| Introduction | Measurement | Comparisons | Conclusion |
The 18Mp CCD previously used in Leica M rangefinders hasn’t fared well in our tests ranking towards the bottom for full frame sensors. The new 24Mp CMOS model is a big improvement however boasting 6 million more pixels than its predecessor and delivering a 1 Stop improvement in image quality, too. With a +1 Stop improvement for both Color Depth and ISO, as well as an impressive +1.6 Stops better Dynamic Range, it seems switching to a CMOS sensor was certainly a good move for Leica. It’s interesting to note too that no RAW smoothing is applied to files from the new model, as was previously the case from ISO 400 on earlier versions.
Two of the best full frame sensors we’ve tested in the fixed lens Sony RX1 and Nikon D800 DSLR offer a +2/3rds of a Stop image quality overall compared to the Leica M. Drill down into the different categories and the superior RX1 and D800 sensors deliver + 2/3rds of a Stop better Color Depth and Low-Light ISO performance. The Leica M’s measured base ISO of 134 compared to ISO 81 and ISO 74 for the RX1 and D800 respectively mean the Sony and Nikon cameras have an immediate advantage and deliver +1 Stop improved Dynamic Range overall with scores of 14.3 (RX1), 14.4 Evs (D800) and 13.3Evs (M). As sensitivity is increased on all three cameras Dynamic Range drops at roughly the same trajectory up to ISO800 where the Leica M takes a further dip. The impact of this is good Dynamic Range of 10Evs and over is only achievable on Leica M up to ISO1600, while the RX1 / D800 are good until ISO 3200 and these better sensors offer the same Dynamic Range at ISO 25600 as Leica M at its maximum sensitivity of ISO 6400.
Compared to the full frame sensor in Canon’s flagship DSLR, the 1Dx, the Leica M just wins overall with 84 points to 82. For Color Depth there’s little in it with 24 bits for the M compared to 23.8 bits for the 1Dx, so it’s left to Dynamic Range and Low-light ISO to separate them. At base ISO sensitivities the Leica M offers significantly improved Dynamic Range +1.5 Stop better than the 1Dx, although this changes quickly as sensitivity is increased.
While the 1Dx delivers consistently good Dynamic Range of around 11.8Evs up to ISO 800 the Leica M performance drops markedly as sensitivity is increased and by ISO 800 the two cameras are the same. The trend persists past ISO 800 too where the Leica M’s Color Sensitivity continues to drop rapidly and the 1Dx offers better Dynamic Range at ISO 25600 than the Leica M at ISO 6400.
Nikon’s top end DSLR the D4 ranks 1/3rd Stop better than the Leica M overall scoring 89 points compared to 84. The D4 is also ½ Stop better for Color Sensitivity at base ISO with 24.7 bits to 24 bits, but as the Leica M’s performance takes a dip as ISO sensitivity is increased the D4 offers +2/3rds of a Stop Color Sensitivity at ISO 1600 compared to the Leica M at the same sensitivity. While the D4 and the Leica M have similar Dynamic Range at base ISO sensitivity of 13.3Evs (M) to 13.1Evs (D4) the same pattern emerges as sensitivity is increased, so while good Dynamic Range of +10Evs is possible on the Leica M up to around ISO 800 the D4 delivers the same results up to ISO 3200.
No doubt you’ve picked up on the pattern here that while the Leica M can deliver comparable results at base ISO it struggles at higher sensitivities. With a DxOMark Low-light ISO score of 1860 ISO (M) compared to 2965 ISO (D4) and 2786 ISO (1Dx) the rangefinder offers around -1 Stop ISO performance compared to these flagship DSLRs.
|
|
Added by Nick932 |
April 21
Validity
I just wonder why all pictures that I have seen taken with a Leica look better than the ones taken from Nikon. However the marks are not very great. It seems that there factors that are not measured. I did return my D800E back. It seems that the Technology is not there yet.
Reply |
|
|
Added by Nerval |
March 15
As good as it is...
The point of shooting rangefinder, and any way investing in a Leica M is not cracking burst shots at sport events, I take it...
One thing though, Leica did really make a fuss about this new sensor and all, well at least here in Europe, got some friends who were invited to a couple of events by the store in London (they're customers and Leica tends to take care (cling onto) their customers, well Leica does not sell potatoes...). And, here is the result, you'd think given what they ask for a M, they would fit it with top notch sensor technology (I'm thinking what Canon or Nikon (I mean Sony) manage to achieve). But the SNR curve is not that impressive and most of all the dynamic range, which is wrongfully labeled as landscape here on DxO, drops quite rapidly to below APS-C sensor average past ISO 800. And well DR at base ISO does not tell you the whole story, it's quite the opposite. At base ISO you'll barely even use more than 12 ev, unless you fancy over-everything radioactive HDR finish. BUT DxO measures DR from where the noise levels cripple the shadows to the highest reproducible gradations in highlights. Thing is, as you increase the amplification of the signal, the highlights are always maintained very well, the shadows however... that's a different story. Meaning, despite the score of ISO 1800 in acceptable SNR, the shadows turn to mush quite fast, which means the camera is already under-performing above ISO 800. Canon for instance has never done too well in the low ISO DR department, however what made their cameras perform "low light miracle" is that they manage to keep a reasonably good DR at high ISO, on par or even better than their Sony (I mean Nikon) counterparts. Sony's chip manage a tad cleaner SNR. My point is it's better to have a better headroom than sacrificing tones for the sake of a bit cleaner file... Cause when you have 8 ev of DR in your Raw file, in general it's game over, and well, the M does it two full stops earlier than its competitors, or so it seems cause the graphs do not display the data, DxO has not updated the site as of now. Isn't it a bit disappointing? (guess it still is an improvement over the M9). Reply | Read all replies for this comment |
|
|
Added by MikjoA |
March 17
Re: As good as it is...
Why dynamic range would be wrongfully called landscape, since having a good dynamic range is mostly important for landscape ?
The curve may be steep, but who would take landscapes at 800 ISO ? Most people shot landscapes at 100 iso or even less if the camera allows it. Anyway, 84 is surely a good score, but I'd expect it to matches Sony's sensor... Reply |
|
|
Added by bobn2 |
March 14
The problem
Looks like rather low quantum efficiency.
Reply | Read all replies for this comment |
|
|
Added by Nerval |
March 15
Re: The problem
Well, the dynamic range has a rather linear curve, so you could say that it follows a more ideal path of analog signal amplification than the 6D for instance... Except that it's not right, because the curve for this Leica is... well... Rather steep.
Reply |
To provide photographers with a broader perspective about mobiles, lenses and cameras, here are links to articles, reviews, and analyses of photographic equipment produced by DxOMark, renown websites, magazines or blogs.