[texandrews]

I think the accelerator looks for clear edges in the image and if it does not find them then the area receives NR filtering. It may be possible to fine tune the triggering level. By no means this is ordinary global NR function but precise local filtering with possible edge enhancement. Very slow to implement on desktop post processing software with traditional CPU.

So far, this is the most interesting thing to me (clearly not to others...)in this thread, and a statement that may indeed describe something like what we are seeing, not just here, but across all the test shots we have seen so far. Also, this would I think have to be a combo of the accelerator unit and new or tweaked processing.


It gives me a headache to have read through all this stuff, and then have to compare it to problematic through botched(DPR) test shots---and of course DPR's arrogance always gives me a headache, as it does when it comes from governments, rich people, academics, engineers, and other entitled groups. This second decade of the 21st century has really gotten depressing....


But back on topic, I think you have indeed hit on something true here. We've seen this before: remember the first FF Leica and the blacks problem? This is similar in that it's maybe a problem that is appearing in a variable way according to the subject matter.


Now, let's all hope that we get a decent test out of Imaging Resource, and a full one here at Pentax Forums. But, if the above is true, then these tests will need to include shots that do and do not induce the problem.

[MJKoski]

You can do similar thing right now with Nik Raw Presharpener. It does not reduce noise but insteads skips it depending on your settings. The filter is VERY slow as plain photoshop plugin and takes huge amount of time to complete with typical 60-80 MP 4x5 scan I use it for these days. Works very very well when tuned for the image at hand.

[normhead]

Now, let's all hope that we get a decent test out of Imaging Resource, and a full one here at Pentax Forums. But, if the above is true, then these tests will need to include shots that do and do not induce the problem.

I've read posts from at least two users who haven't experienced the problem, or any problem and who are happy with the additional noise reduction and better AF. It has yet to be determined if there is an actual problem with the camera or if we have shoddy test procedures employed by the testers.

I end up thinking myself, am I more like the guys who report problems or the guys who don't? my feeling is, I'm much more likely to be using my camera the way the guys who don't have problems do, especially when I look at all the trouble folks have gone through to try and establish that there's a problem and how much erroneous, uninformed information they've posted, most of which is completely open to interpretation. Simply stated. I don't engage in those kind of behaviours. A problem is only problem if I can't fix it with minimal amount of effort.

If you go looking for problems you'll probably find some, with any system. You still have to look at the big picture and ask yourself, do I really care? IN the overall scheme of things, with this thread, I don't care. Sometimes the problem is that people can trick themselves into thinking there's a problem, or even if there is, that it's in some way going have meaning in their lives.

I haven't seen anything in this thread a few seconds with definition and contrast brushes and a cloning tool wouldn't fix, in seconds. I've probably spent more time trying to understand this thread than I'll ever spend correcting these "problems" in Aperture.

Compared to the electrical wires, out of focus flying insects, tree branches etc. I regularly deal with in photos, these "problems" are laughable.

Just be careful you aren't being mislead by people with much different expectations than your own. The fact that they believe there is something wrong, doesn't mean you should. I'd recommend this thread to anyone shooting 4x5 film who just uses their K-1 as their "cheap and dirty" camera. I'm not sure it's relevant to anyone else.

Oh wait, that would be a null set.

[Gimbal]

The difference in focus point is because the camera was set to refocus with AF with the remote control. I could reshoot without that setting.
I had to touch the camera to change the settings - ISO and PS/no PS - and slightly moved things when I did. I would need to use tethered shooting program in order to avoid that problem to meet your high standard, I think. I am just a hobbyist. Never thought I would buy a $2000 camera like the K-1 II.

Do you think my FA50 / 1.4 is not a proper lens ? I got it for $100 two weeks ago. It looks fine to me at f/1.8 . Not so sharp at f/1.4, which is why I stopped it down to f/1.8, but it depends on the subject.
I could have gotten more of the frame in focus by stopping it down further, but chose not to.

So, since you played, let me reveal what the files really are :
1. ISO 800, no PS
4. ISO 500, no PS
You got those two reversed. Do you like #1 more than #4 ?
And if so, did you think the accelerator was in picture #4 and hurting the quality ?

2 is indeed ISO 500 PS
3 is indeed ISO 800 PS

You got those right. I'm impressed you got 50% right. Do you prefer #2 over #3 ?

Overall, #2 seems to be the most popular, as ranked by the other players - LensBeginner and Mallee boy.

I haven't actually pixel-peed the shots myself. All 4 look very good to me. I knew what each one was, so there was no purpose in playing. Happy to play the game with your shots, though.I think I would be very happy with any of them unless I did huge enlargements, and even then, not sure.
I don't think I would have fared well in this game, but it was your game.
I can make the rendered JPGs and DNGs available if anyone wants them. Probably going to be about 500MB.

One thing I can say is that the profiles in RawTherapee significantly changed the color vs what was initially displayed in RT, and when rendered.
This color shift does not occur when I use Aftershot at default settings. But aftershot doesn't support PS, so I couldn't use it for the PS test.

In any case, your other claim was that PS at ISO 800 was making thing worse, vs no-PS at the same ISO 800 .
Do you still see that in these shots ?

One interesting thing is that the rendered PS files - regardless of ISO, and regardless of original RAW or rendered only JPEG, (obviously RAW PS files are larger)- are significantly smaller than non-PS files.
That means some information got eliminated. Yet, people still seem to prefer the ISO 500 PS image #2, which is the second smallest of all 4 JPEGs.
So, the PS must be eliminating some unwanted information somehow.

Here are the sizes of original RAW files :

05/11/2018 17:38 42,859,883 ISO500.DNG
05/11/2018 17:38 157,822,689 ISO500PS.DNG
05/11/2018 17:35 42,784,080 ISO800.DNG
05/11/2018 17:37 158,751,797 ISO800PS.DNG

Obviously, PS RAW files are larger than non-PS.
And here, ISO 800 files are slightly smaller than ISO 500, regardless of PS.

For JPG files :

05/11/2018 18:35 5,779,612 ISO500.jpg
05/11/2018 18:33 4,265,957 ISO500PS.jpg
05/11/2018 18:31 5,221,428 ISO800.jpg
05/11/2018 18:32 3,780,444 ISO800PS.jpg

Here, there is a significant difference between ISO 500 and ISO 800 files. The ISO 800 files are much smaller, ie. compress better with JPG algorithm.
I tried rendered in lossless 16-bit PNG also, but got files that were larger than all 4 original RAW files

I played it too, (without cheating). I looked at the noise in the dark blurry parts and got it all backwards. Picking out the PS was easy but I mixed up the 500 and 800 iso. I guess the NR made the 800 look better then the no NR 500.

[madbrain]

I played it too, (without cheating). I looked at the noise in the dark blurry parts and got it all backwards. Picking out the PS was easy but I mixed up the 500 and 800 iso. I guess the NR made the 800 look better then the no NR 500.

Thank you. This just goes to show that it's hard to claim that the accelerator is doing something detrimental to the image. And I'll feel quite comfortable shooting at high ISO on my K-1 II, even if Pentax never provides an option to turn it off. Of course, it might be interesting to repeat the test with different ISO and closer focus. It is just time consuming. I would like to shoot more interesting targets with my camera, rather than some more rather than tests to prove that the accelerator doesn't cause noticeable IQ loss.

---------- Post added 05-12-18 at 04:57 PM ----------

There is nothing wrong in boosting -3EV ISO100 to make it look like ISO800 and show that ISO100 not only keeps all detail but effectively replaces the need for artificial software gain in-camera.

It is free workaround for anyone looking for a way to skip forced processing right now. In similar fashion I would use max ISO400 to shoot milkyway etc. astro stuff and push in post.

-3EV is nothing but standard push. I can do it with film and digital right now (even witha canon). -8EV is something and still doable by this old 36MP warhorse. Either by using PS or multiexposure.

I simply don't see why you would do that. Personally, I never carry a tripod outside my house. To me, not requiring a tripod is one of the benefits of having a high ISO camera. And the stabilizer means that I don't have to buy and carry expensive and heavy glass, either. Because of those benefits, I'm more than happy to carry a big bulky camera like my Pentax K-1 I I. It still fits in my camera bag.

If I'm in a low-light situation outdoors - and that is actually my favorite type of shooting, I am simply not going to be shooting ISO 100 . ISO 3200 and 6400 were very usable on my K-30 previously, with some minimal NR in post (couldn't apply too much without blurring the images). I have done bracketing with different ISOs at +/- 1 EV, and generally found that I liked to push the underexposed images at ISO 1600 more than the ISO 6400 images. But between the -1 EV and 0EV images, there just wasn't a massive difference. The most important thing I found was using a reasonable shutter speed to avoid blur, as blurred shots just cannot be recovered in post. They just have to be discarded, unless the blur is minor that you can compensate by sharpening in post, and you are doing only 4x6 prints.
ISO 12,800 and 25,600 seem very usable on my K-1 II now. I will likely use 12,800 quite a bit. And have zero hesitation about ISO 3,200 or 6,400 . I had my K-30 set to ISO AUTO 100 - 6400 almost all the time. I have set my K-1 II to ISO AUTO 100 - 12,800 so far, and liking the results. Though I sometimes force fixed ISO, or aperture, or shutter speed. I never shot in green mode on my K-30, and won't on the K-1 II either. Most often, I shoot in P mode. Rarely use the other modes unless I'm doing something very specific like the sorts of tests I have been doing lately and posting to PF.

---------- Post added 05-12-18 at 05:06 PM ----------

Even a little bit of lens slop will throw away focus point when the setup receives a minor bumb. That 50mm you used felt like a plastic wonder when I tried it briefly in 2016 when purchasing the mk1.

Thanks. Plastic wonder seems like a fair assessment of FA50/1.4. My only other 50 is the DA50. It autofocuses better and faster than the FA50/1.4, but isn't as sharp at equal aperture.
I had an FA50/1.7 before. Lots of CA and PF with it. Had loaned it to a friend who broke it. He bought me the DA50 as a replacement. That was years ago.
I hope the D FA 50 / 1.4 will be an improvement over both lenses. I know I just won't have $500 in my budget for it. Maybe in another 10-15 years I can buy it used for $100 too

I think the accelerator looks for clear edges in the image and if it does not find them then the area receives NR filtering. It may be possible to fine tune the triggering level. By no means this is ordinary global NR function but precise local filtering with possible edge enhancement. Very slow to implement on desktop post processing software with traditional CPU.

Yeah, and you would need to pay for such post processing software. I have only used fairly inexpensive software up to this point - Corel Aftershot and Paintshop Pro mostly. Never used Photoshop before. Not really looking forward to it due to price. I heard the learning curve is high too. I would much rather have something quick like Aftershot. Too bad it is an unstable program, as is ON1 which I'm trying now. Still looking for a suitable replacement that isn't ransomware.

[MJKoski]

Practical limit of Sony 36MP was about 3200 ISO no matter what brand and it has not changed. ISO 12800 is usable for web-sized image at best.

Learning curve of photoshop is much lower than one might think. It is not expensive anymore thanks to CC subscription model and has all the tools one will need to handle all kinds of situations. Simple things are simple in the UI nowadays. About every common problem has a tutorial as a video as well.

[madbrain]

Practical limit of Sony 36MP was about 3200 ISO no matter what brand and it has not changed. ISO 12800 is usable for web-sized image at best.

I don't know how you came to that conclusion, but that seems completely wrong. I have taken plenty of pictures on my K-30 at ISO 3200 - 6400 that are great. They did need some NR. But those are photos that I could simply not have taken on my K200D, which maxed at ISO 1600. I would have had to use lower shutter speed and tripod, or more expensive lenses than I could afford with the K200D.
The K-30 was a godsend to me for indoor & night-time outdoor photography. I bought the K-1 II in large part because of the larger sensor, to get more light, but also higher ISO. My own results show there are plenty of usable pictures at ISO 12,800 . 25,600 looks usable too. If I was better at processing, ISO 51,200 and even 102,400 might be usable too.

Learning curve of photoshop is much lower than one might think. It is not expensive anymore thanks to CC subscription model and has all the tools one will need to handle all kinds of situations. Simple things are simple in the UI nowadays. About every common problem has a tutorial as a video as well.

The CC subscription model is exactly what I object to. I'm not a pro, and I find it expensive. For me photo is just a hobby. I don't sell my art or have clients, or even enter any photo contests. I only want to pay for software for upgrades for new features. I especially don't want to hold my collection of photos hostage to ransomware, which Lightroom Classic CC is. I'll gladly pay $120/every year if there are new features I want, but if not, I want to be able to keep using the software without having to pay over and over again. Of course, I'm only really interested in Lightroom, and not Photoshop or any of other programs included in the Adobe photo subscription, so $120/year for that really seems too much. Photoshop may be a great program, but the subscription is just a non-starter for me. Of course, I would never have paid $500 for the Adobe software suite. It is just not true to say that the subscription has made things cheaper - I would say it is the opposite, if one only wants one program. I think there are probably many other cheaper programs which can be replacements for Photoshop. But as far as image management, I have really found all other programs besides Lightroom very wanting.

[MJKoski]

Your photos are not hostage. It is common mistake to think they are borked once sub ends. Develop module and printing just stops working. Your archives and photos stay intact.

One of the best methods to ruin your new hobby is to start with wrong tools. For it kills your creativity and understanding what you can do in post phase. Right tools work with any source material.

[madbrain]

Your photos are not hostage. It is common mistake to think they are borked once sub ends. Develop module and printing just stops working. Your archives and photos stay intact.

One of the best methods to ruin your new hobby is to start with wrong tools. For it kills your creativity and understanding what you can do in post phase. Right tools work with any source material.

Not being able to print is a pretty damn big problem. Are you sure that is the case ? Last year I did a trial, and I was still able to print afterwards.
Maybe they have changed it since. My computer had Windows reinstalled since, which is why I am able to do another trial.

The hobby is not brand new to me. I have digital photos going back 18 years with 3MP Olympus compacts. I got my first DSLR in 2008 - a K200D. I started shooting in RAW with my K-30 in 2012. That's when I became a little bit more serious. I'm still going to keep looking for something other than Lightroom, as I simply object to the subscription model. I hope I will find it soon. So far, nothing else is looking good.

[MJKoski]

You can get LR6 standalone license which is not subbed. Locally they sell for 40-50€ or so. Ofc, then you have to use DNG or modify PEF EXIF as it works with MK1 RAW files.

Then there is one workaround for photoshop :] Get "old" cs5 license for few bucks. It works with most modern plugins and scripts still AND you can do 95% of all possible hacks and edits if you are okay with the traditional UI. Or CS2 for free and run it with virtual machine. Dunno if it is free now, it was freebie for a moment years ago.

Sorry, print should work still after sub. More info here: http://blogs.adobe.com/lightroomjournal/2014/07/what-happens-to-lightroom-after-my-membership-ends.html

[madbrain]

You can get LR6 standalone license which is not subbed. Locally they sell for 40-50€ or so.

Hmm. $149 at B&H . I would like to know where.

Ofc, then you have to use DNG or modify PEF EXIF as it works with MK1 RAW files.

Yeah, a bit of a pain. I did that for a while with my K-30 with Aftershot before there was an official camera profile for it. Would rather buy software that's still being supported, of course.

Then there is one workaround for photoshop :] Get "old" cs5 license for few bucks. It works with most modern plugins and scripts still AND you can do 95% of all possible hacks and edits if you are okay with the traditional UI. Or CS2 for free and run it with virtual machine. Dunno if it is free now, it was freebie for a moment years ago.

Thanks, I'll think about it. Why VM for CS2 ? Does it require an old OS ?

[BigMackCam]

The CC subscription model is exactly what I object to. I'm not a pro, and I find it expensive. For me photo is just a hobby. I don't sell my art or have clients, or even enter any photo contests. I only want to pay for software for upgrades for new features.

I'm in danger of boring people senseless with my posts about this, but after several years of of being a very happy Lightroom 6 user, I've switched to digiKam (for library management), Darktable (for RAW development) and GIMP (for PhotoShop-style image editing). I trialled Darktable for several months to ensure it did what I needed, before making the jump. Mine is installed on Linux, but the same tools are available for Windows.

These are professional tools in every respect. Very close in capabilities to Lightroom and PhotoShop - in some ways, not quite as good (but not far off), and in other ways, a little better. Darktable, in particular, is easily equal to Lightroom 6. Actually, in most respects I prefer it.

They're mature, stable products that are actively developed to offer new and improved features and performance at regular intervals. Oh, and they're very well documented.

Best of all, they're completely free.

If you don't want to pay for Adobe's subscription products (that's how they lost me as a customer), you might try the tools I've mentioned. It takes a little while to learn how things work, but it's not a difficult proposition - and it's well worth the effort

[MJKoski]

Gimp is powerful but has weird and confusing UI. Ferrari engine in Fiat 600 frame.

[BigMackCam]

Gimp is powerful but has weird and confusing UI. Ferrari engine in Fiat 600 frame.

I've never used full PhotoShop - only PhotoShop Elements 14 (which I didn't like much), and that may be an advantage for me, in that I have no expectations of the UI in GIMP. It seems pretty easy to use (I'm running v2.8.x), and for most of what I need to do, quite straightforward. But I can believe it might not be as polished as PhotoShop in terms of the UI. It fits my requirements, though, and I do like the price

[MJKoski]

It could clone photoshop UI as it contains mostly the same features. But it could lead to massive legal problems.