[Kunzite]

I will quote said someone (Rishi), as it didn't sound like a guess: "K-1 II is 'star eater' on steroids" (he posted that on the forums). Later or earlier, he admitted he didn't test the K-1 II at astrophoto, which means he couldn't possibly know.

[SSGGeezer]

So I saw Ian used the 28-105 kit lens to take some images in his article. Therefore I decided to test it out last night with my kit lens.

It is horrible. Why would you use this lens? /Sarcasm




Just some basic edits in Lightroom and Nik Collection in Photoshop.

Astrotracer on. 60 seconds 53mm ISO6400 f/5/6.

I also did some stacking images and a few pano's to remove the slight streaking at the edges. I would never have considered using the kit lens for astro. But that has changed.

Too true, it is horrible! To the Pentax is DOOMED™ crowd and DPR lovers at least. Nice image! Loved his full resolution images that completely disprove the Sony problem.

[reh321]

I will quote said someone (Rishi), as it didn't sound like a guess: "K-1 II is 'star eater' on steroids" (he posted that on the forums). Later or earlier, he admitted he didn't test the K-1 II at astrophoto, which means he couldn't possibly know.

And the problem is that other detractors quoted him as though it was a proven and known fact.

[leekil]

Take a look at the ISO invariance section of the review. His image results are kind of curious, and also his comments about noise, perhaps related to the accelerator chip? I think his way of comparing the different ISOs with a composite image is interesting, and I would like to see a similar type of test for the K-1 to compare.

So this same behavior at ISO 100 seems to be replicated in this K-1 to K-1 II comparison, and the K-1 doesn't have it, only the K-1 II. I wonder if this is something that could be remedied in a firmware update?

Youtu.be

Youtu.be

[Kunzite]

Dear Rishi contacted the author which gladly obliged and updated his article:
"Update, June 19, 2018: A critical issue with the Pentax K-1 Mark II’s noise performance, particularly when stacking multi exposures was brought to my attention by Rishi Sanyal of DPReview.com. I have updated the verdict of my review accordingly and I have added an addendum to the “Image Quality; Noise” and modified the “Conclusion” section on this review."

[reh321]

Dear Rishi contacted the author which gladly obliged and updated his article:
"Update, June 19, 2018: A critical issue with the Pentax K-1 Mark II’s noise performance, particularly when stacking multi exposures was brought to my attention by Rishi Sanyal of DPReview.com. I have updated the verdict of my review accordingly and I have added an addendum to the “Image Quality; Noise” and modified the “Conclusion” section on this review."

The following words are apparently the meat of his change:
"However, the issue does create that bizarre non-random grid pattern that’s visible under magnification. Astrophotographers hoping to stack multiple exposures should approach the K-1 Mark II with consideration of this issue and understand that the K-1 Mark II may not be as good of a choice versus its older and cheaper predecessor, the original K-1. If stacking astrophotos using Astrotracer on the K-1, photographers should consider capturing a large number of frames (greater than 8) and re-aligned them in post before combining them in to a stack. With enough variation between frames, artifacts can be averaged out. Note that this technique will not work on stacks of exposures that are already aligned such as foregrounds."

Then at the end he says:
"It’s too bad that these excellent features are marred by Pentax’s poor implementation of a noise reduction algorithm. Astrotracer is particularly hindered by this issue if used for creating short stacks of exposures."
but he retained the words:
"At ISO settings higher than ISO 400, there is noticeable “baking” of the RAW file that reduces noise but also slightly affects fine detail. We did not feel that the noise reduction compromised the results of our single exposure photos in a very tangible way, we wish the behavior did not exist."

[Kunzite]

He's basing that last paragraph on DPR's "review", as he has no K-1 to compare with, and detect the reduction in fine detail.
For the same reason, he's just guessing (perhaps, just reproducing Rishi's take on this) that the K-1 II is worse than the K-1 for stacking.

[reh321]

He's basing that last paragraph on DPR's "review", as he has no K-1 to compare with, and detect the reduction in fine detail.
For the same reason, he's just guessing (perhaps, just reproducing Rishi's take on this) that the K-1 II is worse than the K-1 for stacking.

More important, the following sentences are still in his review:
" In practice, the difference is very small and I never felt disappointed with the K-1 Mark II’s capability to resolve detail at high ISOs.The noise problem does not present the same problem as the Sony star-eater issue."

[LensBeginner]

He's basing that last paragraph on DPR's "review", as he has no K-1 to compare with, and detect the reduction in fine detail.
For the same reason, he's just guessing (perhaps, just reproducing Rishi's take on this) that the K-1 II is worse than the K-1 for stacking.

Bummer...
Stacking can be done to decrease noise (average/median) or to increase signal (sum).
Is a sum of low-ISO frames, with precautions to avoid clipping, feasible, instead of an average of high(er)-ISO frames?

[Oakland Rob]

I couldn't find the language in the Lonely Speck re "bizarre non-random" grid. Was that the source of the quote?

I did see this in the comments below:

Dan,

I’ve actually been chatting with Rishi and taking into consideration your mentioning of dithering. I’ve checked the images and re-processed without the same issue becoming apparent. As a result, I’ve removed my review’s changes until further investigation is completed. –Ian

[reh321]

I couldn't find the language in the Lonely Speck re "bizarre non-random" grid. Was that the source of the quote?

I did see this in the comments below:

Please look at the first line of the "correction" I quoted in post #51 - that is the way his article read at the time I copied it.

[MJKoski]

Star eater is not the same thing. Where mk2 could cause issues are areas with almost sharp detail. Not pinpoint one pixel sized stars which the Sony issue erases. Star eater does nothing if star is larger than 1 pixel.

So, Sony nr eats away sand grains in Sahara landscape photos for example when exposure time reaches 3.2 seconds.

[Rondec]

Nice that Rishi is searching out other reviewers to make sure that they come to the same conclusions he did. I guess I shouldn't be surprised, but it does seem odd to do so .

Star eater is not the same thing. Where mk2 could cause issues are areas with almost sharp detail. Not pinpoint one pixel sized stars which the Sony issue erases. Star eater does nothing if star is larger than 1 pixel.

So, Sony nr eats away sand grains in Sahara landscape photos for example when exposure time reaches 3.2 seconds.

Welcome back.

[reh321]

Star eater is not the same thing. Where mk2 could cause issues are areas with almost sharp detail. Not pinpoint one pixel sized stars which the Sony issue erases. Star eater does nothing if star is larger than 1 pixel.

So, Sony nr eats away sand grains in Sahara landscape photos for example when exposure time reaches 3.2 seconds.

And the reviewer we're talking about here didn't see an issue with detail loss. Period.

[MJKoski]

Of course he did not see detail loss as he knows how to focus to infinity properly as is shown in the photos in that article. You see, misfocused stars count as junk soft detail to the accelerator as much as any other type of (slightly) defocused detail which cannot be easily excluded by the filter it uses. When there is not enough edge contrast anything is at danger to be seen as noise by the filter it uses. I saw it in my own test shots, Pentaxforums own test shows it and so does the "flawed" DPR review. In two months skies get dark enough here and I have a chance to experiment more.

Ian used the same ISO-invariant method to show where the accelerator kicks in as I did with the fur samples. Pushing lower ISO images to make them appear high ISO images. That ISO100 push is however weird as I cannot reproduce it with my mk2. ISO100 is still the best option for best IQ as it was with MK1.