Dx0

Bill Pierce

Bill Pierce

Well-known
Local time
5:35 AM
Joined
Sep 26, 2007
Messages
1,407
This has been reported on several websites; so, it will be old news to many of you. Dx0, the folks that produce some exceptional image correction software for specific cameras, have released some of their raw image evaluations for specific cameras at a new website

http://www.dxomark.com/

The information is sorted into various performance categories and an overall image quality rating. In that last category the M8 comes out 33rd in a list of 51 cameras. Not very good for a camera released relatively recently and with a relatively high price.

Has anyone who has spent some time on their website any thoughts? Two things the Leica was always identified with were "small, quiet and rugged" and exceptionally high image quality. This forum and others have already dealt with the first. How about image quality?

My feeling is that the lessened anti aliasing and good lenses are a real plus in the fine detail department - and that Photoshop is a great leveler in the tonal separation and medium detail departments. I think this lets the M8 compete successfully with other top cameras, but I don't see the clear superiority in image quality that others see.

Any thoughts?
 
The site is a pretty good resource for people who worry about stuff like this. DxO are also writing about how they do their measurements, and as a company really probably have no bias towards one manufacture or another.

The M8 is a surprisingly poor performer based on these metrics. My Rebel XT looks to be a comparable from an imaging standpoint. As you point out, Lecia makes some nice lenses, but it's not like Nikon and Canon don't as well. You can buy Zeiss class for your canon's now as well.
 
Since my not very expensive dSLR scores much higher than more expensive models from Canon and Nikon on that site I am completely convinced that their metrics are perfect ... ;)
 
You just know that people are going to use these comparisons as the final word in which camera is "better", and use it as a tool to back up their silly arguments.
It looks like Photodo for cameras. I've actually met people who will not buy a lens that did not get a top score on Photodo - ridiculous!
There are too many variables that the DXO testing can't take into account. I honestly don't see it as a very valuable tool and potentially harmful in fanning flames of internet arguments.
 
At first this stunk to me of the Photodo lens test site. But there were some very interesting things I read on the site and it confirmed some of the issues I had with a certain camera, but could only express qualitatively and only in terms of "I think that...."

I'm sure everyone will find an issue on the site to fault, especially if the site "rankings" aren't kind to their camera. But while not everything, technology is an issue (we wouldn't have photography without technology), the technical issues still are important and this site has value. It's not the be-all/end-all, but can provide useful information.
 
I thought it was great, unitl I saw what came out ahead of my camera:)

Actually, I think it is useful, but only in as far as it goes. DXo would be the first to admit this I think. I think it's useful as it reveals how level the playing field is at the top and down the tree - The canon and nikon fanbase need something that says, 'Hey, both companies make cameras with similarly (very) high levels of ablity. Stop[ being childish.'

Mike

Mike
 
You can get into the "too much test report" syndrome in which two cameras are very similar in the overall ratings. One is, by the evaluation, an 89.3; the other, an 89.2. (well within the statistical error we heard so much about during the presidential campaign season) Suddenly the 89.3 is a clear winner, clearly the better camera. And the 89.2 is something loathsome you should stay away from.

That's obviously silly. But, when there is a significant difference in score, we should ask ourselves why and really check out the test results category by specific category, asking ourselves if the difference is pertinent to our photography. (Studio photographers using fast prime lenses aren't as worried about high "ISO" performance as available darkness shooters with slow zooms.) I find these reports very useful, not so much in the overall ratings, but in the ratings that refer to specific image characteristics.
 
I agree. An interesting coparison is the D3/D700 and 1Ds3. There is, in all practical senses, nothing much between their scores, which hopefully reflect the actual overall efficiency of the sensors. In use I supect that each have their strengths whether you need clean high iso or more resolution at low iso for bigger prints.

Mike
 
It's interesting that DxO evaluates the RAW data directly as it is stored on the memory card. This is appealing as their analysis reflects the starting point for the image before it is manipulated and modified. Of course many cameras actually modify the RAW data before it is stored on the data card. From a scientific (signal processing theory) point of view, this is unfortunate. Stll. the DxO analyses avoid the variables introduced by using different post-acquisition RAW software algorithms.

Because the M8 sensor is relatively old, it's no surprise that newer sensors produce superior RAW data. The M8 is a value-added product as you are paying for the ability to use M/LTM glass and benefits of the rangefinder experience. The sensor seems competent enough, and I doubt that photography best suited for a M body suffers significantly (assuming one wants B&W images or religiously uses IR filters).
 
You know, at one time I just took pictures.

I never read test reports and I didn't buy much equipment. I used what I had, shot what I pleased and enjoyed stuff more. Ever since technology began to dominate photography, there's something new eight times a day and I find myself hitting all the photo gear review websites several times a week and seeng new information on updated software, firmware, and hardware. My new Canon G10 (thanks for your input, Bill) was delivered a couple of days ago and I doubt I get to finish reading the instruction book before a 2.0 version is announced.

Image quality is over-rated. That's a paraphrase of something Walker Evans supposedly told his lab assistants who kept reprinting his photos to try to make a perfect print.
 
To echo what dogman just said, i've yet to have someone look at my photos and then go on to complain about chromatic aberration. I guess because this is a gear forum, people are more prone to focus on the more technical and esoteric aspects of taking pictures.
 
Last edited:
M8 Image quality

M8 Image quality

I have been a photo enthusiast for over 40 years, the majority with Leica rangefinders as my principal camera.

The M8 gives me outstanding image quality. I shoot raw +B&W jpegs and the jpegs are usually so good that I do not even convert the raw. This is upto 8.5 x 11, often with moderate cropping. A semi professional friend of mine has a D300 and he is very happy with it but he says the image quality of the M8 is superior.

For me this is a non argument. I would never consider using a DSLR tank even if I wasn't tied to optical viewfinders.

All these comparisons are really useless. Given any opinion I can find someone on the forums who will agree with it. I have to see a print in my hand before I will make a judgement.
 
I'm not sure what is being measured. Is it just the sensor? Does it include the fact that Leica has that thin IR filter. Are they just measuring RAW files or does it matter what happens when converted to a usable file from RAW? I wonder if that is the great leveler. I imagine that a better RAW file provides for better latitude in adjustments, but if you take a "suitable" exposure in the first place...Many people scorned the quality of the 35mm portable cameras in the age of large neg graflex news photographer. Those digital Hasselblads must still blow it all out of the water. I'd bet that a Canon G10 is probably as good as 35mm film. What's enough image quality? Is it ever enough? How large do you want to print your M8 files?
 
Last edited:
There is no doubt people are prepared to pay for the M8's ergonomics over everything else - with all the inherent RF 'advantages'. I don't believe that many people are paying merely for the Leica name anymore. So in that respect, IMO a digital RF from Zeiss or VC would literally kill the M8.

At the moment though, assuming I did have that money to spend, I would definitely buy an M8. Reason: there is nothing in digital that would give me more pleasure to use. Do I think a D700 is THAT much better in quality terms? Yes I do, but it probably wouldn't change my mind - even if it were used purely for commercial applications (unless of course it were for applications unsuitable to a RF). From the company's perspective, with the S2, the M series suddenly isn't professional anymore. Bit like what Hasselblad have done to the V system and it's users.
 
Last edited:
I think many folks have misunderstood what Dx0 is presenting in terms of information. They are evaluating the capabilities of the sensor and its raw images. They do not measure lens quality, focusing accuracy, ergonomics, e.t.c.. Nor do they measure the quality of some image processing program on your computer. Thank goodness. For the rangefinder user, it's not unusual to use a broad spectrum of lenses. Individual lenses are cammed to individual bodies over an unfortunately broad range of accuracy. People convert raw images with a great variety of programs.

If you don't use raw, these evaluations are worthless. What they really talk about is your camera's potential. They have added a q & a section that explains what they do (and don't do) in detail at http://www.dxomark.com/index.php/eng/News/First-week-s-Q-A.

I recommend it heartily.

Bill
 
I think the DxO tests are very useful, if understood properly. Still, I can almost hear the directives coming out of Tokyo boardrooms: "Toshiro-san, the Director wants you to design the next camera to get the highest score on the DxO tests." Shades of computers being designed primarily to ace the PC Magazine benchmarks.

And yes, the tests will be the subject of umpteen Internet arguments, cited with the fervor of a fundamentalist quoting Revelations. :D

Regarding the M8, a few things come to mind after reading the FAQs Bill mentions above:

* The way DxO calculates the scores, cameras with a base ISO of 100 will score higher than cameras with a base ISO of 200.

* If they didn't use an external IR cut filter on the M8, colors would probably be contaminated, and this could partially explain the M8's comparitively weak color scores. (Does anyone know if they used an IR filter or not?)

* They don't appear to be looking at per-pixel sharpness, because that's something optics would bias. They seem to be mostly measuring signal to noise ratio. So the tests measure areas where the M8 is a bit weaker than the big Pro cameras, and ignore the very areas where the M8 shines: fine detail and very fast, superlative optics. Actually, the M8 didn't do badly compared to many APS-sized sensors, even in the high-ISO tests. Once we get to 1250, the M8's performance falls. But we knew that.

Anyway, the tests do show that big, full-frame cameras with big pixel sites score better than APS-C sensors, that newer cameras tend to score better than similar older ones, and if you want an APS-sized DSLR, the Nikon D90 is one nice machine.

To really get useful information out of the tests, read the detailed graphs. and remember, it's not the simple rankings that count, it's the general "band" into which the camera falls.

--Peter
 
You must log in or register to reply here.
Back
Top Bottom