DPR review of K-7

Tom C cakaltm at gmail.com
Sat Oct 3 14:16:04 EDT 2009

On Sat, Oct 3, 2009 at 12:58 PM, Adam Maas <adam at mawz.ca> wrote:
> On Sat, Oct 3, 2009 at 1:16 PM, Tom C <cakaltm at gmail.com> wrote:
>> I basically agree.  I thought because of the sensor, CMOS vs. CCD,
>> that the K20D/K10D difference was considered quite large.
>> Here's my chagrined viewpoint on the K-7, with my own rationale
>> applied, and I reserve the right to change my mind:
>> 1. First I have really been close to purchasing a Canon EOS 5D MK II
>> and still am, so there is an issue of  'do I want to spend a
>> significant amount of money on the K-7'?
> Personally, the 5DmII is neither fish nor fowl to me. If you need the
> high ISO performance, the D700 is better. If you need resolution the
> Sony's are better. Both the A900 and the D700 are faster (higher fps,
> better AF), the Sony's and the D700 are also better built and have
> better viewfinders. Oh, and they're cheaper (especially the A850). The
> only way I'd go for a 5DmII is if I needed a Canon-only lens.

On the other hand dpreview tends to think they've walked down the
middle line on this.

"The D700 set the benchmark for high ISO performance (along with
overall shooting performance at this price level), and the A900 set a
new benchmark for ultimate resolution. The 5D Mark II offers similar
resolution to the Alpha 900, increases the sensitivity range to ISO
25,600, and offers high ISO/noise performance that gets close to the
Nikon D700/D3. In short it - almost - offers the best of both worlds
without costing the earth".

>> 2. I've been very unhappy with the K20D, mainly in the area of
>> exposure accuracy. A very high number of images require adjustment and
>> it's unpredictable at times. I thought Pentax 'got it right ' before,
>> at least that's what all the cheering for their two prior top-end
>> cameras would lead one to believe. So I'm left wondering... have they
>> really got it right this time?
> Canon's not much better aside from the 1 series for exposure accuracy.
> After using pretty much every system on the market I've discovered
> that the biggest factor in how reliably a matrix meter performs is the
> number of cells. Under 10 provides decent performance as they're
> simple and easy to understand what they're doing. Between 10 and
> around 50 tends to get distracted too easily and is mostly useless.
> Over 50 works much better than anything else except spot metering.

Interesting observation.  I've considered just turning on
center-weighted and seeing if I could do a better job.

>> 3. The review on dpreview states the K-7 has "More RAW noise than
>> predecessor (but in line with competition)".  That sounds like a
>> potential 1/4 step backwards in image quality from the K20D. It also
>> reveals "In numerical terms the K-7 teams up with its predecessor in
>> being the noisiest camera in this group of four cameras".
> 4 channel sensors run hotter and therefore noisier than 2 channel
> sensors. The D300 and D90 have the same split as the K-7 and K20D, the
> slower camera is less noisy at high ISO's. They're not huge
> differences overall.
>> 4. The review on dpreview states the K-7 has "Less dynamic range than
>> direct competitors" and "At it default settings the K-7 produces
>> slightly less dynamic range than its predecessor".  Again that sounds
>> like a potential step backwards in image quality.
> That's JPEG only. DPReview's tests are far too biased towards JPEG
> use. The RAW headroom shows up to 10.5 stops, a bit less than the
> competition but more than enough for real-world use.

Good point.  I agree, I don't really care what their JPEG tests show
because it's only relevant if you plan on shooting JPEG's.

>> 5. The review states "Therefore, purely from an image quality point of
>> view, there is no pressing need for K20D users to upgrade".
> But in pretty much all other regards the K-7 is superior. Pentax does
> appear to be alternating processing and body upgrades (K10D was body,
> K20D sensor/processing, K-7 was body).
> --
> M. Adam Maas

If I hadn't purchased a K20D, I'd almost surely be getting the K-7.  I
skipped the K10D largely because of finances at the time.

More information about the PDML mailing list