[JohnBee]

This is a new thread in response to the earlier one but with a completely different setup(from DPReview). I decided to make a new thread because I felt the old one had run its course and that adding this to it would most likely not receive the exposure it deserves.

In case some of you don't know.... DPReview initially conducted a head to head noise test involving a 60D, D7000 and K-5. However, the K-5 output was so poor in fact(PF and blur) that it stirred-up a series of complaint threads and inquiries which motivated them(DPReview) to reshoot the test with a better lens. Which I for one am glad they did, because it completely reverse what was initially thought of the K-5 against the competition. AND... it also served to put to rest the absolute disaster by Imaging Resource's Cmparometer samples. That are so far out of spec, that the blacks in the scene aren't even black anymore but rather a shade of purple(extreme PF).

Well that about does it for my rant and so lets move-on to something a little more positive shall we:

FULL SIZE SAMPLES:

K-5, 1/13s, f/9, ISO100

Click on the image to view it in a larger size

D7K, 1/20s, f/9, ISO100

Click on the image to view it in a larger size


ORIGINAL RAW/SAMPLES:
http://movies.*************.s3.amazonaws.com/pentax_k5/IMGP2482.PEF.zip
http://movies.*************.s3.amazonaws.com/nikon_d7000/DSC_1875.NEF.zip


CROPS:

K-5 CROP: 1 (100%)

D7000 CROP: 1 (100%)


K-5 CROP: 2 (100%)

D7000 CROP: 2 (100%)


K-5 CROP: 3 (200%)

D7000 CROP: 3 (200%)


-

Well there you have it! The K-5 and D7000 head to head in what we could call, fair and balanced conditions. Needless to say the differences really do come down to hair splitting between both systems. However, the lack of PEF converters did not help matters any in this case.

Either way, I'm just glad we had more than just Imaging Resource from draw conclusions from and of course that DPReview had the backbone to re-shoot their own test samples.

---

Last edited by JohnBee; 05-01-2011 at 05:16 AM.

[NaClH2O]

well, I'm looking that these samples from my less than optimal work monitor, but I have to say "a mouse would starve on the difference" I even asked my boss and his response was "there is supposed to be a difference?"
Outside of slight differences in color rendering, I cannot see any difference on a standard 17" business flat screen monitor.

NaCl(but I'm not a good pixel peeper)H2O

[Tommot1965]

be nice to see a K7 in that test result....just so us K7 owners can see if the K5 upgrade is worth it...or sit back and wait for the next body releases?

oh and to my eyes the K5 trumps the D7000 in these images...

[telfish]

If anything the K5 is a little sharper and resolves better than the 7000 in this series.

It's as I said before it all comes down to the lens.

Is it a good lens to start with and is it a lens designed for digital and for crop factor bodies
Is it a good sample. Is it calibrated to the body.
Have they taken care with focusing.

If the answer is yes then you get a decent result.

[JohnBee]

well, I'm looking that these samples from my less than optimal work monitor, but I have to say "a mouse would starve on the difference" I even asked my boss and his response was "there is supposed to be a difference?"
Outside of slight differences in color rendering, I cannot see any difference on a standard 17" business led monitor.

NaCl(but I'm not a good pixel peeper)H2O

Well I'd agree. Especially at 100%.

However at 200%, there is an ever so slight difference between the Pentax system over the Nikon one(stray pixels as I'd call them). But it's not apparent at first glance and I doubt any of that would ever translate to print unless it was either severely cropped or enlarged.

Having said that... the difference may also come down to the RAW converter also. I had to process the Pentax PEF with the Digital Camera Utility 4, which may or may not render things differently than ACR 6.x. So there is still room for variance in output.

Either way... these samples bare no resemblance of traits to the Imaging Resource samples I quoted earlier. In fact... I'd go as far as saying that the K-5 images shot by IR may very well be the poorest I've seen this side of the review table! And I mean that in the nicest of ways

[Jeff Charles]

...
Well there you have it! The K-5 and D7000 head to head in what we could call, fair and balanced conditions. Needless to say the differences really do come down to hair splitting between both systems. However, the lack of PEF converters did not help matters any in this case...

They do look really close. Which software did you use to render the raws?

Edit: You answered the question as I was posting. i have found that LR 3.3 renders the K-5 image a little softer than the D7000's until I apply a fair amount of sharpening. Is it possible that the Pentax DCU is applying sharpening based on the camera setting? I am not saying that this is problem. If the K-5 has a stronger AA filter, more sharpening is appropriate.

Jeff

[JohnBee]

They do look really close. Which software did you use to render the raws?

Jeff

I used Pentax Digital Camera Util to convert the PEF to DNG.
Then I converted both in ACR v6.3.064 and sharpened both images using the exact same profile and settings.

[Jeff Charles]

I used Pentax Digital Camera Util to convert the PEF to DNG.
Then I converted both in ACR v6.3.064 and sharpened both images using the exact same profile and settings.

Thanks. We are leapfrogging posts. You did exactly what I did, except that I used LR.

[Abstract]

Crop 1, the D7000 looks sharper, crop 2 the K-5 looks sharper, crop 3 can be argued.

Hm...

[Gimbal]

Crop 1 and 2 is ok, but crop 3 what is going on there with the K5? Is it saved with 256 colors or something?

Very ruff gradients in the blue, it's like there are only two shades, sea blue and light blue.
One can also see that the raw converter used for the K5 is not as good on diagonal lines as the one used by nikon, the black lines in the border becomes very blocky when running diagonally.

But then didn't you use the same converter? Strange.

Edit, or is the K5 over-sharpened, a lot?


(K5 to the right)

[JohnBee]

Crop 1 and 2 is ok, but crop 3 what is going on there with the K5? Is it saved with 256 colors or something?

Very ruff gradients in the blue, it's like there are only two shades, sea blue and light blue.
One can also see that the raw converter used for the K5 is not as good on diagonal lines as the one used by nikon, the black lines in the border becomes very blocky when running diagonally.

But then didn't you use the same converter? Strange.

Edit, or is the K5 over-sharpened, a lot?

I think there may be some JPG degradation there(I didn't check my settings) but they may have been too conservative coming from the FULL size samples.

[rawr]

FULL SIZE SAMPLES:

Why does the EXIF from each of those full size samples say they were BOTH taken on a Nikon D7000? Something has gotten mixed up here.

Also I wish you would just link to those full size images, or provide a tumbnail and then a link, rather than include the big images directly into the post here (at 4+ MB each per image).

I am being shaped atm and this 10MB page has taken me at least 20 minutes to load....

[stormcloud]

why does the k5 look darker than the d7000 dispite a longer exposure? just difference in raw processing?

[robgo2]

They are, indeed, very close, and any differences might be explained by variations in the lenses and/or processing. If anything, the K-5 samples seem to show a tad more detail, but not enough to matter in the real world. I am eager to see the high ISO samples, as we know that Pentax treats on-sensor noise differently than Nikon.

Rob

[Jeff Charles]

why does the k5 look darker than the d7000 dispite a longer exposure? just difference in raw processing?

I have the same two raw files loaded in Lightroom. I spot checked different patches in each, and they are actually fairly close in brightness, except for the darkest patches. Those are brighter in the K-5 image, making it look less contrasty than the D7000 image. The difference may be due to the values given by each manufacturer to specific "tags" in their raw files.

Jeff