First thoughts on Pentax K-1 36Mp FF camera

Have questions about the equipment used for macro- or micro- photography? Post those questions in this forum.

Moderators: rjlittlefield, ChrisR, Chris S., Pau

Lou Jost
Posts: 5991
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

First thoughts on Pentax K-1 36Mp FF camera

Post by Lou Jost »

I was looking for a moderately priced mirrorless FF camera, and almost bought the Sony A7Rii. At the last minute, though, I learned about the Pentax K-1 (first version, not the Mark II) which, unlike the Sony, offered pixel-shifting technology to improve color fidelity and resolution. I've been a big fan of pixel-shifting since I got my Olympus mirrorless cameras, which do an amazing job increasing resolution through pixel-shifting. Though the Pentax has a mirror, it has a Live View mode which behaves almost like a mirrorless camera, so I decided to buy the Pentax.

I have just begun to use it, and I still have a lot to learn. These are my first impressions.

There is a lot of conflicting and misleading information about this camera on the internet. Mine is updated to Firmware 1.50, and with that update, the camera gains a fully electronic shutter, but with one peculiar behavior. In Live View, on a fully open lens, the exposure is taken without any sound or vibration. The mirror stays up all the time. Great! But as the camera begins to write the file to the card, the mechanical shutter does close, and it opens again after a short delay. It is an insane behavior. The same happens in pixel-shifting mode: the camera silently takes four pictures, and then the mechanical shutter closes and opens. For most macro work, this does provide the vibration-free photography we hope for. However, the extra vibration after the exposure can cause a subject in liquid to move slightly between frames of a stack. If I had known of this behavior, I may not have bought the camera.

There is also no information on the internet about how long it takes to capture a pixel-shifted image. I found that it takes about 14 seconds, exclusive of [shutter speed x 4], to capture and write the image onto a fast SD card. That's a long time if a big stack is being shot. A 200-image stack (typical for me) would take almost an hour to shoot, which is quite long, especially if the subject is alive like a plant. So the pixel shifting has some major practical limitations. The capture time can be reduced to 8s per shot if only a jpg is being saved to the card.

I want to use microphotolithography lenses and monochromatic light, and a pixel-shifting camera should have vastly more resolution than an ordinary camera in that case, so I would probably still have bought this even if I had known how long the pixel shift capture takes. But for most stackers, I think such a slow pixel-shift capture would rarely be worth the trouble.

I have lots of Nikon lenses and no Pentax K mount lenses. No direct Nikon to K mount adapter is possible, because the flange-to-sensor distances of the two mounts differ by only a millimeter. There are adapters that add an extension and incorporate a lens to get infinity focus, but these are optical garbage. I stripped out the insides of one of these, to give me an adapter that lets me use Nikon-mount lenses that focus by pure extension. But my plan is to replace the Pentax mount with a Nikon mount, probably using parts from the Novoflex Nikon-to-Canon mount that I have. Then the camera will be much more useful. But a Sony A7Rii would have had no conversion problems of any kind. So again, unless someone really wanted pixel-shifting badly, the A7Rii would be a better choice.

There were some complaints on the internet about the difficulty of using Live View and getting good metering data with non-native lenses. I have not found that to be any problem at all. There's a green button that sets the exposure so that it looks just like the Live View image, and at least with the subjects I've tried, it works perfectly.

Here are my first stacks of jpgs, using the default settings of the camera, which make for rather flat images. The default settings can be finely adjusted, except for the most important one, the degree of compression. I made normal and pixel-shift stacks. At this magnification, the normal and pixel-shifted images look the same.

Image

Below are 200% enlargements of the yellow-outlined portion of the image above: on the left is an enlargement from the stack of normal images, and on the right is an enlargement from the stack of pixel-shifted images. (These were made from 100% crops which were then enlarged 200% using "nearest-neighbor" extrapolation.)

Image

Here is the same at 400% enlargement:

Image
Last edited by Lou Jost on Sat Jul 21, 2018 4:08 pm, edited 1 time in total.

Lou Jost
Posts: 5991
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

I should add that the lens used above was the Printing Nikkor 105mm version A, used at about 1:1.

ray_parkhurst
Posts: 3439
Joined: Sat Nov 20, 2010 10:40 am
Location: Santa Clara, CA, USA
Contact:

Post by ray_parkhurst »

The camera does not appear to be adding any of the weird artifacts that I saw on the earlier Pentax or the A7RIII. It's hard to tell how much of the additional detail is due to more sharpening vs actual additional data, though there are indeed features on the composite that are missing on the single image. Overall I'd say this is a much more believable rendering than I saw on the A7RIII.

Did you try my suggestion of shooting with a 2x teleconverter, then downsizing 2x, to compare the added detail to see if it's real? The A7RIII did not pass this test (though I did it with an objective rather than a TC), imparting "detail" that was not actually there.

mawyatt
Posts: 2497
Joined: Thu Aug 22, 2013 6:54 pm
Location: Clearwater, Florida

Re: First thoughts on Pentax K-1 36Mp FF camera

Post by mawyatt »

But as the camera begins to write the file to the card, the mechanical shutter does close, and it opens again after a short delay. It is an insane behavior.
Lou,

Think this is common to most cameras with Electronic Shutters, the mechanical shutter is closed to block any possible image contamination during readout since the Electronic Shutter cant block something like a bright flash, or high intensity light source. The D850 does allow complete Electronic Shutter in Live View without mechanical shutter, but blocks the hot shoe flash signal.

Your PN105 is looking good :D

Best,
Research is like a treasure hunt, you don't know where to look or what you'll find!
~Mike

Lou Jost
Posts: 5991
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

Ray, I was worried that I would see artifacts like the ones you saw. What a relief when I saw the results!

Your suggestion is a good one. I should try that. But I did search for structural details that differed between shifted and non-shifted imagery, and in the clearest case that I found, the unshifted one had the false detail! I do think the jpgs are getting over-sharpened by the camera, so I will eventually re-do this with the RAW files once I have a good RAW converter for the 150Mb shifted files.

Mike, your explanation makes sense for the single-shot sequence, but what about the four shots in pixel-shifting mode? Surely the readout of the first shot has to happen before the next shot starts. But the shutter does not close until all four pictures are taken.

And yes, it is nice to have at least one lens that never lets me down.

Macrero
Posts: 1199
Joined: Thu Sep 01, 2011 8:17 am
Location: Valladolid , Spain

Post by Macrero »

I've been waiting for this thread, thanks for sharing your first impressions!

Pixel shift crops looks very good, PS clearly works. To my eyes detail is real and there is no artifacts. As you pointed out, pity is that Pentax did not implement a "real" e-shutter in Live View and 14" for taking and processing a shifted frame is quite excessive. One of the reasons I sold the Sigma SDQ was its slowness (the main reason was the mechanical shutter).

Which software have you used for processing the PS images? What about batch processing Pixel shift images?

I'm honestly still undecided which one to get, the K1 or the a7R II :? The virtually identical current price tag makes decision even harder...

Best,

- Macrero
https://500px.com/macrero - Amateurs worry about equipment, Pros worry about money, Masters worry about Light

mawyatt
Posts: 2497
Joined: Thu Aug 22, 2013 6:54 pm
Location: Clearwater, Florida

Post by mawyatt »

Mike, your explanation makes sense for the single-shot sequence, but what about the four shots in pixel-shifting mode? Surely the readout of the first shot has to happen before the next shot starts. But the shutter does not close until all four pictures are taken.
Lou,

Agree, the readout must take place before the next pixel shift image capture. So not sure what the thinking is here.

All this electronic shutter stuff seems quit mysterious to me, not the way it works but why it works in certain ways.

I'm sure a logical explanation is somewhere behind this, but cant put my finger on it just yet.

Will be very interesting to see what Nikon comes up with in their new mirrorless camera to be announce in a few days.

Best,
Research is like a treasure hunt, you don't know where to look or what you'll find!
~Mike

Lou Jost
Posts: 5991
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

Macrero, I didn't want to keep you in suspense, so that's why I put up these first impressions even though I still have a lot to learn.

Pixel shift does work and it is great except for the time delay. If you normally stack jpgs the time is cut in half. But the RAW files (even unshifted RAW files) are really impressive. Of course you never have to use pixel shifting if you don't want to. It is just nice to have the option.

ChrisR and I were discussing an interesting mount option, the Adaptist mount (https://www.imaging-resource.com/news/2 ... ut-adapter), in a private message. Since it may be relevant to you and others considering this camera, I'll post part of my response:
...they have run out of stock long ago. I jhad corresponded with the owner and he said he is trying to restart production but doesn't know when.

In any case the mount, when used with Nikon lenses, does not duplicate the correct Nikon flange distance. Infinity focus does not take place when the lens is focused at infinity but instead when it is focused somewhat closer. So the close-range correction system of these lenses will not be working optimally.

However Adaptist does sell 0.5mm wide shims for the mount, to help Nikon lenses mount properly. These aren't designed to correct flange distance, but they might be used in conjunction with other work-arounds to help space a Nikon mount on the Pentax. I have many of these shims from Adaptist in case they will be useful.

This is something like what I currently have in mind, plus shims to get infinity focus at the right place:
http://forum.mflenses.com/pentax-p30t-a ... 29918.html

Lou Jost
Posts: 5991
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

Macrero, I have tried the bundled DCU5 software, but it crashes every time. It is horrible. It's inconceivable that a good camera would be bundled with such trash. Very bad move on the part of Ricoh-Pentax. They should be chastised for it.

I've also tried ACR but I could not see much difference between the shifted and unshifted outputs. Some versions of ACR had been found to only use the first of the four images, and I may have one of those versions.

RawTherapee has been recommended but may also have the same problem as ACR. Need to test more.

Batch processing shouldn't be a problem if any of these can be made to work on single files. All allow batch processing.

Lou Jost
Posts: 5991
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

Mike, yes, all I can think of is my original adjective: "insane".

mawyatt
Posts: 2497
Joined: Thu Aug 22, 2013 6:54 pm
Location: Clearwater, Florida

Post by mawyatt »

Lou & Others,

Pardon my ignorance as I know nothing about pixel shifting nor have a camera that does such.

My question is: Couldn't you use Zerene (or other stacking tool) to just stack the individual pixel shifted images? Seems Zerene could do an image alignment to compensate for the physical sensor pixel shift, just like normally does.

Best,
Research is like a treasure hunt, you don't know where to look or what you'll find!
~Mike

Beatsy
Posts: 2139
Joined: Fri Jul 05, 2013 3:10 am
Location: Malvern, UK

Post by Beatsy »

mawyatt wrote: My question is: Couldn't you use Zerene (or other stacking tool) to just stack the individual pixel shifted images? Seems Zerene could do an image alignment to compensate for the physical sensor pixel shift, just like normally does.
Just to add to that question, if I may. If that would work, then for those of us without pixel-shifting cameras, would it be possible to take a few images with a very slight "up, down, left, right" movement between frames? It would likely be a few pixels total movement but with a good chance of sub-pixel displacement between frames (after alignment to pixel level, anyway).

Lou Jost
Posts: 5991
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

Mike, Zerene doesn't work on the RAW files, and Pentax doesn't give us access to the four individual photos anyway.

What this kind of pixel-shifting does is just fill in the Bayer-caused "holes" in the RGB values for each pixel. It replaces a triplet with two zeroes by a quadruplet with no zeroes (two independent G measurements) at each sensel. This kind of stuff has to be done at the RAW level, before debayering algorithms, if it is to have maximal effect.

In cases where the correct alignment is known exactly, I also think it would be counterproductive to use any method that has to estimate the alignment.

Macrero
Posts: 1199
Joined: Thu Sep 01, 2011 8:17 am
Location: Valladolid , Spain

Post by Macrero »

Lou, I guess I will get a K1 and try it when I have time for photos, I'm very busy with work and life this summer... Thanks again for sharing your initial impressions about the camera!

I will use the camera in studio on bellows only, so the K-mount is not a problem for me.

The consensus seems to be that RawTherapee handles best the K1 Pixel shift images. I will try various programs and see which one works best, though.

- Macrero
https://500px.com/macrero - Amateurs worry about equipment, Pros worry about money, Masters worry about Light

Lou Jost
Posts: 5991
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

Beatsy, yes, you should be able to get improved results by that procedure, but you'll need many more than four shots for it to work as well as a controlled algorithm. Its success would depend on the "law of large numbers", which is not a pleasant law when the numbers are numbers of photos per slice.

People (including me) have tried this in normal photography, with some success:
http://www.photomacrography.net/forum/v ... p?p=184546

https://petapixel.com/2015/02/21/a-prac ... photoshop/

It seems like you need around 10-20 shots at least. An algorithmic approach using known sensor displacements is much more efficient.

Post Reply Previous topicNext topic