[maw]

Pentax vs Canon 5DS

Instead of hair, it seems interesting to me, this mini test although not a true one in all its aspects, but it is a real simulation of what you can really get with the Pentax K1 MrkII at night.

Pentax by night

[aikaarska]

Pentax vs Canon 5DS

Instead of hair, it seems interesting to me, this mini test although not a true one in all its aspects, but it is a real simulation of what you can really get with the Pentax K1 MrkII at night.

Pentax by night

Yep, but "the problem" still is that we don't know how well original K-1 handle same situations if you compare to K-1M2...They have exactly same sensor inside. Only in-camera image processing differs. But is it just a good thing? That was the question from the beginning of this thread...

[Rondec]

Yep, but "the problem" still is that we don't know how well original K-1 handle same situations if you compare to K-1M2...They have exactly same sensor inside. Only in-camera image processing differs. But is it just a good thing? That was the question from the beginning of this thread...

I think we will know soon enough as Pentax Forums has a copy of the Mk II and they will be shooting comparison shots to try to clarify things. They already have an article up about handheld pixel shift or dynamic resolution or whatever Pentax is calling it.

MJ Koski has high expectations for his gear and it may be that the K-1 (original or Mk II) just can't meet those. He was frustrated by color shifts when shooting multiple exposures in camera -- he was shooting several hundred images and the camera wasn't dealing well with it. I understand the desire to push your gear to the max, but I don't know ranting about it is helpful when you are shooting outside of a camera's specifications.

[MJKoski]

I found out what happens, at least with DNG files. Not yet camera RAW support for PEF files. The effect is very subtle but manifests itself with images where there are near-sharp details around sharp areas. ISO100-ISO400 are fine but after that, camera performs some form of edge enhancer function AND starts to blur out areas which are not 100% in focus.

Here is a crop from -3EV ISO100 exposure which is boosted 3 stops:


Here is same scene with 0EV ISO800 exposure:


Switching back and forth between the two the effect reveals itself. Boosted ISO100 exposure has tiny amount of more noise but keeps fur details intact in out-of-focus areas. ISO800 exposure starts to melt the areas almost in-focus but not fully so. This is why the original dead rabbit crop looks messed up. This is ISO100-400 camera for tripod use and to keep maximum details one should push the exposures in post.

I add this difference of the two image layers (calculated in photoshop):



It is more easy to see which areas are affected and they are the out-of-focus areas while sharp areas stay near black.

---

Last edited by MJKoski; 04-21-2018 at 03:36 AM.

[Kunzite]

Is there anything unusual in having differences between two images?

[BigMackCam]

I found out what happens, at least with DNG files. Not yet camera RAW support for PEF files. The effect is very subtle but manifests itself with images where there are near-sharp details around sharp areas. ISO100-ISO400 are fine but after that, camera performs some form of edge enhancer function AND starts to blur out areas which are not 100% in focus.

Here is a crop from -3EV ISO100 exposure which is boosted 3 stops:


Here is same scene with 0EV ISO800 exposure:


Switching back and forth between the two the effect reveals itself. Boosted ISO100 exposure has tiny amount of more noise but keeps fur details intact in out-of-focus areas. ISO800 exposure starts to melt the areas almost in-focus but not fully so. This is why the original dead rabbit crop looks messed up. This is ISO100-400 camera for tripod use and to keep maximum details one should push the exposures in post.

I add this difference of the two image layers (calculated in photoshop):



It is more easy to see which areas are affected and they are the out-of-focus areas while sharp areas stay near black.

Firstly, thanks for coming back with more examples - this is really useful and helpful

I can definitely see the effect you're describing. It looks very subtle to me, and I doubt I'd notice it at all without looking at 100% reproduction... but it's there.

Have you tried developing the ISO 800 shot from RAW to high quality JPEG in-camera, and also using DCU? I'd be fascinated to see what Pentax's own JPEG engine and software produce. Might be better, worse, or much the same... but interesting, for sure...

[Class A]

ISO100-ISO400 are fine but after that, camera performs some form of edge enhancer function AND starts to blur out areas which are not 100% in focus.

I'm not disputing your results, but here are a couple of things to make sure of:

• Are you sure your post-processing software isn't using different levels of noise reduction depending on the ISO setting? Lightroom may even show the same noise reduction setting (e.g. "0") but apply different noise reduction levels under the hood. Adobe is well-known for such techniques. ACR is one of the worst RAW converters to use for all sorts of comparisons. I recommend trying to replicate the results with a different RAW converter, such as RawTherapee.
• Have you tried adding a little noise to the ISO 800 shot? Sometimes added noise can create the illusion of more detail being discernible.

Personally, I do not like any massaging of RAW data. If Ricoh believes they have good image improvement algorithms, they should (optionally) apply them to JPGs and offer them as part of DCU. If such algorithms are applied to RAW data (as in DNG or PEF files), they should be strictly optional, in my view. Having said that, I understand that all data in RAW files has been "treated". All manufacturers do it to some extent and sometimes users or sites like DxOMark find proof of it.

As far as I'm concerned, the jury is still out whether the accelerator unit has a detrimental effect. I still hold some hope that it doesn't.

[MJKoski]

Above crops have LR detail module turned off. If something still happens by Adobe, then it is Adobes fault too. I just do not recall seeing similar behaviour with files from my previous mk-1 bodies. Maybe PEF support works better then.

I can share the RAW files tomorrow when I get back home. In DNG format (straight from camera ofc). Then you can play with them as you wish.

[normhead]

So bottom line, what do you think? Can you work with the files and get acceptable results. On my K-1 for wildlife, 800 ISO is big deal.

[Rico]

I found out what happens, at least with DNG files. Not yet camera RAW support for PEF files. The effect is very subtle but manifests itself with images where there are near-sharp details around sharp areas. ISO100-ISO400 are fine but after that, camera performs some form of edge enhancer function AND starts to blur out areas which are not 100% in focus.

Here is a crop from -3EV ISO100 exposure which is boosted 3 stops:


Here is same scene with 0EV ISO800 exposure:


Switching back and forth between the two the effect reveals itself. Boosted ISO100 exposure has tiny amount of more noise but keeps fur details intact in out-of-focus areas. ISO800 exposure starts to melt the areas almost in-focus but not fully so. This is why the original dead rabbit crop looks messed up. This is ISO100-400 camera for tripod use and to keep maximum details one should push the exposures in post.

I add this difference of the two image layers (calculated in photoshop):



It is more easy to see which areas are affected and they are the out-of-focus areas while sharp areas stay near black.

I am still not seeing what the problem is the two crops look virtually identical to me.

[MMVIII]

I don't have a horse in this race, however, some observations. This is pixelpeeping on quite high level and I notice that I can't see any difference on most of my screens, just on my imaging monitor.
I am afraid that this is running more into a principal discussion with those seeking the 'purity' of raw files and those happy with the results of clean high ISO.
I am pretty sure that any subtle effect would not be noticed at all by 99% of photographers. AFAIK no one has described this with the KP using the same processing engine, but there just seems to be agreement that it's IQ is outstanding in real life! But I'm sure we will run into arguments with the K-1, just because it can be well compared to the original version. I just hope some contextual sanity keeps on top. I don't mean it offensive, but a narrative that is once established in a forum (and here it seems to be a finnish forum) is gonna go on for ever, even if the evaluation of it's basis might have went on and changed the situation.
Saying that, I think that the OP's work is stunning, and in most cases I'm happy if tools are properly tested, if only that I dont have to go this route and destroy my equipment to know it's limits.

[MJKoski]

To be fair, if scene has large depth of field no-one will see anything strange. Only when printing large and using high ISO the *current* DNG conversion with Camera RAW is tricky, or Raw Therapee for that matter. I can reproduce this finding with any image where there are lot of high frequency content and thin depth of field by using ISO800+. Hairy subjects seem to be good canditates.

Firmware update would be nice with on/off setting for this magic chip. Post-processing easily produces as good or better results with right tools.

[normhead]

I am still not seeing what the problem is the two crops look virtually identical to me.

The top one is crisper, the issue is it's not a an area where if you are appreciating the artistic value of the image that you're likely to pay much attention to.

But the main issue fo me is like that old joke.
"Doctor it hurts when I do this."
"Don't do that."

I've never found cause to reduce an image 3 stops and then boost it 3 stops in post. Those are never good images. The simple fact is, that's the same as claiming the image is 800 ISO, when it's really 6400 ISO.

Bottom line, it's not good photographic practice, unless you are desperate.

I mean really, who does that? Apart from MJKoski
You have to find data relevant to your way of operating a camera. This certainly isn't relevant to me or I'm guessing most people. Most of us realize, it's usually best to pay attention to your light meter, and follow it's advice as closely as possible. That let's you take advantage of the characteristics the engineers who designed the camera built into the system to help achieve optimal performance.

MJKoski has other ideas, I'd be interested in hearing what they are.

PS... most of us will be happy to accept this lack of performance when 3 stops under-exposing and then going 3 stops over in post, if it means we get better performance when we operate the camera according to specification and don't depend on PP to try and bail us out.

After all, who wants a camera that's good at what it's not designed to do if it means they don't get a camera that's better at what it's designed to do? For the serious photographer, the question is how do i get the best image? Trust me, it's not under-expose 3 stops and then boost three stops in post.

The exercise simply isn't meaningful for those seeking the best images.

Although I am waiting to here from MJKoski as to why he feels these tests are in any way meaningful. I won't close the door on this until he's had his say. Maybe he thought of something I didn't. It's possible,

[texandrews]

I guess I'm probably a fanboy, although I haven't posted on this thread. Honestly, I have a hard time seeing the issues that some people are claiming. There may be a bit of softening, but I'm of the impression that neither photo is great nor would they look good printed at 100 percent. My feeling is that pixel peeping really high iso images is a fairly frustrating experience -- regardless of the camera. But looking at other threads where folks have posted high iso shots of night skies, I don't detect any stars being eaten or the like.

But I'm not an expert on the subject either. I'm happy with my K-1's performance and might get a K-1 II down the road if auto focus is a bit better.

My looking parallels yours so far.

[MJKoski]

I lost photographically good images thanks to Sony filtering back in 2015. I never expected Pentax to go that route. And I have only one purpose for a digital camera - to produce something usable when 4x5 cannot be operated in satisfactory manner to keep quality up. My standard print size is 1 meter wide and up from there and this looks like something that actually will show up. Now I know I will use max ISO400 in any case, also in astrophotography with this new mk2.

The purpose of the comparison was based on the idea that no filtering happens at base ISO and when boosted one can use ISO invariance capabilities to crank up the exposure and avoid any damage done by overzealous filtering.