[rawr]

Thanks Haaken. Very useful.

Use of the target for Ektacolor papers also reminds me that the perception of noise can vary dramatically by output medium - screen, print etc. In comparable modest 8x10 or less size prints from the K20D and the D90 it would probably be hard to pick them apart under even the highest ISO's. This is something that falls out of the DXOMark scope.

[ivoire]

I found this article to be interesting and thought i'd post the link.

Digital Photo - D-SLR State Of The Art, Part II | DPmag.com

On the 3rd page, there is this (and I quote) under the topic

"Balancing Noise and Sharpness"......

"Whatever methods you use for controlling noise, we still prefer a sharp shot with noise over a blurry shot without it. If you need to increase your ISO to get an adequate shutter speed, don’t let concern about noise stop you."

AND, they printed that line twice. I've looked at a number of high iso pics from several 'pro' cameras compared to the k20d and k7. While there may be more noise in the pentax shot, the detail is better than many of the competitors. So, I'm done with my noise about noise and intend to get out there and just enjoy my Pentax camera.

[Haakan]

I am happy my efforts are appreciated (my wife were somewhat less pleased with me sitting by the computer)

I agree that these cameras are all producing high ISO images that are very hard to distinguish from each other, and that other factors (e.g. print media, screen, and correct exposure) probably will have a bigger impact on the perceived noise than which of these cameras were actually used.

Best regards,
Haakan

[Class A]

And assuming that most of the APS-C cameras are using roughly similar sensors, it then comes down to what you do with the signal.

There are still some differences in sensor technologies.
It appears that there is still some room for improvement for the Samsung sensor (based on a pre-production model) used in the K20D/K-7, with respect to dark noise.

If there is a trade-off between noise and detail, I'll take detail and make noise a choice that I make in pp.

Yes, any camera not doing that is patronising the user (not good).

...I thought I would try to make my own comparison of the high ISO image quality using RAW files from Imaging Resource.

Great contribution, Haakan!

I set my browser window size such that a "page up" and "page down" button press would exactly replace one version of an image with another.

I personally think the (chroma) noise handling (looking at ISO 1600 and 3200) is done best in this order: K20D (best), D90, 50D (worst). I did not expect this.

[Haakan]

Based on the often very heated discussions about K20D and K7 noise performance, and the discussions around DxOMark, I was also very surprised that even at pixel peeping level the differences were so small.

I personally also felt that the overall noise on K20D was slightly less than for the others, especially looking at the lighter part of the image. But one can also see that for the darkest part, there is slightly less noise in the D90.

But again, looking at the overall impression of the images the differences were indeed very small, and I would think that they all would actaully produce very comparable images.

Best regards,
Haakan

[eagle]

Instead of reading all this, I'm going to go take some pictures.
I'm getting results at ISO 4000 that I'm happy with. Next generation sensors will see that increase I imagine, but not everyone will be happy, again....

[eva2000]

Based on the often very heated discussions about K20D and K7 noise performance, and the discussions around DxOMark, I was also very surprised that even at pixel peeping level the differences were so small.

I personally also felt that the overall noise on K20D was slightly less than for the others, especially looking at the lighter part of the image. But one can also see that for the darkest part, there is slightly less noise in the D90.

But again, looking at the overall impression of the images the differences were indeed very small, and I would think that they all would actaully produce very comparable images.

Best regards,
Haakan

well it has been found K-7 out of the box applies too much sharpening hence noise - lower the sharpness a few levels and high iso noise is much better from my own K-7 tests as well from https://www.pentaxforums.com/forums/691001-post29.html

[Haakan]

well it has been found K-7 out of the box applies too much sharpening hence noise - lower the sharpness a few levels and high iso noise is much better from my own K-7 tests as well from https://www.pentaxforums.com/forums/691001-post29.html

I fully agree, most of the times the comparisons are done with default JPEG settings, and those can differ quite a lot between different brands, based on what their philosphy is (optimizing for printing vs screen etc).

This is why I did the comparisons based on RAW images, with a third party RAW converter that is said very neutral, and also used no NR or sharpening.

So when done this way, taking away the impact on JPEG default settings, it turns out that they all do perform almost identically.

So at least for me, the noise at higher ISO, will not be a decision factor when selecting next camera (I like my Pentax for a number of other reasons)

Best regards,
Haakan

[Class A]

well it has been found K-7 out of the box applies too much sharpening

Note that the problematic sharpening levels were not "out of the box" but maxed out by the user.

With a view to the forthcoming DPR review I wished the out-of-the-box settings were more DPR friendly but we'll see...

[eva2000]

Note that the problematic sharpening levels were not "out of the box" but maxed out by the user.

With a view to the forthcoming DPR review I wished the out-of-the-box settings were more DPR friendly but we'll see...

my K-7 default out of box set to bright profile with FINE Sharpness set