[normhead]

I disagree that these could (should) be applied automatically as presets. In my experience, the individual files would need their own tweaks, and that would require looking at different bits of the image individually. So, by my calculations, add another 1-2 minutes per shot that you have determined is worth working on. I think it's at least an hour's extra work for 60-100 images, and for some shoots could be more than that. The time adds up fast. NR is a PITA in quick turnaround situations, again in my experience. YMMV.

Or you can just realize that an average preset is just things you've noted need adjusting in avery image, and understand that a preset is a starting point, you still need to work on the file afterwards. A good preset saves work on things that always need to be done. But each file needs some custom attention. That's why it's called a preset.


It's what you do before you "set" the image.



---------- Post added 06-02-18 at 11:40 AM ----------

Switching off *all* processing in Darktable except for white balance (so not even loading a base curve) results in the following histograms - first the K-1, followed by the K-1II... there are indeed minor differences.

I wonder if you took 5 images with the same camera and settings if they'd all be the same. Outdoors in daylight your odds of that happening would be small.

---

Last edited by normhead; 06-02-2018 at 08:42 AM.

[house]

You're welcome

And now, here's the good news. I've been working with the ISO 6400 K-1II image in Darktable for a couple of hours, and it seems to respond rather well to additional processing, such that detail can actually be recovered (or, more accurately, emphasised - as it doesn't seem to have been lost in the first place).

Here's a crop of the K-1II image with no processing options applied except for VNG4 demosaicing, base curve and white balance:



Note how the magenta coat looks smudged and lacking in detail.

Here's the same file but with chroma noise reduction and edge boosting using Darktable's "Equalizer" module, plus a little extra sharpening at low radius:



Notice how the magenta coat now shows considerably more detail and less smudging than in the first crop? Believe it or not, it's chroma noise reduction that is primarily responsible for bringing out that detail. The extra sharpening just gives it a final lift.

For comparison, here's the original K-1 image, with chroma and luminance noise reduction added to approximately match the K-1II's out-of-camera noise levels:



I'd still say there is just the tiniest bit of extra detail in the original K-1's file, but it's inconsequential IMHO.

Now, as I mentioned, this processing was carried out in Darktable. I'm not sure what other RAW development tools are capable of, and I no longer have Windows 10 and Lightroom installed, so I can't test that. But at least we can say that with appropriate RAW development software, it seems there's more detail in the K-1II's images than is apparent in DPR's test shots

Why vng4? Amaze is better at bringing out detail.

[BigMackCam]

I disagree that these could (should) be applied automatically as presets. In my experience, the individual files would need their own tweaks, and that would require looking at different bits of the image individually.

But that's the case with any image file, whether it comes from the K-1, K-1II or any other camera. By applying this processing as a preset (or, as I implied, a series of ISO-specific presets), you're just getting the image(s) to a good starting point.

You could choose not to apply default presets to any image batches, in which case those from the K-1 will have a tiny bit more detail but more chroma and luminance noise. Conversely, those from the K-1II will have a tiny bit less detail, and less noise...

[house]

It has never been established that there are any cameras with no NR applied in their high ISO images

Photons to Photos establishes precisely that. No need arguing this point, it's rather well researched by that site.

[BigMackCam]

Why vng4? Amaze is better at bringing out detail.

Comparing between the two demosaicing algorithms in Darktable, they brought out equal amounts of detail on this test shot.

Generally, I prefer VNG4 because it produces less obvious artefacts. AMaZE can result in a slight "digital grain" to solid colour and grey areas, so whilst it may resolve details slightly better in some circumstances, there's a trade-off.

[swanlefitte]

BigMackCam,
I really appreciate your efforts. Interesting stuff. So thanks for taking the time.

[Kunzite]

A score of 80... if they'd fix the AF test, maybe it'll get to 81. If they fix other things...

[texandrews]

Or you can just realize that an average preset is just things you've noted need adjusting in avery image, and understand that a preset is a starting point, you still need to work on the file afterwards. A good preset saves work on things that always need to be done. But each file needs some custom attention. That's why it's called a preset.


It's what you do before you "set" the image.

Yes, I do understand that. With me, and my practice, as I tried to make clear, NR is the thing that requires more massaging on a per image basis than any other adjustments. Thus, the K1MKII changes (I did not say "improvevments") I think are going to be beneficial for me in the noise department, according to what I use my K1 for, as opposed to spending more time in post for these images. I probably would not go for such an upgrade, based on what I've seen, on my Z.

[ffking]

A score of 80... if they'd fix the AF test, maybe it'll get to 81. If they fix other things

Totally off topic, but I've been meaning to ask for ages- how come your english is so good? - better than most people here!

[texandrews]

But that's the case with any image file, whether it comes from the K-1, K-1II or any other camera. By applying this processing as a preset (or, as I implied, a series of ISO-specific presets), you're just getting the image(s) to a good starting point.

You could choose not to apply default presets to any image batches, in which case those from the K-1 will have a tiny bit more detail but more chroma and luminance noise. Conversely, those from the K-1II will have a tiny bit less detail, and less noise...

See my response to Normhead; I find that NR is the problem child that for me defies using it as a preset. This is the main reason the K1MKII upgrade looks good to me, and somewhat worth it for the shots I need to do at ISO's above 640. Again, YMMV.

[BigMackCam]

See my response to Normhead; I find that NR is the problem child that for me defies using it as a preset. This is the main reason the K1MKII upgrade looks good to me, and somewhat worth it for the shots I need to do at ISO's above 640. Again, YMMV.

Understood. Just bear in mind that if you want to recover the loss of detail people are talking about in the K-1II files, you'll still need to perform chroma noise reduction as I demonstrated above, or apply some other technique.

In fact, I like the output from both the K-1 and K-1II for slightly different reasons. Initially, I was convinced by the K-1II, but my first processing attempts on the K-1 files were giving me better results, so I switched to a slight preference for that model. However, now that I'm getting to grips with processing of K-1II files, I'd be equally happy with either camera. The differences are tiny, though there are very slight advantages and disadvantages to both in terms of IQ, IMHO. But the K-1II also has improved AF (from what I've read) and hand-held pixel shift, which is pretty neat - and I rather like the noise reduction, though I'd prefer if it was possible to switch it off if desired.

This is all pretty academic for me, as I won't be moving to Pentax for my full-frame needs in the foreseeable future - but I'm sure I'll make the jump one day

[BigMackCam]

BigMackCam,
I really appreciate your efforts. Interesting stuff. So thanks for taking the time.

I'm glad it's of interest, Chris. Many thanks for the kind words

[Rondec]

I think it is all but assumed that if you are shooting at iso 6400 or above that you will need some noise reduction. The question is whether you can do it in camera or choose to do it in post. Probably much of that decision depends on the individual photographer -- their skills and what software they have available to them. But studying these green feathers, every single one of them looks a bit blurred and has some loss of detail due to using higher than base iso.

[BigMackCam]

studying these green feathers, every single one of them looks a bit blurred and has some loss of detail due to using higher than base iso.

Indeed. And by the time you add just enough NR to the K-1 files to get a similar noise profile to the K-1II, the detail loss is - if not exactly the same - very similar. For those who value lower noise right out of the camera, the K-1II looks like a great choice. It's all about individual priorities.

[Not a Number]

You could always use a filter to add random noise.