[biz-engineer]

Does the shutter speed cause the vibrations wavelength to change or is it constant?

The question become irrelevant when there is no shutter shock.

[Oakland Rob]

Having seen those nice Dynamic Pixel Shift images earlier on, thinking I'll give them more of a shot. Seems kinda dependent on the context; my initial ones weren't really worth it so I stopped using it.

Is there anything that can post process raw Dynamic Pixel Shift? Aside from PDCU, which I find unusuable? I tried shooting DNG in camera, then creating TIFFs from those, and results weren't really worth the effort. I know Lr still only uses the first frame; dunno about Rawtherapee as I'm a think a version behind.

[jbinpg]

Nope, I believe in-camera and PDCU are your only options for Dynamic Pixel Shift as of now. Rawtherapee may be looking at it, though.

[Rondec]

Nope, I believe in-camera and PDCU are your only options for Dynamic Pixel Shift as of now. Rawtherapee may be looking at it, though.

I don't even think DCU can do dynamic pixel shift. The issue seems to be having the data on camera movement so that the images can be combined in the right way. The two options are to shoot RAW and convert it to a TIFF in camera or to just shoot jpeg and let the camera do the work at the time of shooting.

Biz-engineer is probably right though, that the main benefit is going to be in higher iso situations where you just can't hand hold a longer exposure. Taking four shorter shots should be quite feasible. If you are on a tripod, just use a longer exposure and/or shoot traditional pixel shift.

[derekkite]

Nope, I believe in-camera and PDCU are your only options for Dynamic Pixel Shift as of now. Rawtherapee may be looking at it, though.

The latest versions of Rawtherapee supports pixel shift, allowing you to merge the four images, select one, a bunch of things. The interface is reasonable. I don't know if it works on anything except linux distributions.

[Rondec]

The latest versions of Rawtherapee supports pixel shift, allowing you to merge the four images, select one, a bunch of things. The interface is reasonable. I don't know if it works on anything except linux distributions.

It works great for traditional pixel shift. I think the question has to do with "Hand held pixel shift," or "Dynamic Pixel Shift" (I think that's what Pentax calls it). Anyway, I haven't had much satisfaction in trying to develop it out of camera.

[Oakland Rob]

It works great for traditional pixel shift. I think the question has to do with "Hand held pixel shift," or "Dynamic Pixel Shift" (I think that's what Pentax calls it). Anyway, I haven't had much satisfaction in trying to develop it out of camera.

Right, exactly. The handheld dynamic version. I agree that PDCU still doesn't seem to do it.

I seem to recall at least speculation that there was metadata from the movement involved in the processing; it wasn't just a matter of stacking.

The new up-and-coming Olympus E-M1x is reported to have handheld PS. Be interesting to see.

Frankly I'd trade DPS for focus bracketing....

[jbinpg]

There has to be frame alignment done as well for DPS which is harder to do, I assume.

[Ash]

DPS is an excellent technology no doubt and has to be responsive to be able to take the image when the image has shifted one pixel only between each frame.

[monochrome]

There has to be frame alignment done as well for DPS which is harder to do, I assume.

The DPS processor uses the data stream from the IBIS engine to calculate the alignment parameters.That’s why DPS actually needs hand displacement to work.

[biz-engineer]

The DPS processor uses the data stream from the IBIS engine to calculate the alignment parameters.That’s why DPS actually needs hand displacement to work.

K1 II users have to be aware of the purpose of single exposure, pixel shift and dynamic pixel shift:
- some parts of the image I see in the viewfinder may move within 1 seconds (2x4x1/4th) or more: I use single exposure mode
- what I see in the viewfinder doesn't move during at least 1 sec. or total exposure time, I have a tripod with me: I use pixel shift
- what I see in the viewfinder doesn't move during at least 1 sec. or total exposure time, I don't have a tripod with me and the scene doesn't require fine tone curve correction: I use dynamic pixel shift

I've seen Sony A6K in camera stacking more for low light shooting without tripod: ISO is set high enough not to induce image blur, multiples noisy frames are captured and then stacked to reduce the noise.
The K-1 II does a better job by stacking images but also eventually using slightly displaced frames as oversampling technique also having the potential to render a little more detail depending on lens sharpness and aperture, so it works best at medium high ISO. And generally, pixel shift or dynamic pixel shift won't render more details at aperture f8 and smaller due to diffraction, only noise will be reduced.

Typical case where PS or DPS bring better image quality: wide angle lens , 20mm , aperture f4.
Beyond 35mm FL [for landscapes type of shots], for all in focus frame, PS or DPS won't bring more sharpness.

[Mikesul]

Biz-engineer. Are you saying that focal lengths narrower than 35mm are not useful for Pixel shift? So there would be no use to use PF with the DFA *50 1.4mm? It think that is what you wrote but I am just checking.

[biz-engineer]

Biz-engineer. Are you saying that focal lengths narrower than 35mm are not useful for Pixel shift? So there would be no use to use PF with the DFA *50 1.4mm? It think that is what you wrote but I am just checking.

For all in focus landscape, I defined my own rule of thumb for aperture settings on the K1, and in that case, FL longer than about 35mm do not benefit of pixel shift. Obviously are other cases such as photographing of a flat surface such as a painting, or photographing objects with a lots of blurred background.

[Mikesul]

For all in focus landscape, I defined my own rule of thumb for aperture settings on the K1, and in that case, FL longer than about 35mm do not benefit of pixel shift. Obviously are other cases such as photographing of a flat surface such as a painting, or photographing objects with a lots of blurred background.

Thanks for the clarification. I appreciate your insight. This may also explain why some testers would see no benefit if they are using a longer focal length for landscape.

[biz-engineer]

Thanks for the clarification. I appreciate your insight. This may also explain why some testers would see no benefit if they are using a longer focal length for landscape.

For me, pixel shift is fun to try but never made any meaningful difference vs single exposure. There's more value into Astrotracer and 36Mp FF for under 2K. Dynamic pixel shift is more value for me, because I remember traveling and wanting to photograph some architectural thing at dusk and lack of tripod was a problem; with DPS I could have bumped ISO and get away with a good picture. And since this is a thread about the K1 II, I'm happy that the K1 II is free of shutter shock, for me that's a great improvement (I don't know if the improvement came from a firmware change , like a slight delay between mirror/shutter or something fixed on the K1 II shutter mechanism).