OK, here are test shots from the K-1 II .
All with D FA 28-105mm at 28mm and F7.1 . I used TAv mode.
First one, 1/40, ISO 10,000 .
Second, 0.5s, ISO 500 :
Third, 1s, ISO 250 :
Last, 2s, ISO 160 :
I took the best of 5 shots for 0.5s and 1s .
For 2s, I needed at least 15 shots to get it this sharp. The difference is mostly seen when you pixel peep at 100%. At full screen on my 32" 4K monitor the different is difficult to perceive for the first 3 images. The first image has more noise, and the last one is clearly softer.
These are not from PixelShift. I have not tried disabling the SR.
If you click "view image" on each image, at least in Firefox, then click the magnifying glass, you can pixel peep the entire 36MP resolution.
---------- Post added 05-11-18 at 04:13 AM ----------
And here are the tests with the Panasonic GX85, which has dual IBIS, both from the body and the 12-32mm lens.
All shots were taken at 12mm (24 mm equivalent), and f11 . After shooting, I realized I shouldn't have stopped down the lens that much with this small sensor. The UI on the GX85 is still very unfamiliar to me, and it lacks a "Tav" mode, which the Pentax has.
I was having to switch to A mode to set the aperture with the single dial, then back to M mode to set the shutter speed with the single dial. Now that I think of it, the GX85 has a touch screen, and there was probably a way to set the aperture in M mode by touching the screen. But none that I could find by using the dial. Maybe I'll just have to RTFM. I got the GX85 camera only a few weeks ago, and used it mostly for video on tripod so far, or still photos of my cats during daytime. The stabilizer option had been left at the default all along.
First, ISO 25,600, 1/40s :
The field of view is different from the other 3 on this one because I went back upstairs to reshoot at f11 . My original shot at 1/40s was at f3.5 . I wanted all 4 shots to be at the same aperture.
Second, ISO 1600, 0.5s . This was the best of 5 images :
Third, ISO 800, 1s . This was the best of 5 images :
Fourth, ISO 400, 2s . This was the best of about 15 shots :
What's really surprising on this one is that none of the shots at the lower shutter speeds (0.5s, 1s or 2s) look very good at full screen on my 4K monitor, and they are just awful when pixel peeping.
The 1s image looks the best to me at full screen. I could probably have achieved better results at 0.5s with more tries.
The 1/40s shot is also awful, but not due to the stabilizer - wrong choice of aperture, which forced me to use the highest ISO of 25,600 on the camera.
If I was a review site, I would repeat somewhere between f9 or f7. I would also sober up. Anyway, to me, the stabilizer GX85 is just not helping much. I have not tried disabling it. Just night and day with the much bigger K-1 II at identical shutter speed in terms of blur.
I put it on a scale, and the K-1 II with the original battery, plus battery grip, 6 eneloop pros, 2 SD cards, and the D FA 28-105mm with reversed hood weighs 4.4 lbs.
The GX85 with the 12-32mm lens, battery, SD card, and the lens cap attached with a lens keeper weighs 1.1 lbs, a quarter of the weight.
Either I am much better at handling large cameras than small cameras, or the K-1 II really blows the GX85 out of the water in terms of image stabilization, despite being in-body only vs body + lens.
The first theory certainly makes sense, since I have been shooting Pentax for 10 years, usually with a relatively heavy superzoom.
The GX85 was also a $500 camera with two lenses (12-32 and 45-150), where K-1 II was $2400 with one lens (28-105) and battery grip. Huge difference in price.
Again, with these GX85 shots, you can click "view image" and magnifier in Firefox in order to pixel people to the full 16MP resolution.
All shots from both cameras were in RAW - DNG for the K-1 II, and RW2 for the GX85. They were processed to full-size JPGs in Aftershot. The K-1 II profile is one I created from the official K-1 profile. The GX85 profile is the official one from Corel. I didn't apply any kind of processing, all defaults when generating the JPEGs. As you can see, the 12-32 lens has a pretty bad distortion at 12mm. Aftershot doesn't have any lens profile for either the Pentax D FA 28-105 or Panasonic Lumix G Vario 12-32mm, so there was no lens correction applied either.
I don't have an Olympus to compare with, but I sure hope the stabilizer does better than my GX85.