|
All images © Bob Atkins
This website is hosted by:
|
Author
|
Topic: DxOMARK (Read 5171 times)
|
whizkid
|
|
DxOMARK
« on: November 09, 2010, 06:48:44 AM » |
|
Recent DxO figures for the Nikon D7000 and Pentax K-5 are a shock to this Canonite. I'm not sure what real world results will be but Canon needs to get on the stick. Canon's recent 60D was a step backward on several fronts while the competition is advancing important imaging areas and adding to performance. Anyone with thoughts on this among us Canonites? I know DxO disclaims their results for other than raw sensor output and that doesn't tell about other important aspects but things are buzzing in some forums (especially Pentax ). Of special interest to me was the dynamic range scores for landscape use. Based on their 14 point DR scores something should be noticed maybe helping reduce some burn-out.
|
|
« Last Edit: November 09, 2010, 06:51:12 AM by whizkid »
|
Logged
|
|
|
|
|
Bob Atkins
|
|
Re: DxOMARK
« Reply #1 on: November 09, 2010, 08:39:08 AM » |
|
I have not yet seen an actual review of the K5 so I really can't comment on the DR. The numbers certainly look good on bench tests. We'll have to wait and see how it performs in the field.
|
|
|
Logged
|
|
|
|
KeithB
|
|
Re: DxOMARK
« Reply #2 on: November 09, 2010, 10:46:32 AM » |
|
How are the Canon's deficient in taking actual pictures?
As far as measuring "sensor-only" goes, unless DxO is taking the sensor out and bonding it to their own mounts. there could be some RAW processing going on behind the scenes. I have heard that the k5 does some noise reduction in the RAW data.
|
|
|
Logged
|
|
|
|
Bob Atkins
|
|
Re: DxOMARK
« Reply #3 on: November 09, 2010, 12:12:40 PM » |
|
There's no doubt the K5 numbers are good, but I'll wait on a more detailed analysis to make comments on them.
You usually increase dynamic range by lowering noise, thus making the underexposed parts of an image recoverable by suitable processing. However looking at the S/N ratio of the K5 vs the 5D MkII, there's not a huge amount of difference, and in fact at lower ISO settings the 5D MkII noise appears to be lower.
The other way to increase DR is to use some pixels which are less sensitive to light, so you can expens the range at the upper end. That's the approach that Fuji took with their "Super CCD" technology. They used both small (low sensitivity) and normal sized pixels and combined the output to yeild a higher DR image.
It's not yet clear to me what approch the K5 (and D7000) takes to increasing DR. I think they both use Sony manufactured sensors, so maybe it's a Sony technology, though if so it's surprising that Sony aren't using it themselves unless there's some sort of patent issue.
|
|
|
Logged
|
|
|
|
KeithB
|
|
Re: DxOMARK
« Reply #4 on: November 09, 2010, 12:37:16 PM » |
|
I have heard that Pentax and Nikon have superior Processing (having more experience than Sony) and can wring more out of the sensor.
|
|
|
Logged
|
|
|
|
|
|
|