DxO Labs Launches DxO Mark
News by woodyjnr
DxO™ announced today its launch of DxO Mark™, a beta product which aims to provide detailed test information based on RAW data. Many websites base their findings on JPG output, which only tells us half the story. It is fair to say DxO Mark™ may change the way both the industry and consumers evaluate and compare cameras. DxO Mark™ is, in their words, "the first publicly-accessible database of objective and in-depth RAW sensor image quality measurements".
The test data concentrates on 8 main categories (ISO sensitivity, SNR %, Dynamic Range, Tonal Range, Colour Sensitivity, Full SNR, Colour Response and Full CS), and interestingly includes test data not only from DSLRs but also from some high end compacts too. For those of you wondering, yes they have tested Sony Alpha cameras and the A900 does rather well (to find out more click the link at the bottom).
AlphaMountWorld.com commends DxO™ for pushing the boundaries and providing some of the most comprehensive test data out there.
It is a fascinating site and resource and can be found on the following link: www.dxomark.com, detailed test procedures can also be found on their site. If you value unbiased and thorough test data then it's well worth checking out.
AlphaMountWorld staff.
Fascinating and Fair
From what I have seen here: http://www.dxomark.com/index.php/eng/Image-Quality-Database
So far, looks like many of the higher MP DSLR's shoot at lower ISO values than stated. I find that very interesting.
Definitely worth your time to poke around here on this site. Now we just need to get DXO to get the A700 up there.
Carl
Love it
Also can't wait for them to put up the A200.
I am very sceptical with
I am very sceptical with some of those DR figures, said this before, but there is no indication of IQ, and when you pull images about, quality suffers. Looking at the imaging resource tests, you can see they add a quality part too it, high, medium, low etc. I doubt you would get poster prints off the A900 with the max DR number on there. Even on IR, they say look at the real shots, and not too much at the numbers.
Best way is to have a look at what is going on in the real world, might do a few film v A200 tests, just on this alone, even then..it's hard to directly compare, as the mediums are very different. Too early to tell on the A200 for me, DR wise appears better in the highlights compared to the Km5d, not sure it matches it shadow wise..will see. Metering is a little more conservative, than the minolta, not a bad thing for WA shots..
A200 def looks better than the 450d I was playing with recently, though that tends to overexpose a bit..
You can never have enough DR, not in my books, erm add +3 stops in the highlght end and we get to a happy place!
Their test methodology is clearly stated...
Their test methodology is clearly slated on the website and yes the figures do need to be taken carefully and used in conjunction with real world samples. However the dynamic range of digital SLRs has been improving at a vast rate for years- it is now very, very good with a RAW image.
I shall post this link to an image I took a few months back because I feel it really demonstrates the huge DR that is available in RAW. I was shooting a sunrise image and didn't have a ND grad to hand quick enough to capture the boat, I thought I had lost all highlight detail- indeed the screen and histogram did indicate that a large amount was "white". When it came to developing it, I was stunned at how much detail was recoverable. A lot of this will depend on the software you use also- I'm using Aperture which has very good adjustment tools (recovery being particularly useful in dealing with over exposed areas).
Image is here: http://www.alphamountworld.com/forums/photo-sharing-forum/wells-next-sea-sunrise/2008-7/1180
Having used both the KM5D and A300, the DR of the A300 is much higher.
Andrew
Only done a quick look at DR
Only done a quick look at DR on the A200. It is most def better in the highlight end (which is good), but does not appear to have the same depth in the shadow region. Not that I can really complain, the 5d could clip a bit at times..the A200 seems less likely to do this. I would imagine the A300 shares the same sensor.
If the tests mean anything, not sure but here they are from IR: (a200)
http://www.imaging-resource.com/PRODS/AA200/AA200IMATEST.HTM
Here is the 5d:
http://www.imaging-resource.com/PRODS/KM5D/KM5DIMATEST.HTM
Almost the same.
So far at least, the A200 does better with jpegs DR wise than the minolta, and it does not seem to lose the DR with higher ISO, or least not as much. Not that I am defending the KM, because DR could have been improved..however, it was good for a digital, and pulling up the shadows worked well, the A200 isnt as happy with this. I would rate the A200 a bit better DR wise..but not "loads" I think the DR has improved, but not to a huge level..if you look at the IR test on the Minolta, it says "severe overexposure", and quality is low. Low to me means not much good for print. So really these 11-12 stops quoted on sites, don't have any real meaning to me. The high and med-high seems more realistic.
I did shoot the 5d side by side with film, and you can get near it (but the quality is not as good overall), and with some effect on the tones and colours, very hard to match neg film. So to my eyes, no we are not at neg film levels. And more to the point, neg film is very tolerant of overexposure..which is useful with strong contrasty light. It's those extra few stops highlight end, that makes all the difference, and give the gradual roll off on film. That is what I am aiming for..and we ain't there yet!