack some punch if used right.

Blurring luminance noise makes the image appear cell-phone like as happened with the rabbit scene. Combine blurred texture with edge enhancement or clarity filter and the outcome is more or less a mess.

We're kind of back where we started, with me disagreeing wholeheartedly on your cell-phone comparison I see no "mess" in the example shots I've reviewed. I do, however, see some evidence of what you describe. When all's said and done, it's going to be a subjective thing for the individual photographer. Some will prefer the K-1 files, others will prefer the K-1II. You sit in the K-1 camp, and that's fine. Different strokes for different folks

In other words Democracy, fortunately Pentaxforums is all this...until now.

[Oakland Rob]

I appreciate some of the reviewers' concerns, but frankly in the real world the other variables just seem to overwhelm some of those. Like the obvious differences in the lenses; they outstrip whatever the NR is doing in DPR's test scene. Even in just that scene I see sections I like better with either the K-1 or K-1ii depending on what bit I'm looking at, whether it's on the edge, etc. TL;DR: I just can't care about differences that small since they almost never will affect any images I want to make.

That being said, I guess reviewers have to talk about something, and camera makers' hype invites stuff like Where's Waldo hunts through handheld PS images for defects. And yeah, if I did action or video I'd prefer a Nikon or Sony. Perhaps these cameras are all so good there isn't much point to sweating the smaller details, esp since one has alternatives, even within the same brand.

[Class A]

Ricoh should have made the "accelerator" processing optional. Everyone would have won.

As it is, the K-1 II got slammed by DPReview and that won't help sales at all.

I haven't looked at DPReview's review in detail, it may be flawed in some aspects, but as there is no doubt that the accelerator processing mainly constitutes some form of noise reduction, I cannot complain about DPReview when they point this out and criticise it. Restrict such measures to JPGs or at least make it optional for RAW files.

Let's hope Ricoh will give us the option to turn off the accelerator processing via a firmware upgrade or at least make it optional in future camera models. For better or for worse, DPReview's opinion matters for sales and it is completely unnecessary to give them ammunition like that.

[maw]

Images downloaded in DNG unfortunately, but the difference is real. 100% cropping, ISO 12800.

[madbrain]

Ricoh should have made the "accelerator" processing optional. Everyone would have won.

As it is, the K-1 II got slammed by DPReview and that won't help sales at all.

I haven't looked at DPReview's review in detail, it may be flawed in some aspects, but as there is no doubt that the accelerator processing mainly constitutes some form of noise reduction, I cannot complain about DPReview when they point this out and criticise it. Restrict such measures to JPGs or at least make it optional for RAW files.

Let's hope Ricoh will give us the option to turn off the accelerator processing via a firmware upgrade or at least make it optional in future camera models. For better or for worse, DPReview's opinion matters for sales and it is completely unnecessary to give them ammunition like that.

What is odd is that DPR did not criticize the accelerator in their reviews of the K-70 and K-P vs previous APS-C cameras.
So, is the accelerator only hurting with the K-1 IQ ?

[rawr]

Images downloaded in DNG unfortunately, but the difference is real.

Interesting. Both here and in the case of rabbit fur, their NR algorithms do seem to have a problem with rendering fluffy organic matter.

[madbrain]

Images downloaded in DNG unfortunately, but the difference is real. 100% cropping, ISO 12800.

Can we really be sure the difference is solely the camera, though, when they didn't even use the same lens in both tests ?
Also, others on DPR forums seem to have pointed out other differences in the methodology.

[maw]

Interesting. Both here and in the case of rabbit fur, their NR algorithms do seem to have a problem with rendering fluffy organic matter.

This is just a sample image and may not be very explicative, I explain not having other real images in Pef, I looked for an area where there were fine details and poor contrast "as you see the green is dominant" to try to understand how the accelerator operates.

My opinion is that in continuous tone zones it cannot separate fine or small finely details. Vice versa it does better than the K1 and the images appear apparently more sharp,
with high ISO. Of course it would take multiple files in Pef possibly and evaluate the high ISO, to understand how it will behold.

[Class A]

What is odd is that DPR did not criticize the accelerator in their reviews of the K-70 and K-P vs previous APS-C cameras.

Maybe they should have criticised them, maybe not. I don't know what kind of smoothing issues the K-70 or K-P may or may not have.

Mandatory RAW file manipulation is bad in my book, whether its Sony with their "star eater" issue, Sony with their prior lossy RAW compression, or Ricoh with non-optional smoothing.

If the K-70/K-P deserved criticism in that regard then perhaps they were lucky to get off the hook. Even if that's the case, though, then it does not invalidate the fair criticism by DPReview regarding the mandatory RAW manipulation by the K-1 II.

[ogl]

What is odd is that DPR did not criticize the accelerator in their reviews of the K-70 and K-P vs previous APS-C cameras.
So, is the accelerator only hurting with the K-1 IQ ?

There is software for this accelerator. Ricoh decided to produce the leader of low noise, but....the solution is too rough.

[rawr]

Really...?

Clearly DPR needs to add an in-house squirrel to their test team.

[Kunzite]

Images downloaded in DNG unfortunately, but the difference is real. 100% cropping, ISO 12800.

Is it? The K-1 II is slightly out of focus - they botched the test.
Check the corners, especially the top-right one is awfully OOF.

[BigMackCam]

Images downloaded in DNG unfortunately, but the difference is real. 100% cropping, ISO 12800.

I don't doubt that the effect is there, but there's something not right with the test photo conditions, as @Oakland_Rob noted. If you compare the central area of the K-1 and K-1II shots, they're both sharp, but move out to that greenery in the top right and the K-1II shot looks blurred even at base ISO. The lens is playing a part here, IMHO. That's not to say the noise reduction isn't having an impact - it obviously is, but I don't feel we're comparing like for like optically and that's skewing the results somewhat...