[ElJamoquio]

??? Where do you see 6db of difference? - at any ISO. The largest I could see was about 5, but most of the time along the scale the difference is about 3-4db - eg at 3200, 6400 ISO etc.

I'm sorry, I don't see 6 dB of difference, I was just giving everybody a reference.

Also one has to note the sometimes sizable discrepancies between measured ISO and claimed manufacturers ISO.

Of course, which is why no one on earth should buy a Fuji.

[ElJamoquio]

Have you done this comparison? I'm always a bit peeved, when I do something and someone tells me to do something different.

You don't have to do a gosh-durn thing, of course, but your picture has the Pentax zoomed out. That's negative help, because it gives an unfair advantage to the Pentax.

Shooting at anything other than your base ISO you are giving up dynamic range and increasing noise. For me, what you can do with the best image achievable is what is important. Which degraded image is better is less important to me.

Fair enough...

Looking at the charts at Imaging Resource, I'm guessing the K-5 IIs is tested at about 2600 lw/ph, while the D600 tests out at about 3000. That's about a 13% increase in resolution..the D600 should look better, I suspect that at values like 2600 lw/ph and 3000 lw/ph,

Those numbers sound way too sketchy for me. FYI imaging resource, IIRC, tests cameras with the kit lenses. Not too impressive a 'standard', and certainly shouldn't be taken as the limit of resolving power of a camera system.

[rawr]

FYI imaging resource, IIRC, tests cameras with the kit lenses.

No. IR mainly use a Sigma 70mm f/2.8 Macro, which is conveniently available in most mounts:

Eg from their D5000 review:

A note about focus for this shot: We shoot this image at f/4, using one of three very sharp reference lenses (70mm Sigma f/2.8 macro for most cameras, 60mm f/2.8 Nikkor macro for Nikon bodies without a drive motor, and Olympus Zuiko 50mm f/2.0 for Four Thirds and Micro Four Thirds bodies).

It is PhotoReview.com.au that use the kit lens that came with the camera. Or, if not the kit lens, whatever the best lens was the manufacturer sent them along with the review camera. Which at least shows what users may expect in normal use of the camera.

[ElJamoquio]

No. IR mainly use a Sigma 70mm f/2.8 Macro, which is conveniently available in most mounts:

Eg from their D5000 review:

Thanks; I believe I recall someone here referring to pictures or specs that WERE from the kit lens, e.g.

Nikon D600 Optics



Below are the results of our optical tests with the Nikon D600 and the bundled optional Nikkor AF-S 24-85mm f/3.5-4.5G VR kit lens. The test images shown on most other pages of this review were taken with very sharp references lenses, so we use this page to explore kit lens quality.

http://www.imaging-resource.com/PRODS/nikon-d600/nikon-d600GALLERY.HTM

Looking at the charts at Imaging Resource, I'm guessing the K-5 IIs is tested at about 2600 lw/ph, while the D600 tests out at about 3000. That's about a 13% increase in resolution..the D600 should look better, I suspect that at values like 2600 lw/ph and 3000 lw/ph, the ability of the human eye to discern such fine differences is being taxed. The D600 may have slightly more detail, but you have to look really hard to see it, and depending on whether or not there is detail the right size to be defined, may be irreverent to your pictures. .

...because the only 'chart' I saw for the D600 was on the kit lens page. I didn't look very hard, though. normhead, care to show the source for your conclusions?

[rawr]

I didn't look very hard, though. normhead, care to show the source for your conclusions?

Look under the 'Exposure' part of the IR D600 review for resolution numbers:
Nikon D600 Camera Exposure - Review

(They used the Sigma 70mm there).

[ElJamoquio]

Look under the 'Exposure' part of the IR D600 review for resolution numbers:
Nikon D600 Camera Exposure - Review

(They used the Sigma 70mm there).

Thanks. Do you know where the same figures are for the K-5 or the K-5 II? I looked; the Exposure part was missing for K-5II and didn't have resolution figures for the K-5 section.

Edit - found the 'similar' numbers for the K-30. The resolution was 2200 or so for the K-5II compared to 2800 or so for the D600.

[rawr]

Do you know where the same figures are for the K-5

IR read the K-5 numbers as 'about 2,100 lines from converted RAW files'. The K-5 resolution numbers are at:

Pentax K-5 Camera Imaging - Full Review

For a full set of recent Pentax numbers, their test of the K-x is also below. They said '~ 1,700 lines of strong detail' for the K-x.

Pentax K-x Camera Imaging - Full Review

[rawr]

Because I was bored, I prepared a little chart of some of the Imaging Resource resolution tests, to help compare various camera formats. Not much in these numbers reveals any huge natural advantage of FF over APS-C in terms of resolution, although the Canon 6D does seem to do very well with it's modest 20MP.

The table shows the resolution of the camera alongside it's megapickles.
The chart plots just the resolution.

[Ash]

Interestingly the 12bit K30 scores better than the 14bit K-5. Not sure but there may be some experimental variation causing a measurement error to give such results. Confidence ranges and standard deviations would be interesting to see, unless this was all based on an n=1 study comparison.

[Kunzite]

Large error margin? Measurement error? "Lighter" AA filter? Better demoisaicing and sharpening algorithms?

[rawr]

unless this was all based on an n=1 study comparison.

Just about. I think they get one camera to test, they mount the same Sigma 70mm on it, and then shoot their test charts. Someone like LensRentals could probably do testing on a larger number of bodies, but IR don't have the resources to do that, I think.

Anyway there are other testers out there to cross-check with. The IR numbers seem generally sensible.

[normhead]

You don't have to do a gosh-durn thing, of course, but your picture has the Pentax zoomed out. That's negative help, because it gives an unfair advantage to the Pentax.

I've actually shown that if you blow the Pentax image up it doesn't really change much. I've done it for myself and posted the images in other threads. You might want to look them up, or verify the assertion yourself. But I can tell you having done it, enlargement is not going to be the negative factor you think it is. Digital is not film. The degradation of image enlargement caused in film is tempered by advanced enlarging algorithms in digital. My best advice here is do the work yourself. It'll be done your way and you'll be comfortable with the results. I can guarantee you from my work, if you spend a few hours trying to understand and issue and post your results on the forum, you will be dismissed by people who's work towards understanding the issue is composed of reading the work of "experts" and expounding on their take of what the experts said. Even though in many cases what the experts said has been mis-represented to conform to their own unsupported point of view. The thing is, once you've down the work, you know what you're talking about. Before then, you don't. it's as simple as that.

Interestingly the 12bit K30 scores better than the 14bit K-5. Not sure but there may be some experimental variation causing a measurement error to give such results. Confidence ranges and standard deviations would be interesting to see, unless this was all based on an n=1 study comparison.

I have read a paper done by a guy from the University of Chicago suggesting there was no benefit in going to 14 bit in APS-c cameras. It was so far over my head, I'm not going to try and explain it...I just want to point out, some really smart people think that way.

[badsykes]

You guys make me feel like a lab rat...
If you want something "cool" for peeping pixels get the 645D and you have 3050 resolution...

[ElJamoquio]

Anyway there are other testers out there to cross-check with. The IR numbers seem generally sensible.

Absolutely, there are other test setups to check. And thanks for compiling that info, I might make some charts based on that.

I don't mind the cameras being only tested once. It seems to me like the sensor should NEVER be a problem. The AA filter also seems like it'd be pretty well controlled.

I'd be more concerned about using different lenses to test different cameras. One reference lens should be used IMO - something like a 'normal' hasselblad or whatever has a really long registration distance.

I appreciate the IR numbers, but when a APS-C (D3200) and FF lens (D3X) have the same resolution capability, there's been a bad test. MAYBE if there's no AA filter on the APS-C... but the D3x scores only 80% of the D600? No way.

[normhead]

You guys make me feel like a lab rat...
If you want something "cool" for peeping pixels get the 645D and you have 3050 resolution...

The stock reply to that is "The difference in price between a 645 and a D800 is such that it's not economical." That is said by everyone who doesn't own a 645D. The fact that FF users are arguing absolutes in order to justify the D800 (better IQ whether they'd actually use it or not) but then change their argument to argue price and refuse to argue absolutes in terms of the 645D, and the contradiction that implies, always seems to escape them. My question is "do you need maximum resolution or don't you?" Once you start talking price and value then the K-5 comes back into the equation. It's a discussion that goes back and forth between absolute IQ and value, that in the end comes down to, if you can afford a D800 and the high end lenses that go with it, go for it. If you can afford the 645 DSLR go for it. If you can't afford either don't sweat it, you'll probably never notice you don't own them, except in "my sensor is bigger than your sensor " bragging contests.

I appreciate the IR numbers, but when a APS-C (D3200) and FF lens (D3X) have the same resolution capability, there's been a bad test. MAYBE if there's no AA filter on the APS-C... but the D3x scores only 80% of the D600? No way.

Is this because you've compared images or is this a theoretical construct? You also might want to consider things that don't show up in tests. I was looking at a used 5D a while ago, and the common thing users raved about was the "colours". In reviews and comparisons I've never seen this addressed in any way. It's quite possible that Canon has something going for it that no one has found a way to measure, that makes them look worse in the composite scores than they really are. There is simply no other way to rationalize how Canon users feel about their cameras. All those pros shooting Canon, they aren't all sheep. Some of them know what they're talking about, and they can't all be wrong. There is a reason why people love Canon cameras that doesn't show up on the test charts.