[Urkeldaedalus]

I really expected the K7 to looke a worse at 3200. I was a bit surprised. These were all raw correct?

ISO3200 is about my absolute cut off for the K-7, and that's only if I get a good exposure (pretty important at this ISO). The nice thing the K-7 had going for it was that the grain looks pretty good, especially when printing instead of viewing it on a computer.

[MAC]

The only other RAW converter I have is PSP Pro X3, and it's so slow that I don't think I'd have the patience to do all of that work with it. I have Lightroom 3 on the way, perhaps when it gets here I can redo this comparison, although at that point there are likely to be more (and better) comparisons up.

[maxpiz]

Very interesting!
But I think it's also very important result of this new sensor at low ISO. So it would be useful to see the crop of the shots even at 80, 100, 200 and 400 ISO.
Thanks anyway!

[johnmflores]

I'm compiling a collection of K-5 photos that illustrate the K-5's low-light/high ISO capabilities:

Flickr: Discussing Low-light/high ISO Samples in Pentax K-5 dSLR Fan Club

I encourage all K-5 owners to join the group and contribute their own examples!

[rparmar]

What is with the crazy blue fringing on the K-5?

K-7 looks sharper which does not surprise me, as the K-5 absolutely must be doing more NR in-camera to achieve these results at high ISO. K-5 seems to have a more "plastic" look, just as the K-x does.

But if you absolutely must have the high ISO then the K-5 it is!

Thank you for these shots. Can we now have a nice set of the K-5 and K-7 taken with the Mamiya?

[MAC]

As it turns out, the Pentax software applies fairly heavy NR. I've reprocessed these with PSP Pro X3. My apologies for the misleading results in the original post. These are the ones I've done so far, I'm working on the rest but this method is slower and it'll be a little while.

EDIT: The artifacts at the edges of the K-5 images are a result of PSP not playing nice with the K-5 files. They appeared when the files were opened in PSP, and I thought I had cropped them all out. Apparently i missed some.

ISO 100










ISO 6400

[Painter]

With the new images it looks like the 100 ISO result is cleaner from the K-5. How that would show up in a print I'm not sure but you might be able to print larger without any further processing. Thanks for the effort.

[audiobomber]

With the new images it looks like the 100 ISO result is cleaner from the K-5.

I don't see that at all. The letters Natural 1600 for example are blurry in the K-5 ISO 100 sample. It could be a focus issue, or it could be some innate processing of the raw file. At f8 I'm more inclined to suspect the latter. I know my K-x is a little blurry at low ISO's compared to my K20D.

[raz]

No,is not all about sharpness, I think the images from the K7 have more contrast. Look at the side of the camera, in the K7 image is almost black.

The k5 image looks more exposed - look at the specular on the plastic lens - the white is clipped - and also the side of the camera is more "white", like more light was captured.

Very different results from these 2 sensors if all the shutter/aperture/iso were the same.

[MAC]

Alright, I've finally finished redoing all of the pictures, and i now have crops for the lower ISOs too.

ISO 80







ISO 100










ISO 200










ISO 400










ISO 800










ISO 1600










ISO 3200










ISO 6400

[Edvinas]

Well, it seems that K-5 has only one f-stop advantage in noise...

[peteng]

Some lowlight, high ISO would be great too...

[MAC]

ISO 12800







ISO 25600







ISO 51200

[raz]

Well, it seems that K-5 has only one f-stop advantage in noise...

I would say almost 2 stops. But the idea is that 3200 and 6400 are absolutely usable.

And these are without NR, right ? Image these samples out of K5 with a little NR.

[MAC]

Yes. I did everything i could the second time around to avoid NR