[interested_observer]

I'm not really interested in the ISO bump to 819,200 from 204,800. I am interested in better image quality, with substantially lower noise. Here is an example at 12,800. I would be really interested in examples at 1600, especially for night shooting - no moon, ambient low light. The K1 is extremely good, but if it can be substantially better, then it might really be worth the $500 upgrade option. These are the first set of examples I've come across.

• Pentax K-1 Mark II Announced | ISO 819,200 & Hand-Held Pixel Shift | SLR Lounge

---

Last edited by interested_observer; 02-24-2018 at 10:53 AM.

[BigMackCam]

Impressive. I wonder if that's RAW or JPEG? If it's RAW, fantastic...

[jatrax]

Agree. 819,200 sounds great on a spec sheet. But what the camera does at 1600 is far more important to me.

[tuco]

Noise isn't everything. How much dynamic range of light are you capturing at ISO 12,800? 6 or so (usable) stops of light? You need to take subtract 2 to 3 stops from advertised DR because that is Engineering DR and not photographer's DR. That's terrible when it comes to tonal scale and hence image quality. And that's true with a lot of cameras.

---

Last edited by tuco; 02-24-2018 at 11:03 AM.

[interested_observer]

I would also be interested if it addresses the white dot issue in any way? In that the white dots appear to be ISO/noise related the accelerator chip may be their solution.

[BigMackCam]

I don't think you are interested in IQ. Because noise isn't everything.

Well, I'm definitely interested in all aspects of IQ, including high ISO performance (a lot of my shots are at high ISO settings). Luminance noise isn't everything, I'll agree... colour noise, however, can have a very detrimental effect even on smaller reproduction images. If that's better controlled in the K-1II's RAW files, and it's not at the expense of general image quality, then it's a good thing IMHO.

As for dynamic range, it would be swell if ISO 12,800 gave the same dynamic range as ISO 100, but that's clearly not going to be the case. Either way, there's sufficient dynamic range on my K-3 and K-3II at ISO 6400 to make it very usable. ISO 12,800 and higher are still useful. I'll be interested to see what the K-1II is capable of...

[SteveinSLC]

Yes, this is probably the big decider for me. The numbers above 25600 are silly marketing numbers and largely useless.
But if the quality takes a significant jump in the range from 1600-12800, that would be a big deal to me and make the upgrade well worth it.
That photo certainly looks promising, but I'll definitely wait to see a few more results.

[interested_observer]

I guess I should go and read about the KP (essentially the accelerator chip's capabilities) to see if the increased noise reduction is spread across the entire ISO range or is just oriented towards the higher ranges. I have not read anything about the KP, so I know very little - especially about any potential increase in image quality.

[tuco]

As for dynamic range, it would be swell if ISO 12,800 gave the same dynamic range as ISO 100 ...

Every time a new camera gets tested boasting high ISO the first thing I look at is DR vs ISO. And I continue to see little improvements there. I'm waiting for that tech breakthrough on that.

[photoptimist]

Unless the accelerator is disabled at low ISO, it should provide significant improvements at all ISOs in the shadows. Those samples show huge improvements in noise at 12800 which is to say that images taken at 1600 will show similar improvements in shadows only 3EV down from middle gray.

But to me the bigger "so what" for high ISO imaging is whether photographers who say they don't need ISO higher than 1600 say that because they've never had a camera that could produce a decent image at higher ISO. That is, they've unconsciously or deliberately avoided high ISO scenes either by putting away the camera when it gets dark or pulling out a flash or tripod.

There's absolutely nothing wrong with seeking or creating low ISO photographs. But there's a lot to be said for expanding the frontiers of high-quality, ambient light photography to times and places that aren't in full sun or that aren't readily (or permissibly) lit by flash or tripod-friendly.

Innovations like the K-1ii aren't so much about improving the images you already take (although it does that). It's really about expanding the conditions under which you can get great images.

---

If Canikony introduced a sensor and camera that gave one or two extra stops of low-light performance, it would be hailed as game-shanging innovation. Why no similar accolades for Pentax?

[BigMackCam]

But to me the bigger "so what" for high ISO imaging is whether photographers who say they don't need ISO higher than 1600 say that because they've never had a camera that could produce a decent image at higher ISO. That is, they've unconsciously or deliberately avoided high ISO scenes either by putting away the camera when it gets dark or pulling out a flash or tripod.

There's absolutely nothing wrong with seeking or creating low ISO photographs. But there's a lot to be said for expanding the frontiers of high-quality, ambient light photography to times and places that aren't in full sun or that aren't readily (or permissibly) lit by flash or tripod-friendly.

Exactly this!

Those of us who shoot at higher ISOs are used to dialing in wide apertures and slow shutter speeds to keep the ISO at manageable levels, but that often means compromising on depth of field, motion capture and the like. Better high ISO performance reduces the need for creative compromises

Like others, I'm not interested in the higher max ISO level, but rather the performance at ISO 3200 - 51,200. And I'd be especially grateful for a real quality boost at 6400 - 12,800.

I can't really say I like the change. Both aren't useable IMHO, but the K1 image still looks "organic" and recognizable, while the K1ii image looks like a cardboard cutout. The new algorithm seems to create hard sharp lines, then make a global decision about colours inside and outside those lines. It's not just "sharpening", it's chopping detail out. I can't see how this is useful.

[biz-engineer]

These are the first set of examples I've come across.

Both noise and MTF should be compared with and without NR. Not sure if perceived image quality is better with NR or without NR.

[swanlefitte]

To me getting photos in very low light is like trying to run up a wall. (The wall is an obstacle in the ninja warrior competition. ) the boost on my ks2 sometimes gets my fingers just high enough to make it. Its looking like with the mark2 i can get my elbows over the wall most of the time. Thats the difference between getting a shot and getting The shot.

[photoptimist]

I can't really say I like the change. Both aren't useable IMHO, but the K1 image still looks "organic" and recognizable, while the K1ii image looks like a cardboard cutout. The new algorithm seems to create hard sharp lines, then make a global decision about colours inside and outside those lines. It's not just "sharpening", it's chopping detail out. I can't see how this is useful.

Actually, the K-1ii is not chopping detail out. If the K-1ii image looks like a cardboard cutout it's because the original subject is a night shot of a cutout metal sculpture on the side of a building (see the samples at High-resolution design | PENTAX K-1 Mark II | RICOH IMAGING). The K-1ii is actually much closer to the true scene than the K-1.