View Full Version : E5 DxOMark ratings

6th December 2010, 07:01 PM
Interesting results over at DxOMark on the new E5.


The above compares the E5 to the E3 and Nikon D7000 although you can easily select other cameras to compare to.

6th December 2010, 07:10 PM
Umm that is interesting..... but I would say that numbers alone surely do not add up to image quality.

Edit: Weather sealing aside (which is a big bonus for me) and even though it is somewhat heavier than my E-30 I will always pick up my E-5 first even though according to Dxo they are apparently pretty much the same.

6th December 2010, 07:14 PM
If I am reading it right sensor quality on the E5 is down slghtly compared to E3.

As such image quality must be due to jpeg processing improvements.

6th December 2010, 07:23 PM
I would say not just Jpeg because I definitely see improvements in RAW compared to the E-30. Processing in CaptureOne 6 (which is a nice improvement on 5 btw :)) I can see RAW files are cleaner, holding more detail and sharpness. So I would conclude that Oly as worked some magic with their imaging pipeline which influences both Jpeg and RAW.
Again I would say don't get too hung up on numbers, it's the image quality that is important after all :)

6th December 2010, 07:41 PM
The nice thing about DxOMarks is that it is independant so I was waiting for their results to be published.

Other positives e.g. more megapixels and bigger screen etc on the E5 but sensor wise no improvement. I was hoping for the dynamic range to match that of Nikon, after PP Nikon raw files I can see how the DxOMark comparison holds true.

6th December 2010, 08:59 PM
meaningless set of numbers if you ask me

I used the site to compare cameras that I own and can see differences in and they come out pretty much the same so IMO either the method or the conclusions is flawed


6th December 2010, 09:15 PM
If you believe these figures the E-3 is better at high ISO than the E-5 and that's what we've all seen in the pictures posted so far, right? *laugh*laugh

I actually want to believe these stats as it will save me money, I might as well keep the E-3. :rolleyes:

7th December 2010, 12:36 AM
I consider all this comparison similar to that of comparing cars. My car has a theoretical top speed in excess of 130 MPH but I will never use that capability.

I did however compare the E-5 with the E-PL1...

It confirmed what I had already determined.

The E-PL1 is brilliant :cool:

Nick Temple-Fry
7th December 2010, 02:47 AM
A quick play with the figures available would suggest that for each family of sensors (for example the 12.1mp sensor seen in the Pens, 620 and now the E-5) figures are fairly close.

That signifigant differences only occur between disparate sensors (eg the E-3 and the E-5).

That if we take the experience of users of the E-5 (both 'professional' testers and group members) the E-5 represents an appreciable improvement in noise/iso performance. This perceived improvement would seem to be at odds with the DxO figures as presented.

The presentation by DxO takes 'measured iso' and compares it with another metric (snr, ev).

If you take 'claimed' iso and compare it with the metrics then you get a far better correlation with the observed results of users.

One of the things banged into my head by tired professors when I was studying physics was if experimental method produces results that diverge from what is actually able to be observed then you (as the conductor of the experiment) have a problem. Regardless of how good/pure/elegant the experimental method was. Basically it meant that you have failed to understand sufficiently the problem.

DxO results would seem just to read the raw count of the photons (or charge) that fall into each pixel pit. But we know that a picture is more than that, that an image taken on two different implementations of the same sensor can manifest apparently different levels of exposure (just in terms of brightness ignoring colour cast etc that might be due to other factors), that's why raw developers do more than count the electron piles. Basically the result is influenced by more than one type of data, but DxO would seem to be measuring only one paremeter.

Throughout Olympus has said the improvement seen in the E-5 is through the management of the sensor in its processing engine (with the exception of the aa filter), not in the basic sensor design (capturing/counting photons). Could it be that the DxO methodology is too pure/elegant to really reflect this?


Roy Clarke
7th December 2010, 04:42 PM
In Response to an information request asking when DXO will be supporting the E5 as it did not show on their latest roadmap, I received the following:

"Bonjour Roy,

Our intentions, to the best of our knowledge going out ~ 6-months is as you stated. There have as yet, been no plans made regarding support for the Olympus E-5 that I can share with you. Once these decisions have been made, they will be shown on this listing."

Not much help coming from the French quarter then.



7th December 2010, 05:20 PM
I'm a big fan of Capture One, great for pulling every last detail our of a RAW file. Version 6 supports the E-5 very nicely.

7th December 2010, 07:28 PM
I must say that those numbers and conclusions seem very much out of step with the majority of reviewers.

Naturally I hared on over to the D7000 Club on Flickr to view the amazing dynamic range and IQ which must of necessity (if this DxO report is any good) mean scores of brilliant photos. And guess what - much the same as most other groups (I also looked at the practically defunct D3X group while at it) and not really as good as the E-System Community pool on Flickr. But DxO must be right: I now really regret buying this crummy E-5, out of which by day 3 on a grey Tuesday I get rotten IQ like this in test shots:
...must try harder on those highlights though...

9th December 2010, 01:51 PM
The DXoMark figures for the E5 now appear to have been removed from the site....

9th December 2010, 02:07 PM
The DXoMark figures for the E5 now appear to have been removed from the site....

Yes seems to be the case, I wonder why though....