comparing images from Pentax K-3, K-5 IIs, and Nikon D7100

<p>Using Imaging Resource compare images, results seem to agree with similar comparisons by dpreview between the K-5 and D7100 at ISO above 800. The K-5 IIs appears to retain a bit more edge sharpness at higher ISO settings compared with the K-3 (which is still very good) and even moreso compared with the D7100. Judging by comparisons at lower ISO and the house poster image, it seems to me the D7100 employs more conservative in-camera sharpening than the others, even though that of the K-5 is already somewhat conservative. All images are JPEGS at default camera settings. I was amazed upon finding the K-5 doing so well with noise control using a large 16mp sensor, and now incredibly I find the K-3 reduces noise just as well with its 24mp sensor, but evidently by using more agressive NR. </p>

<p>Below ISO 800, however, image sharpness and detail from the K-3 appears to be outstanding, while the K-5IIs is no slouch. I did not notice any moire in images from either camera, even in the fabric patterns that are presented. Doesn't mean it can't happen , but indications are for infrequent occurance.<br>

</p>

<p>There has been quite a bit of discussion on the "other forum" about the K3 noise above ISO 1600 and, at the same time, showing in several images how this noise has a sort of grainy nature which is rather easy to "deal with".<br>

On the other hand, being a K5 user myself and having tried the K3, I find this a bit annoying given that it becomes a matter of having to PP more, if you want to have a "noise-free" image.<br>

Another "plus" for the K3 is its cropping power, so to speak. Using a DA*300/4 and cropping up to 50% still gives one heck of a nice picture.</p>

<p>I thought I'd jump in today .... I haven't posted for quite a while and logged in to see what's up regarding Pentax on this forum.</p>

<p>Cheers!</p>

<p>JP</p>

<p> </p>

<blockquote>

<p>On the other hand, being a K5 user myself and having tried the K3, I find this a bit annoying given that it becomes a matter of having to PP more, if you want to have a "noise-free" image.</p>

</blockquote>

<p>I think it is almost fascinating how our standards and expectations of digital technology have risen so quickly. Complaints about noise at ISO 1600 from a 24MP cropped sensor camera would be unthinkable five years ago.<br /><br />Sure the post processing demands more effort, but you are getting so much more and I'm not just talking about the number of pixels per image. Due to its improved AF capabilities and faster frame rate, the K-3 can be viewed as a performance camera rather than solely an art camera. <br /><br />My Canon 7D is very much in the former camp: the dynamic range may be less than the K-5 series, but its superior AF and FPS allows me to get shots that I simply couldn't with the K-5. And it requires a fair amount of post processing to obtain final images to my standards.<br /><br />So here is the K-3, merging both of those worlds to a fair extent. While the AF may lag some behind the 7-D, it may just be good enough. Plus the superior sensor has allowed me to obtain usable images at ISO 12,800, under the lights even with an F4 zoom supertelephoto. My 7D was turned off above ISO 3200. <br /><br />So the value of the K-3 become more multidimensional if you have firm requirements to test it against. Some folks who don't have requirements for "performance" (in my case sports and birds) may have a hard time recognizing the whole picture.</p>

<p>ME</p>

<blockquote>

<p>All images are JPEGS at default camera settings.</p>

</blockquote>

<p>So, the real comparison is where the manufacterer has focussed on in the in-camera processing, rather than the actual ability of the camera? Sharpening opposes noise reduction, noise reduction opposes details.... one can't have it all. A far more interesting comparison would be RAW files developed as identical as possible in a leading-class RAW development tool (Lightroom, CaptureOne, DxO for example).</p>

<p>As for the "effort" to get noise free images: all of the above mentioned tools work with profiles that automatically adapt noise reduction / sharpening depening on camera model and ISO. It really isn't more work, it's just different defaults.</p>

<p><em>I think it is almost fascinating how our standards and expectations of digital technology have risen so quickly. Complaints about noise at ISO 1600 from a 24MP cropped sensor camera would be unthinkable five years ago.</em><br>

I am not complaining, just stating the fact that to me, at least that is how I see it, any PP is an "effort" ! :)<br>

Of course, the K3 is one heck of a camera and ... I have ordered it and I am really looking forward to take it out for birding with the DA*300/4.<br>

JP</p>

<p> </p>

<p>I remember when the K20D came out and we finally got to the "close enough" level of ISO800 Fuji film ;-)<br>

Even a m4/3 GX1 has a usable ISO1600 nowadays and it's a generation behind what m4/3 is capable of today...</p>

<p>Usable images at ISO 12,800? Wow! That is amazing. </p>

<p>Yes, the K20D was considered decent for noise control. I found the K200D to be even a little better (though limited to ISO 1600), and both had in camera user control for NR for the first time. Both also retained very good detail at higher ISO. But the K-x, K-r, K-5's and now the K-3 go well beyond what we thought was good.</p>

<p>I mentioned the tests were done at camera default settings for JPEG images and thus do not apply to RAW shooting and PP to advise RAW shooters of this. These are valuable comparisons for people who shoot a lot of JPEGs, and those like myself and Jacques who do rely on in camera settings most of the time, and are inclined to avoid PP as much as possible. The dpreview test will of course be of further interest. They run tests for both JPEG output, and for RAW as processed using a standard for all comparisons. </p>

 

<blockquote>

<p>Usable images at ISO 12,800? Wow! That is amazing.</p>

</blockquote>

<p>Dunno - with a little careful PP work, I can get very usable 12,800 ISO images from my Canon 7D: and using DxO Optics Pro 9 for the Raw conversion stage, it's actually pretty easy.</p>

<p><a href="http://www.capture-the-moment.co.uk/tp/tfu29/upload/DxO_12800_crop.jpg">This</a> is what 12,800 ISO from the 7D looks like at 100% with Optics Pro's new "PRIME" NR. The "standard" NR isn't far behind, either. </p>

<p>And this is without further NR from dedicated NR software, which - let's face it - wouldn't be unreasonably applied to a 12,800 ISO crop sensor image.</p>

<p>Because it's rather <em>related</em>, I don't wanna bother to open yet another thread... Surprise, surprise; DXO just published their <em>analysis</em> of the K-3, a <em>sensor specific</em> test. Rated an overall <strong>80</strong> as opposed to the long standing <strong>82</strong> for the K-5. And here I was, waiting for first reliable test reviews before ordering the <em>latest</em> (not <em>greatest</em>..?).<br /> Weird science...</p>

<p>While machine tests provide their dimension of truth, and offer valuable information to include, I have never been inclined to regard such results as the last word indicating what to expect in actual use. Pop Photo uses the latest DxO machine to evaluate noise performance in their camera tests, but a good rating from their test does not always coincide with image quality as actually seen. For instance, their result does not seem to reflect how much detail is visibly preserved as noise is reduced, even though the lines of resolution they report indicates a good performance there as well.</p>

<p>It is interesting, however, that the K-5 is still rated higher than the K-3 by DxO. Gives one pause to exercise caution, or realistically to expect the K-3 to provide superior performance just at ISO 800 or less. Still a great camera with all its numerous advancements. </p>

<p>I also wonder about K-3 file sizes compared to the K-5, both for JPEGs and RAW. Those from the K-5 are already plenty large and cause my olde computer to run somewhat slower for any PP I may wish to employ, compared to images from say my K200D.</p>

 

<p>As Michael Kuhne noted, the DXO ratings seem to represent a sum and overall results can be skewed by performance gains or losses in a given category. If you want to see a good example of this, compare the Pentax 645d vs the K5 series of bodies or the K3, for that matter. The noise figure of the 645d is considerably superior to the K5/K3. Yet the K5 and to a slightly lesser extent the K3, all exhibit better dynamic range. Even so, DXO rates it a tie at 82 for both the K5 and 645d. But when you compare the image files from the two cameras (I own both) the 645 is far superior in resolution and tonality. </p>

<p>I'm a Nikon D7100/Sony A77/A57 user and just read the Imaging resource review on the K3. Wow. 24 MP, 8.3 FPS, 60 image JPEG buffer / 23 image RAW buffer, and large prism. I'm jealous. I shoot a lot of sports and surfing and while the A57 and A77 are excellent with the Sony 70-400, I need to upgrade my D300 Nikon body to a D400 (which doesn't exist). I have a lot Nikon mount glass. The D7100 is an excellent camera, it's just to slow with too small of a buffer. Is Nikon really following behind technologically or are they purposely crippling their DSLR bodies as GM used to do with their automobiles. </p>