[pingflood]

A

However, since the testers accounted for vibration, the 50D sample does have an all round unexplained sharpness issue at low ISOs, that is not lens based since the same stop was used at all ISOs (except 100) and it gets sharper as it goes up.
.

How, exactly, was vibration accounted for? Different cameras tend to react differently when it comes to vibration, so what might've not been a problem on a D90 could be a problem on the 50D. Just sticking a camera on a tripod doesn't fix everything...

If the images come out looking sharper at ISO 1600 than 100, then there's a 90% chance the test wasn't done right, and a 10% chance the camera is broken in some very weird way.

[mgoblue]

They said a tripod and cable switch. That's pretty typical user behavior I'd think and suitable for the test. I don't know much about the 50D, but I'd guess it also offers mirror lock up which might solve the problem and I'm not sure they used it. But just because the 50D didn't perform flawlessly doesn't suggest to me the test wasn't done right -- none of the cameras performed flawlessly -- though for most purposes they can all do a terrific job. It does look like motion blur, so the 50D shutter may be jerking the camera more than the others. It could also be that an AC or something was only running for the 50D tests and that shook the floor a little.

[pingflood]

Not saying the 50D is flawless, just saying that something's not right when the same lens at the same aperture gives a sharp result at one ISO and not at another. So IMO this test really doesn't qualify as valid for comparison -- not fair to any of the cameras.

[kittykat46]

Its more than just vibration or lens aberration.
Take a close look at the ISO 100 shot. The Darth Vader figure is black, and the shadows would, in reality, be really deep black.
I can see the K-7 is reproducing the black shadow faithfully.
The D300 and to a lesser extent Canon 50D are producing a reddish-grey interpretation of what should be deep black. Either the test is flawed, or the D300 and 50D do not really deserve all the praise heaped on them......

[Big G]

I disagree.

Firstly, you can't see the background on the 50D on the ISO 100 crop shot (except for a very teeny tiny triangle in the corner), and secondly I think the K-7 shot lacking in detail in the black background (almost looks very slightly underexposed).

The K20D and K10D have the very similar black background detail as the D300.

[Billgscott]

Ater a lot of thought I decided to add to my K10 with the K20 instead of the K7 and spend on glass.

Seeing this review has made me vey glad I did, but a little sad that the K7 wasn't the leap forward all had hoped for.

[cooldude14es]

I dont know if I mentioned it before: the camera's owners were helping during the tests in order to minimize possible errors.

As you may imagine this review/comparison has generated controversy upon spanish Nikon and Pentax users (Canon users didnt say a word as far as I know). Some guys on the Spanish Nikon Forum discredited the guys who made the tests and the whole thing became a mess.

The testers admitted that there was a problem ONLY on the ISO 100 test, and thats why they posted images from Nikon and Canon with different speed and aperture, even so they are not completely satisfied with the final test.

I know personally the spanish pentax forum moderator and there is no need to say that he is a wonderful guy, nice and kind and it was never his intention to make a biased test, he did it honestly and as good as he could, always with the help of Nikon and Canon users, as I said before.

So, in order to prove their honesty and objectivity, there is going to be a second test in a few days to come. They learned from the first test so some conditions/rules will be changed and some other will be the same. Here is the complete list:

- Exposition will be determined by a exposimeter (is this word correct in english? ) and it will be applied to all cameras.

- A much better tripod will be used.

- Shots will be done in RAW without noise reduction.

- The same software will be used on all the photos developing, same parameters, camera's users will be attending the review, jpeg output, max quality, sRGB.

- Same lens again for all cameras: Sigma 105 f2.8 Macro.

- Shots at ISO 100, 200, 400, 800, 1600, 3200 and 6400.

- Front/back focus tests will be done in advance.

I will keep you informed. Thanks,

Alex.

[PolishMike]

Sounds awesome, Alex!!

Any time I see a camera/lens review people always complain about the methodology, no matter what. It's great to see some people (no less than an unbiased democratic body of different brand owners!) get together and actually respond to the complaints thoroughly.

(I'm willing to bet money on the fact that as soon as the second test is published someone will find fault with it )

[cooldude14es]

Yes, thats soooo true... there is always somebody complaining about it.

Lets see what happens this time!

Cheers.

[pingflood]

Alex,

In order to completely rule out vibrations, have you considered using strobes (flashes) to light the scene instead? You'd get the consistent light, and completely rule out problems with the tripod/mount setup. Also, you could temporarily put down a focusing target of some sort to 'fine tune' the focus manually before doing the shots. Since the cameras have live view it should be easy to get this identical.

[Big G]

I'm very much looking forward to this "rematch" !

[cooldude14es]

Alex,

In order to completely rule out vibrations, have you considered using strobes (flashes) to light the scene instead? You'd get the consistent light, and completely rule out problems with the tripod/mount setup. Also, you could temporarily put down a focusing target of some sort to 'fine tune' the focus manually before doing the shots. Since the cameras have live view it should be easy to get this identical.

It's not me who is doing the review, but I will let the testers know your advice.

Thanks.

[mgoblue]

Its more than just vibration or lens aberration.
Take a close look at the ISO 100 shot. The Darth Vader figure is black, and the shadows would, in reality, be really deep black.
I can see the K-7 is reproducing the black shadow faithfully.
The D300 and to a lesser extent Canon 50D are producing a reddish-grey interpretation of what should be deep black.

The cameras are all exposing differently with the very same settings -- that accounts for the shadow differences and probably most of the color differences and my bet is most of the noise differences too, but not the blur or inconsistent sharpness (why they expose so differently with the same settings is a good question!). The Canon is over exposing to my eye pretty badly...and the K7 maybe be underexposing a wee touch but for digital I quite like its exposure. Blown highlights are gone forever, but shadows can often be recovered with software. Importantly, though, neither may expose like this with its own meter creating the exposure setting in the real world.

It seems to me on board metering makes the most sense in the test because I'd say most photos are shot with on board metering these days... With a tweak up or down.

I'm looking forward to the new test. While I see the value in these tests, what they always seem to prove more than anything is the huge number of variables at play when it comes to 100% crops. And in the real world, at least in my experience, all the cameras perform so well...

[robgo2]

If the Pentax shot at the *same settings* produces a darker and therefore noisier image then it is very much a flaw of the Pentax - it means that a photographer trying to take a shot in low light and being limited in terms of shutter speed would get a poorer shot.

Not necessarily. A camera may have a tendency to underexpose as a matter of design, so as to avoid overexposure and clipping of highlights. It is up to the photographer to know how his camera's metering works and to make appropriate adjustments. If the K-7's images are indeed underexposed relative to the other cameras, then the comparison is less meaningful than it may seem.

Rob

[robgo2]

I looked at the comparison images last night, and after reading all the comments this morning decided I must have been sleepy and come to the wrong conclusion, so I checked them again this morning. Maybe it's my monitor or my eyes or my brain, but to me, the K10 images through ISO 1600 show more detail and less noise than those of the other cameras tested. Check out the small surface irregularity in Darth's right eye in the medial highlight and also the closeup of the watch rim. They appear to be sharper in the K10 images. Also look at the noise in the blue background to Darth's left. The K10 is showing less noise at ISO 1600 than the D300 shows at ISO 200. As stated previously, maybe it's my monitor, my eyes, or my brain, but it certainly isn't due to being sleepy, because the images look the same this morning as they did last night. I realize that with the K10D's 10 mp sensor the image is slightly smaller than with the other cameras and noise would therefore be less prominent, but what I'm seeing leads me to think that there has to be some fault with the procedure used in this comparison.

CN

Regardless of what you think you see in this study, the K10D is marginal at ISO 800 and pretty awful at ISO 1600. I have used one extensively over the past 2.5 years, and this is its greatest weakness. Vertical pattern noise is a constant problem and is almost impossible to deal with in PP. From what I have seen posted on the web, the K-7's high ISO performance is considerably better.

Despite what I have written, I love my K10D. It has given me some superb pictures.

Rob