Just a quick snap shot taken in very low light (nighttime, just a few street lights here and there). I didn't try to remove the noise or PP the shot, as this was just a snapshot to see how the K5 worked in basically no light. Not great, but considering the amount of light (see exposure info below) and the fact that the shot was a stop underexposed, I didn't really have high expectations.

K5+28-75/2.8 @ 28mm, f/2.8, 1/8s, ISO51200, handheld

[nanhi]

Dear Sir, I am surprised at the noise in this other wise lovely photo of your sweet baby - vow big attractive eyes.
We were told the Kr and the K5 where Pentaxes show piece where noise and high ISOs were concerned.
Here is a hand held shot of my house in the thick of night with street lights. K20D with the18-55 AL II Kit lens at 18 mm - photo 4. I have attached a Nikon D60 shot too - photo 3. Both reduced to around 260 KB for attaching here.

[dgaies]

Dear Sir, I am surprised at the noise in this other wise lovely photo of your sweet baby - vow big attractive eyes.
We were told the Kr and the K5 where Pentaxes show piece where noise and high ISOs were concerned.

Thanks

Regarding the noise; please keep in mind that this was shot at ISO 51200, f/2.8 with a 1/8s exposure.

That's almost 3 full stops below what the in-body light meter in the camera could even register. I'm not sure what settings were used on the images you posted, but I'm not sure how comparable they really are to this shot.

Here is another version, processed in RPP and Denoise. The light levels were kept intentionally low to try and recreate the actual lighting present at the time of the shot.

[Steelski]

Hello Nanhi. The photos you posted are all well and good, but they are really not difficult, nor demanding in terms of high ISO.
The D60 has the Sony 10MP CCD chip I think. You can see in the example you posted from it, that it smudges, there is color bleeding everywhere..
You have also not stated anything about ISO used, shutter speed or apature. The exif files are not intact.
I can tell you that there is between 2-3 stops difference between what I can muster with my K20D and the K-5 shots I have seen. The example above was shot at ISO 51200 which is 3 stops higher than what the K20D can use.
I just tried to get a similar result in similar light. What did I get.
At 6400 ISO 1/8th and F2.8 I got a similar exposed image to the original K-5 image, maybe slightly less exposed. So that would mean that I would have to drop light by another 2.5 stops (there abouts) to get something similar. I thought, I will drop it by 2 stops and see what I get.
What did I get, mush, the 6400 images from the K20D are unpushable. there is no dynamic range left in the blacks to get a result even on the same light level of a K-5 25600 image. I then thought.... It must be the RAW NR that is attacking the K20D files, so I did a full ISO series. I pushed ISO 100 to 25600 levels,,,,, all I got was very strong noise. Same with 200 and 400 and I almost got a usable result from 800, but fact is that 1600 was the best as it is the last real ISO. I got identical results from it and ISO 3200 and 6400.
I thought that If I exposed for ISO 25600 with ISO 1600 and pushed 2 stops that would work stops then I would get real K20D ISO 25600.. which is true. Except there is no real information in the shadows, I can not even get the same light levels if I push 8 stops. there is simply black in the shadows.
If I expose correctly at 6400 it looks good. But it looks like 51200 on the K-5, If I expose for the 51200 equivalent, I get nothing usable. Not even if exposed for 25600, or 12800. There is no room in the K20D file, there is a usable 6400 if exposed correctly.
The usability of the K20D dies 3 stops before that of the K-5, its as simple as that. No DR = no dice.
Your examples show that its possible to get results, but what you do not realize is that the K-5 can get the same/there abouts at light levels 8 times lower (3 stops)... and no, I will not be publishing the mushy K20D images, they are of a teddy bear in very poor light, the 6400 is well exposed, any underexposure (like the one needed for even higher ISO) is murder to the image.

[nickthetasmaniac]

I'm really impressed by this. Not so much because of the lack of noise (considering the situation) but rather the form that the noise takes... I know reviews mention 'grain-like noise' every now and then, but honestly I would've sworn this was film if I didn't know better!

No banding, no splotches (is there another word?)!!! Great!

[wshi]

The PP'd image is cleaned up beautifully. It's amazing how such an image would have been impossible to capture by any other Pentax camera.