M9 Dxo tests

[Bill Pierce]

The Dxo tests are in for the M9 sensor.

http://www.dxomark.com/index.php/eng/DxOMark-Sensor/DxOMark-reviews/DxOMark-review-for-the-Leica-M9

The results are not very good. It will be interesting to see the reaction of the Leica community.

 

[Keith]

I guess when it comes down to it it doesn't matter a damn ... if you're happy with the images from your camera these tests mean little!

That said Leica is certainly behind the pack with a sensor that rates on a par or less with the lower priced pro-sumer DSLR's from what I can see. I guess a few who have paid the asking price for the M9 will be wringing their hands in anguish!

I couldn't help checking the D700 results though ... tra la la! 😀

 

[photogdave]

The photos look great in real life. Who cares?

 

[Keith]

The photos look great in real life. Who cares?

Leica owners don't care much I supect as their choice of camera is based on different needs ... but sales of the big two's DSLRs must be influenced in some way by this sort of stuff surely?

 

[rya]

I wish there was a very quantitative way (like DxO is with sensor quality) to measure the rangefinder experience and M-mount glass characteristics against other cameras. I don't think sensor comparisons could make that great of a difference until there is another full-frame digital rangefinder.

Medium format film is technically better than 35mm but that doesn't mean (most) people want to carry a MF system over 35mm.

That was my reaction.

Edit: I do not wish there was a way, but it would demonstrate my measures of a camera.

 

[maddoc]

A full-fame sensor camera with the size of a M Leica (almost) and being able to use all these fantastic lenses ever made in screw- and m-mount, I could care less about the measured performance of the sensor. But then ... I don't care about MTF charts either. 😀

 

[rya]

Gabor said it better.

 

[Gary Sandhu]

The comparison between the m8 vs m9 is telling -- 3/4 a stop noise difference. Maybe just use a flash at iso 160? 🙂

 

[tanel]

The test seemed a bit overdone. Something like looking at the corners of the uncompressed and unedited picture and whining about sharpness. But in real life the difference is barely noticed.
Then again, I haven't had the opportunity to use M8 or M9. But the fact that people tend to overdo in these tests remains, in my opinion.
Besides, it's not the pixels that makes a good picture. 😉

 

[semilog]

Here is a limitation of the dxomark tests: they are consistent across different sensors. Why is this a problem? Because the requirements for the sensor in a rangefinder are quite different than the requirements for the sensor in a DSLR. Specifically, the rangefinder's sensor has to accomodate light coming from much more acute angles. Put the sensor from a D700 into an M body, and it's extremely likely that the across-field shading and color shift problems would be horrible.

On the other hand, the wide angle lenses available in M-mount blow the wide angles for DSLRs out of the water in performance (usually), speed (sometimes), and compactness (almost always). You pay your money and make your choices.

Me? I'll keep shooting film and save the expensive sensors for work (in that world quantum efficency approaches 90%, CCDs still trump CMOS, and sensors start to perform at interesting levels once they're cooled to -40° or so 😛 ).

 

[gnarayan]

The results are not very good. It will be interesting to see the reaction of the Leica community.

The reaction to quantitative tests are always the same if they don't tell us what we already believe.

1. The tests themselves are biased and wrong
2. The tests are not complete and do not account for some factor
3. The tests do not include variance and my particular sample is not represented by the test
4. The tests are too objective and I prefer to compare based on my own subjective criterion that I feel are more real
5. The tests aren't wrong but are too complicated so I choose to ignore them
6. The tests are fine but irrelevant to the real world so I ignore them
7. The tests are fine and relevant to the real world but I simply don't care
8. The tester/reviewer is inexperienced, didn't test long enough and their opinion is meaningless
9. The tester is malicious and biased and skewed test results


I don't know why you'd consider the reactions here to be interesting. They will inevitably be one of the above. A lot of the objections that will be raised in here will be perfectly quantifiable or already quantified. Some will be irrelevant. A few people might actually have a point. You won't learn anything but you will get a lot of noise.

A lot of people think that the boy that pointed out the that the emperor had no clothes on was merely rude.

It is actually worse if the test tells you what you DO want to believe. Then the tendency is not to even verify if there were any issues with the testing to begin with.


Cheers,
-Gautham

 

[Roger Hicks]

The reaction to quantitative tests are always the same if they don't tell us what we already believe.

1. The tests themselves are biased and wrong
2. The tests are not complete and do not account for some factor
3. The tests do not include variance and my particular sample is not represented by the test
4. The tests are too objective and I prefer to compare based on my own subjective criterion that I feel are more real
5. The tests aren't wrong but are too complicated so I choose to ignore them
6. The tests are fine but irrelevant to the real world so I ignore them
7. The tests are fine and relevant to the real world but I simply don't care
8. The tester/reviewer is inexperienced, didn't test long enough and their opinion is meaningless
9. The tester is malicious and biased and skewed test results


I don't know why you'd consider the reactions here to be interesting. They will inevitably be one of the above. A lot of the objections that will be raised in here will be perfectly quantifiable or already quantified. Some will be irrelevant. A few people might actually have a point. You won't learn anything but you will get a lot of noise.

A lot of people think that the boy that pointed out the that the emperor had no clothes on was merely rude.

It is actually worse if the test tells you what you DO want to believe. Then the tendency is not to even verify if there were any issues with the testing to begin with.


Cheers,
-Gautham

Dear Gautham,

A perfect analysis, except that I'd say that (at least) one of the nine applies whether the results confirm our prejudices or not.

Cheers,

R.

 

[Pickett Wilson]

To the Leica Man, objectivity is superfluous. If you just paid $7,000 for a rangefinder camera, it's "the best camera in the world!" 😉

 

[Ronald M]

Looks as if this is nothing but an M8 sensor made bigger with some change to the microlenses on the edges.

If you do the comparision against Nikon D700 & 300s, it hold up well against the 300s.
Not so well against the full frame 700.

All that said, one can use the faster Leica glass at lower ISO where Nikon people have to stop down and use higher ISO.

One must not forget the purpose of thecamera is reportage, at job at which the camera still excells compared to Nikon, even full frame.

Low ISO shots are more than adequate, high ISO useable. Years back, I use low ISO film almost exclusively adding light where I had to have more. Adopt the same usage.

IMHO most of these low/available light shots look bad anyway because of the harsh lighting and high contrast and are not worth wasting electrons on anyway.

If I needed a tool such as the M9, I would buy one. The quality is more than sufficient.

 

[Tracnac]

Humm some strange reacts here... for me bad sensor means bad image even with the best optics (especially dynamic range and high-iso noise)... this just confirm that the M9 is electronically poorly design.

But I agree if you like your images this test is void.

Yvan.

 

[user237428934]

When I look at the test then it shows what everybody already knew: there are far better cameras on the market for high iso photography. This test does not show that it's the only full frame rangefinder on the market.

 

[Sonnar Brian]

This is a camera-level test, not a test of the complete "system". A complete system also consists of post-processing hardware/software.

Leica uses a Kodak CCD, not a CMOS sensor. The Leica M9 is a small, compact camera without a huge battery and power-sucking DSP board to clean up a lot of noise present in the detector. Large DSLR's have huge (by comparison) batteries and much more powerful onboard processing. A camera-level test is going to look better. You can download the long-spec sheet on the Kodak CCD's. Does anyone have links to the CMOS sensor used in the Canon and Nikon DSLR's?

 

[user237428934]

Humm some strange reacts here... for me bad sensor means bad image even with the best optics (especially dynamic range and high-iso noise)... this just confirm that the M9 is electronically poorly design.

But I agree if you like your images this test is void.

Yvan.

😀 What a nice rant.

Compared to all full frame 35mm cameras (Sony 900/850, Canon 5/1, Nikon 700/3) the M9 is the last in line in this test.

Compared to all cameras it shows a very good result because it beats every crop DSLR around (even Nikon 300 and Canon 7)

But who cares. I don't think that someone does not buy a M9 because of the DXO test. There are other criteria for M9 buyers.

 

[Turtle]

I think it boils down to whether the results are good enough for you. Certainly the M9 is the only FF digi RF and there is no doubt whatsoever that there are many uses for which SLRs are far less viable. If this translates into shots gained that would never have been taken then this is the counterbalance to these tests. When I consider whether to get a M9 or wait for the 9.2 or M10, this will be how I think, not in directly comparing with the Eos 5D Mk III or the newest FF SLR of the time. That performance would count little if I could not harness it for my type of shooting. I DO use a SLR with a 85mm for portraits, in preference to a RF, but I use my 35mm RF in preference to my SLR or MF RF for light, fast, on the hoof work where it trounces the other cameras in utility, for me.

I feel the M9 is a bloody good start with most things in place. Now that the concept is proven it should be relatively easier to implement gradual improvements in the areas where it is now somewhat less that the top SLRs. I don't think the camera would benefit from much more MP, but lets imagine what a 22-25mp M10 with better high ISO and dynamic range would be like. I would be hard to know what more to ask for and so I believe that with the next incarnation we will be commencing down the road of diminishing returns for most people. Once you have stunning A2 shots full of detail and solid performance at ISO 2500, what more is needed.

Utility, ladies and gentlemen. No amount of technology or force is much use without it.

 

[Renzsu]

All I care about are the results from the camera/lens system and imho they can't be beat by any of the dslr's on the market. Clearly the ISO performance isn't up there with the Canikon ones but I don't believe for a minute that image resolution and sharpness are less than on those dslr's.
All of this is moot though when you're looking for a small compact full frame camera.. then this is basically your only option.