[gary1952]

I spent most of the day trying to evaluate with pictures, which camera delivered the highest quality pictures. Discounting the cost entirely.
Take a look if you have time and give me your impression. Good or bad or indefferent.

IQ and noise level test K-7 vs. K-x Photo Gallery by Jerry at pbase.com

[redeleon]

Kx images are quite lovely than the K7

[Adam]

There's no doubt that high-ISO shots from the K-x are better than those from the K-7. However, the margin of difference isn't as big as people make it out to be, and I wouldn't say that your test was very much a controlled experiment.

If you want to do a controlled test, you also have to take into account the electronic noise-reduction that's by default set to 'low' in both cameras. Either disable it completely, or take a series of test shots at each setting.

An ISO 100 vs ISO 200 test would also be warranted.

[rawr]

Imaging Resource "Comparometer" ™ Digital Camera Image Comparison Page

Compare K-x <> K-7 sample images at all ISO's under contolled lighting etc.

I don't know about the K-7, but those K-x birds on the pbase page seem pretty noisy for 1600 ISO on the K-x. It ought to look a lot cleaner, viz:

[gary1952]

Adam, Good point. But in both cameras, I completly turned off the noise supression control. In my line of photography it is best to process as it comes out of the camera raw. If any post processing is to be made.
So both cameras were shot with sensors spitting out the original image. No processing.

Nice catch Rawr. Is that an oriental robin?

[Adam]

I'm still not convinced. See this shot; it's awfully noisy for being taken at ISO 800: http://farm5.static.flickr.com/4112/4844157708_9001a4f02a_b.jpg
(k-x)

[hangu]

Birds make poor test subjects, unless they've been stuffed.

When talking about noise, detail retention is just as important. Noise can be reduced dramatically in PP, at the cost of loss of detail. The K7 images exhibit more noise but also more detail according to this test. This means more detail will be retained at the same amount of noise reduction in LR compared to the K-x.

Before someone rips me a new one, keep in mind that I'm not saying the K-7 is better at noise suppression. I still think the K-x is better. I'm saying 1) this test is flawed and 2) we seem to be completely ignoring the fact that detail retention is part of noise reduction.

The K-7 has a higher MP count than the K-x, which allows it to have slightly more noise than the K-x at 100% magnification and still come out ahead.

[hangu]

In my line of photography it is best to process as it comes out of the camera raw.

Now this is a test I'd love to see. NR in PP Vs. NR in-camera.

I've recently started using NR in PP for a few reasons such as:

• Faster camera operations and faster playback on camera because the camera has to do less processing before displaying an image you just shot
  
• I can't imagine the on-camera software being anywhere as good as LR3 on my desktop.

These are just assumptions, however. I could be wrong.

[ll_coffee_lP]

Now this is a test I'd love to see. NR in PP Vs. NR in-camera.

I've recently started using NR in PP for a few reasons such as:

• Faster camera operations and faster playback on camera because the camera has to do less processing before displaying an image you just shot
  
• I can't imagine the on-camera software being anywhere as good as LR3 on my desktop.

These are just assumptions, however. I could be wrong.

I'm the same. After trialing Topaz Denoise for a month I decided to purchase it, set up actions for it in photoshop, and have completely removed any processing (that I can of course) in camera (I shoot RAW only). I now get the benefit of faster shooting while in the field, and clearer workflow measures (i.e. I don't have to think about whether I should use noise software or not...I just do it) when I return home and begin processing.

c[_]

[rawr]

...Nice catch Rawr. Is that an oriental robin?

Thanks. No, wrong hemisphere . That's a juvenile Butcherbird.

As to the K-x vs noise vs detail vs K-7 discussion, whatever NR goes into my K-x in PP are just the defaults of whatever LR or DXO decide to apply, even up to 12800 ISO. LR3 in particular is very impressive with the K-x, even at high ISO.

This is in marked contrast to how I have had to deal with my K200D high ISO's sometimes.

Also high ISO isn't just about noise and detail degradation. People always make the mistake of thinking that's all high ISO is about. But that's only one element of the high ISO IQ equation. It's also about holding onto a decent dynamic range as you slide up the ISO scale - as per the DXOMark measurements. Not much point in having highly detailed noise free images if they have no highlights or shadows, lack life and look as flat as a pancake. The K-x seems to do very well re both high ISO noise AND dynamic range, I've found, which has been very useful for birding in particular.

[CWyatt]

The exposure will significantly change the amount and type of noise.

So both cameras were shot with sensors spitting out the original image. No processing.

Aren't these JPEG?

[yeatzee]

All images seemed soft, and the noise was way more than im getting with my K-7.

[JohnBee]

I spent most of the day trying to evaluate with pictures, which camera delivered the highest quality pictures. Discounting the cost entirely.
Take a look if you have time and give me your impression. Good or bad or indefferent.

IQ and noise level test K-7 vs. K-x Photo Gallery by Jerry at pbase.com

First off, some of the K-7 images are sharpened(even to much for my tastes), while the Kx shots are not. :/
Secondly... the Kx is every bit as good as a K-7 with regards to definition, grain and DR at nominal ISO's and so I don't think the test shows anything to that regard.
However... I get the feeling that test images were shot just short of the Kx's most impressive attribute(why?)...

As for noise, it isn't so much the noise itself that stands-out, but the effects that it has on an image that seems to matter most.
I found the Kx's potential really seemed to come to life at/ or around ISO4000 and upward, while the K-7 seems to top-out at/or around ISO3200. But the difference in grain(noise patterns), DR and recovery(in the files) between the two is night and day. I just hope Pentax puts whatever they put in the Kx into their new camera this year. If not... I fear this may mark the time for Pentaxians to move-on.

For what it's worth, the Kx grain is a real pleasure to work with. Though not as impressive as that of a D700(for example), it does however remain quite manageable(with good PP) "almost" to it's maximum sensitivity.

[JohnBee]

those K-x birds on the pbase page seem pretty noisy for 1600 ISO on the K-x. It ought to look a lot cleaner, viz:

But its a good noise(take a look). The images contain alot of detail and the noise is none destructive. So that is a positive grain and a real pleasure to work with.

When it comes to noise analysis, it is the "mid and low frequency" stuff that we need to watch for and of course, the presence of noise artifacts.
But as for this noise... we might as well say it doesn't exist(click fix).

[rawr]

..For what it's worth, the Kx grain is a real pleasure to work with. Though not as impressive as that of a D700(for example), it does however remain quite manageable(with good PP) "almost" to it's maximum sensitivity.

Have you also noticed how LR3/ACR6 improves on the K-x grain even further, reducing the impression of noise not by luminance smoothing but by making the 'grains' smaller. It's quite impressive opening a high (on the K-x 6400 or 12800) ISO DNG in LR3 and seeing the normal big smudgy 'dots' of the preview JPEG turn into much much finer grains, which combined with only the slightest (default) luminance smoothing that LR3 seems to apply, makes for a very pleasing 'natural' result indeed.