The M9 Sensor is more than adequate by George Sutton
This post is a response to the recent DxO report on the Leica M9 sensor. I chose to respond this way because I can include photos. Photos are, after all, the final word in this whole discussion.
The overpriced and under featured M9 body only exists because it has a full frame sensor and mounts Leica M lenses, but that is enough to be one of the best cameras made. The M9’s biggest drawback is a lack of versatility but in circumstances where it performs well it produces some very good images. I am not disputing the DxO results but to me the take away is that there is not a great deal of difference between high end sensors in actual use. I say that after owning and using a M9, Canon 1Ds, 1Ds III and now a 5DIII, and an Olympus OM-D. To me, the telling thing is the big diss DxO gives the 1Ds. When it was first released the 1Ds was probably the best camera made. It was way ahead of anything Nikon offered (they have played leapfrog since) and it even surpassed medium format cameras for detailed image quality (there were no medium format digital cameras at that time). Yes, that was then and now there are better cameras but the 1Ds still produced great photos. What I have learned in the meantime is that the single most important factor in a camera’s quality is the lenses. The biggest drawback to the 1Ds was the soft to unusable corners in many Canon lenses back then. Nothing, in my experience at least, equals the quality of a Leica M lens. The following illustrate this point.
One of the toughest camera tests for me is shooting a city at night.
The shots below are all taken at f8 and the camera’s lowest ISO on a tripod with cable release and are close to 100% enlargements for the Leica and Canon and about a 125% enlargement of the OM-D. I selected f8 because it produces star like effects around lights and is typically the sharpest aperture for any lens. The images are somewhat flat because it was hazy and I was shooting from a few miles away. The shot with the Leica was taken with a Leica 90mm f2.8 that I bought used. My guess is that the lens is 10 to 20 years old. The shot with the Canon 5DIII was taken with a new 24-70mm f2.8 II zoom at 70mm. That lens is generally regarded as the best medium range professional zoom currently made and it is very sharp corner to corner. On the OM-D I used a Lumix 12-35 f2.8 zoom at 35mm (equivalent to 70mm on a full frame camera), which is generally regarded as the best medium range zoom for a 4/3 camera. Detail in the buildings is close for the Leica and Canon. The OM-D is worse but that is mostly due to the smaller sensor. Printed 8×10 these differences would be barely visible. The biggest difference is the lights. Note the clear multi point stars produced by the Leica. The Canon is close but the rays emanating from the lights are slightly less distinct. The OM-D is the worst. The star effect is there but the lines are distorted and broken with what appear to be concentric circles radiating out from the light. The star effect can be eliminated by shooting the lens wide open. Wide open the Leica and Canon both did a great job of capturing the light as it was. The OM-D did not do as well. I tried different lenses on the OM-D including a prime and got a similar effect each time. If I were shooting this for sale, I would shoot it with both the Leica and Canon and pick the best. If I could only shoot one it would be the Leica.
I offer these only to illustrate the point that in use the M9 sensor is quite adequate to get a great shot. I am including one more shot to make this point (the last landscape image). The landscape is cropped from the original by about 30%. It was taken with the M9 on a tripod with a Leica 35mm f2 lens at f11. I don’t know if this can be seen in the image here but I have printed the cropped image at the largest size my printer will do, 17×22, and individual bushes about a foot wide can be clearly seen on the desert floor more than a mile below. I haven’t used every camera and lens made but of those I have used I have never seen this level detail from any other camera. That is mostly due to the lenses but the sensor has to be up to the task as well and in my experience the M9 sensor is more than adequate for the job.
Leica
Canon
OM-D
Great site Steve.
When I go window shopping, things that attract me visually come first. I move towards it, feel it, and check the price while inspecting it. If it’s out of my reach, price wise, I put it back and move on. Sometimes I just can’t resist and must have it at all cost.
There are things that I can accept a lower quality for savings. but, there are things I must have regardless of whether it is slightly better than the next one at twice the price of it. When I first saw images from a Leica M9, I was struck by that magical look and feel from the images.
I love driving fast cars and a VW would be good enough; there is no need to get a Porsche GT. It’s easy to justify this because I can’t afford the later. The Leica M9 and now the new Type 240 are tools that I know that I can exploit to the fullest from an art perspective because I love photography as an art. Give me a Porsche and I’ll just be showing off as I’m not a great race car driver who can exploit her potential.
Is it me or does anyone find the term “more then adequate” pretty sad when describing a camera that cost 2-3 as times as much as anything in the 35mm sensor world. The leather on this Hermes bag is more then adequate…..etc, etc, etc. The spin doctors and defenders of the absurd are hard at work. I see little differences in those photos that can’t be adjusted in PP.
P.S. – Please don’t attack me as a Leica hater. I own a Leica M3 with Summicron 2.0 lens and I enjoy it. But Leica lost its way. They don’t innovate much. They were a great optics company that can create a mechanical camera with the same rigor as a Swiss watch. But in the digital world they have been exposed as an overpriced camera. Extremely overpriced….
Shooting Leica M is about the experience of shooting a simple, well built, rangefinder camera, which is an entirely different experience than shooting an SLR or EVF-based camera. I’m not saying that the rangefinder experience is better or worse than other methods, but, if that’s the method of working that you like, then Leica is the only digital game in town.
A base Porsche 911 has less horsepower than a Corvette, but the feel of driving a Porsche is much different, so, if that’s what you’re into, it’s an easy decision. The engine isn’t the only important part of a car, but rather a single piece in the chain.
Either way, just about any sensor from any m4/3 and up camera over the last several years is likely good enough for the large majority of photographers. I’ve owned several cameras that “rate” higher than my M9 (which I bought used,) but it’s been a non-issue for me, and I actually prefer the M9’s output in most situations. Sensor IQ has plateaued, for the most part, so all of these relatively minor differences aren’t going to make much difference in normal usage.
p.s. new Leica film cameras are still $5K, so it isn’t as if the digitals are hugely more expensive.
Your points are valid. I do enjoy the M3. Just find the M9 extremely expensive. Would probably look into a used one once they hit 1500 or so. I believe without the Leica name 1,500 is what they are worth.
$1500 is a lot. It includes no glass and has a narrow ISO range.
Justifying a Leica on price/cost is a losing proposition. Plus, if one enjoys that type of shooting/experience then no justification is needed.
Same as the 911. While you couldn’t pay me to buy a Leica, my C4S cabriolet arrives this summer!!
As far as nice starbursts go, you don’t need a Leica for those. This pic was taken with an E-PL2 and Canon FD 50mm 1.4 at F16
http://farm8.staticflickr.com/7076/7251161962_4536331803_b.jpg
With my OM-D
I would not use jpg- I would shoot raw, and I would use an *excellent* lens.. my choice is the 35mm Carl Zeiss Jena Flectogon.. and I would shoot at F4.. and my shot would NOT be that washed out, unsharp sample you provided. Try the excellent 45mm Olympus prime, or the 75mm.. all available for a mere *fraction* of what the Leica gear costs.
Of course.. the Full Frame is *supposed* to be better, and the APS-C would be a fairer comparison.
OMG what are you lot all on! Im off to have fun with a crappy Polaroid and Ricoh FF-3 without a thought given to this endless stream of drivel regarding sensors and meaningless tests and stats!
There are writers that use a pen and paper, a Smith Corona, and those that use a word processor. At the end of the day, a great story is a great story. And photographers are supposed to be great story tellers. Anyone asking me what camera I used is probably not seeing my photographs. ” An M9, Leica and Carl Zeiss lens user.”
Who/what is dxo? ;-)) why would anybody care about it? now, seriously , whats the percentage of photos made at exposures values placed under ~6 or 5.
Still love mine, just shut up and shoot
It might be interesting to see them sized to the same horizontal resolution. a few more pixels may not change the overall look but the Leica image is posted at almost 1600px, the Canon approx 1350px, and the EM5 at approx. 1100 x 700. Maybe post a link to the original RAWs? I don’t have the new version of that Canon 24-70 and would like to take a close look at the pixels.
So many pissy and insecure Leica M9 owners Jesus…
Just keep the bloody camera and stop whining.
Does it hurt so much that DxO said M9 was worse than an OM-D?
No, because despite what DxO says about the M9 versus OM-D; the M9 renders its entire focal length and kicks the crap out of the OM-D IQ in the ISO’s we all shoot at 90% of the time.
Nothing, in my experience at least, equals the quality of a Leica M lens.
I think that’s inaccurate. I think it’s more accurate to say that when taken as an entire lens lineup nothing equals the quality of Leica.
Case by case, it’s often different, however. A number of Zeiss lenses manufactured in Germany are frequently touted as superior to Leica. And there are some examples of Canon and Nikkor lenses that are unmatched by anyone, anywhere.
Perhaps this is Captain Obvious speaking here, but I think it has to be emphasized that most manufacturers build lenses to a price point. Does Leica know more about engineering lenses in the 21st century than Canon or Nikon? Almost certainly not. But it’s doubtful that the latter two could justify (or sell) a $5000 50mm f/1.4 lens (though I’ve no doubt whatsoever they could build it — and that it would match the Leica for performance).
It’s about positioning. Leica has positioned itself as a manufacturer of high-quality, luxury cameras and lenses. No expense spared. Period.
Canon and Nikon service much larger and more varied markets. Both are manufacturers of high-quality professional cameras; prosumer cameras; amateur cameras…and so on. Subsequently, both are much, much larger companies than Leica as well (particularly Canon), and so must be dedicated to serving a broader base.
Interesting comparison. Just FYI, you would have gotten a more meaningful sample out of the OM-D if you had shot at f/4 rather than at f/8. Due to the shorter focal lengths used to get a FF-equivalent field of view, m4/3 lenses tend to perform best at wider apertures, and can be diffraction-limited at narrower apertures. The 45/1.8 at f/4 would have been a great comparison against the Leica 90/2 at f/8, and would have had shown similar depth of field.
Apologies if someone else has already pointed this out above. There are way too many comments.
“More than adequate” sensor is a pretty serious accusation for a $7.000 camera
Of course the M9 is good at low iso. This is consistent with the DXO measurements. The M9 isn’t bad. It’s just that latest (Sony) CMOS sensors are really good. Most obvious at high iso and dynamic range. It wouldn’t be fair to give them similar scores.
I have to laugh at the title of your article, ‘The M9 sensor is more than adequate’. You know….it SHOULD be more than adequate…just 6 months ago M9’s were being sold for $7,000 and even today a new ME is $5500…using a sensor that has been outdated for years.
With all that said…I love the M8/M9 and maybe I’m in the minority but I think high ISO capability is probably the most overrated feature on a camera.
I’ll tell you what an M8/M9/M9-P/ME/M digital rangefinder has….it’s mojo!! They’re fun to use and at the end of the day you can’t put a price on that!
Well said. I continue to enjoy my M9 despite its inferior sensor and laughable electronics.
Even the M8 beats the pants off of any current offerings from Nikon or Canon…. errrr well at base ISOs anyway.
You might want to try a Canon or Nikon if you haven’t shot one in a while. It’s hard to fault the 1D or 5D as it is the D800.
BTW: I didn’t mean to imply that the M8 wasn’t very capable… low ISO and all… but sometimes it’s easy to forget the Canon and Nikons are ‘systems’ as well. You can slap a 200 f/2 on a 1Ds and it can follow focus like none other. It would trash any Leica, any time if you were after fast moving subjects.
Both Canon and Nikon + Lens can often out-resolve most others if pure sharpness is your thing. High ISO? yep.
The M8 + Leica lens is obviously distinct and great in it’s own way but I don’t think anyone could claim it beats the pants off current high-end Canon / Nikon cameras. You might have to sell you pants to buy it though…
Base ISO or otherwise, this is simply a wildly inaccurate statement.
I was already expecting top quality from the Leica or the Canon, but the EM-5 just blew me away. It’s the worst alright, but the IQ on that camera is AMAZING! More than enough for 99% percent of casual users ( a completely random number I pulled outta my butt) I’d bet. At f4 it would be even better. It is really an amazing camera, best smallish camera for the price BY FAR, IMHO.
I think Leica are the best digital (and only) rangefinders of the world. And the best analogue RFs. Their lenses are exceptional. Once said that, I don’t see in the pictures above any advantage for Leica. In fact, Olympus and Canon seem better in some respects. As RFs, digital Leicas are the best; regarding IQ, the thing is not so clear at all.
Hey Steve. It is great to be able to post comments on your site. However would it be possible to make the posts editable (in case you make a mistake or want to alter the comment). DP Review give a 15 min window to edit your post.
Thanks and no need to post articles justifying your passion. Just ignore the negative, sad people.
I don’t give a monkeys if you are an official Leica guy or not. Like most I enjoy the daily posts, the comments and your enthusiastic reviews. Keep it up man! Go well!
adequate is the word
Not an equal test but clearly the Leica shot is better. However taking into account the price of the body and lenses you would hope so. I would have expected the 5DIII to be similar (needs a prime lens). OM-D E-5 not in the same class as a FF camera (despite it’s good IQ). DXO scores don’t tell all the story at all. To use a car or bike analogy two engines may produce similar horsepower but one vehicle may perform much better due to how that power is made and how it is put to the ground. eg a 600cc sport bike puts out about 120bhp and a 900 sport tourer about the same. However they have very different characteristics. The 600cc sportbike develops its power high up in the rev ranges and needs to be kept on the boil unlike the torquer and easier 900cc sport tourer. Cars with modern 8 speed autos and down sized engines can be quicker than the higher horsepowered cars they replace (eg Bently’s new V8 vs old V12). So it is aways the whole system. WIth a camera it is mainly sensor and glass but these days processing is very important (both in camera and out) and how that camera handles can be the determining factor as to whether it is purchased or not.
Point taken.
But the shots on the Canon body should also have been taken with the same Leica lens.
For those who have not used an M9, try one, they can be found (used) at good prices. Once tried, you will realize that not only is the sensor more than adequate but superb. There is not a single metric I have seen or heard that can account for the gorgeous images it produces. DxO garners much attention but is the last test site I would ever consult in a purchase decision, no disrespect for Dxo but I’d rather look at images and make buying decisions. How often we see photos made from the lowest rated equipment that are beautiful and prove the point that equipment alone doesn’t make the photograph.
Ummm… some may not consider $4500 a ‘reasonable’ price to pay to try an M9.
Presumably you could re-sell it and get your money back if you’re not happy – but to many that’s quite an outlay of cash (and may, in some cases, even be grounds for divorce) 🙂
Maybe you can rent one at a reasonable price?
Low ISO the M9 is still king to me, there is a clarity and smoothness that I have not yet seen playing around with my M 240 DNGs. Alot of the increased dynamic range seems to have been at a cost to contrast and pop, and it renders too warm for my taste. But the 240 can still play with a tripod and slow glass once the lights dim.
However the 24mm Leica Summilux for me negates a great deal of that advantage, as I can hardly detect noise in the M9 below ISO 800. Aggressive OIS is a must for video, so once again I am left with a larger LCD, better battery life and some added functionality. I guess the one thing i truly do not miss about the M9 is its tendency for moire.
This comparison proves that there is really no need to blow $10,000 on a Leica body and lens when $1000 is more than adequate. Leica people somehow feel their camera system is superior when the results clearly show that it’s not.
Here’s an idea. Buy a really great camera and lens system for under $2000 and take the remaining $8000 and travel to amazing places and actually SHOOT AMAZING PHOTOS instead of wandering around your backyard testing your $10,000 system on your pets 🙂
Amen….
sorry, but you are wrong, comparing 100% and 125 % view is something that is not ok, and to say that f stop of 8 is the sharpest for every lense is ablosute not true, specially for mft lenses. two stops down, ok, but with mft and f stop 8 you are off!
Yeps, see other reponses de moi,
BTW, Eike…..much more inportant…..great shots on your website, dynamic, cool color and the rest…
Greets, Ed.
thanks a lot 🙂
…and the fanboys rush to justify their cash outlay.
Landscapes at night at low ISO from a tripod can be easily done by just about any camera. That’s an easy test. It’s when you look at high ISO and challenging real world situations that the limitations of the M9 sensor start to show themselves.
It’s a gorgeous camera capable of fantastic quality images, no doubt, but within some narrow parameters.
Nice article! I pretty much agree with George Sutton’s statement that, “single most important factor in a camera’s quality is the lenses”. But I would add that in my opinion, the single most important factor in the quality of a photograph is the talent and skill of the photographer.
If a photographer in 2007 took a great image with a Leica M8 with a 50 year old lens, I can assure you that the same great photo could technically be captured today.
The DXO tests have their use as it gives us a comparative list of how certain sensors stack up against each other within a specific set of parameters. But one must always remember that laboratory tests of sensors is one thing, and the taking of good images is another.
I don’t buy into the “single most important…” statement. Lens, agronomics, shooter skill, sensor or film are all important. Combine a great lens with subpar film and the photo will suffer. The opposite is true. Put it another way….. A car’s engine and acceleration is only as good as its brakes….
I happily make money shooting with my M9 and lenses, and try to shoot with the Leica as much as possible versus using my DSIII. I have to use the Canon when I need to tether, or if I need a long lens. That said, I just purchased a Sony RX100 as a small carry-camera and I will say IMO the noise level of the raw files at 800 and 1600 asa coming out of this $650 camera is cleaner than the M9. The M9 files are way sharper, the images are contrastier with better color, but the noise. Some paying clients don’t like noise. I’ll continue to happily use my M9, but bring on the M240. And thanks Steve for your continuing “M-analysis”!
…the best camera in the world is the one in your hand…as we all know
You can’t measure the quality of a Leica lens at F8 or F11. The strength of Leica lenses is that they are unbeatable at wide aperture. At F8, any old glass holds it….
I own an M9 and a Summicron 28, and at F2, it is just superb, magic, no japanese equipment can match that.
I think yo will find Titch that there are several Canon, Nikon and Zeiss lenses (primes) that will match their Leica counterparts at f2.0. They will render in a different way, they will not be less, nor better.
The 2.0/28 Distagon is one good example, and so are the 1.4/24 and 1.4/35 Nikkors.
+1
Absolutely true.
Agreed with the Zeiss, Canon comments… but don’t forget about those great Olympus FT’s! That 150mm f/2 is nice. Schneider, etc. many have made world class lenses. One of the lens rental companies posted a review of a specialized prime (EF mount, I forget the manufacture) that was basically the reference lens for all primes at that mm and was mostly used for evidence gathering, etc.
Sometimes it’s easy to want to think of your favorite sports team (or company) as the best… especially when they are one of the best!… but we often forget there are some other really great lenses out there that often outclass the more popular high-end lenses. Of course, at that level, it might be hard to tell without specialized equipment.
I try not to go the way of “the stuff I have blows all others out of the water” etc etc., even though I might think that it’s very good. The flaws in my own technique have far more impact than any in my gear.
Ps: I was referring to full frame lenses only. The upcoming new very expensive Zeiss lenses should be interesting.
Yea… I’m kind of looking forward to the Zeiss X-Mount lenses myself. Hopefully they bring something interesting to the market.
I have a 21/2.8 in EF mount that I would love to have an equivalent in X although that 14mm Fuji X looks equally great. I fear the problem with Zeiss and X mount (or other non-M) is that if it’s manual focus anyway, it’s hard not to just buy the Leica version and use an adapter. Of course they have auto-Aperture with EF, but if I had to do it again, I would probably buy the Nikon version and use an EF adapter as I prefer the a-ring anyway.
Ziess does sometimes have a significant cost savings over Leica if sharpness is you main need. (the ZM 50 f/2 is very sharp and less thank $1k, the f/1.5 a classic and still around $1k).
Yep. The flaws in your technique will show up long before any perceptible sharpness or resolution differences can be discerned between Leica glass and quality Canon/Nikon/Olympus/Fuji glass.
It’s a bit like the motorbikes (MV Agusta) I used to ride on trackdays. I enjoyed myself (and crashed a few times), but the capabilities of those bikes far outweigehd my own.
That could be just me of course… 😉
Some beautiful shots George – I love night photography!
One thing that stands out to me in these photos is, at the base of the buildings (where the images transition to black) it looks like more shadow detail is lost in the M9 shot, compared to the Canon shot.
But I’m not sure if that’s sensor related (perhaps it’s more overall exposure, as the M9 image does look darker than the rest). The M9 image is nicely sharp.
I guess overall I agree with your conclusion – the images are all quite good! 🙂
Steve, you totally miss the point. DxO tests cameras, not lenses. I can put almost any lens on the OM-D, including a Leica lens that gives me the frame you took. DxO does not test camera systems, it tests a camera body’s ability to record image quality on its sensor.
Had you used the same lens on each camera, you would have come to similar conclusions as DxO. And these conclusions show that the Leica sensor may be full frame (for what’s that worth today) but performs on the level of the OM-D sensor (as measured by DxO) which is at a quarter of the size effectively. Additionally, you chose to compare a fixed focal length lens with zoom lenses. You could have picked a prime lens for the other cameras too.
The single golden statement, that it’s the lens that defines the picture, used to be true 100% in the good old golden (rather silver… haha) days of film, mainly for the fact that anybody could put any film in their cameras and every “sensor” so to speak was the same. The only way to differ was through the lenses. Today, that’s hardly true. The sensor as a factor in image quality is just as important (if not more important) as choosing a good lens. Also, and this adds to this, today almost all manufacturers produce very good lenses with differences getting so small it’s hard to see. Take Cosina/Voigtländer for example. Their M lenses at a fraction of the cost of Leica glass offer almost the same quality in the image and the build.
Leica recognises this in their sales strategy. For them, it gets harder to compete on the merit of technical excellence only, so they are branching into markets where they sell on exclusivity in limited editions and branded items which are more fashion and luxury than cameras for real photographers. This is where Leica’s margin is.
As for Leica fans (and fanboys), they need to start recognising that other manufacturers have upped their game in the last years beyond what Leica offers at their current price point. This doesn’t make Leica bad per se in absolute terms, but in comparison (and DxO is all about comparison) and in relation to price, Leica digital cameras are not only expensive, they are a total waste of money.
To be clear about this, I like Leica. I would love to buy an M6 or a MP, maybe even a M3. I own a CLE which is cute little camera. I would never buy a digital Leica, mostly based on the fact that I can get equal digital image quality based on lenses and camera bodies available at a fraction of the cost of a Leica system while having the same amount of fun using it (probably more, let’s face it, manual focus using a rangefinder is mostly limiting to most people – I know, I use my CLE a lot but prefer my OM-D).
What’s a bit disappointing is the way you try to fight a battle you can’t win. You add more apples and oranges comparisons by talking about the lenses and then not even use the same lens types across the systems, ignoring the main issue that the comparison is about the cameras, not the lenses and the same lenses can shared across different cameras. In my eyes, you are damaging your reputation as a trusted source for reviews as you fail to see the obvious.
Tobias, First, Steve did not write this so don’t blame him. I think you miss my points here. I have taken three cameras meaning body and lens combined and shot the same image to see how the shots compared. That is how I would decide which camera to use to take a real photograph. What I wanted to see was the relevance of the laboratory tests to real world use. I used the best body/lens combination I had available for each particular camera. I have other Canon primes including the 85 f1.8 and the 100 f2.8 macro and I have found the new 24-70 zoom as good as the primes. The one possible exception is whether the E-M5 would do better with an Olympus prime but the Lumix 14-35 is very highly rated and should be very close to any prime. I do have a Lumix 20mm f1.7 and I see no difference between it and the Lumix zoom at that magnification. I also need to retry the E-M5 at a faster aperture to see if shooting at the same depth of field makes a difference as others have suggested (and I appreciate that suggestion–that is how reading these comments can be helpful). The point is that I found the relevance of the DxO test surprisingly weak with regard to selecting a camera for use. I have actually sold another photograph of this night scene and, as I mentioned, if I were asked to take another shot to sell I would shoot both the Leica and the Canon and pick the best. In the end, the buyer doesn’t know or care which camera was used, they just want to be wowed by the photo. I completely disagree with your statement that any manufacturer makes lenses as good as Leica. I don’t need to use Leica and I would sell it immediately if another camera would do as well in the way I use it. I have had the opportunity to compare it with the best cameras of other formats and, as I said above, I have never seen the same pop or level of detail in any other camera. It doesn’t happen every time and the Leica is useless for many situations, but it produces a winner often enough to earn its keep. That is partly the lenses and I am guessing partly the absence of an AA filter. The Canon also does very well and is much more versatile. I get smooth highly detailed images with rich colors and I have sold several, but I have always been frustrated with soft corners in Canon’s wide angle lenses and as I age the size and weight becomes more of a nuisance to haul on a trip. I fully agree with the criticism that the Leica is overpriced and at its price should include the same features Olympus and Canon deliver at much lower cost like sensor cleaning and IS (I really like the IS in the E-M5 — I think it produces clearer photos in most instances). But the fact remains, in my experience, the Leica produces really good photos and within its useful zone I have not found any camera that will take a better photo. I don’t own it to strut around with it, as you seem to think, in fact the only people who have ever commented to me about the M9 wondered why I was still shooting an old film camera and one other asked if it was “one of those new little Fuji cameras”. As the photographic equivalent of a fur coat or big diamond ring, it fails miserably.
Tobias: “DxO tests cameras, not lenses.” That’s wrong. Check here:
http://www.dxomark.com/index.php/Lenses/Camera-Lens-Ratings
They only did not yet review any Leica glass.
Look at the KeyBank sign and you will see that the Leica kicks it. The Canon and E-M5 were similar.
Hi George, first of all stunning last shot. I feel your correct about the difference when printing relatively small and at this point in time A4 is probably where most people are printing. Is it me or has the Canon got a bit more DR in the shadow areas and is generally a tad brighter. Thanks for your post and yes I agree the M9 has plenty of mileage yet!
Hi Peter, I would not draw any conclusions from these shots about DR except they are all good. I tried to equalize the exposure of all three and sharpened them before converting to jpegs. I also adjusted the shadows for the Canon but not the Leica. Both the Leica and Canon have more latitude in the shadows if you wanted to draw them out. I am really impressed with the DR in the E-M5 for a smaller camera but once again, though quite good, the shadow detail in the micro 4/3 is not as good as the full frame sensors.
Eh, DR as measure by DxO (and that is measured on base iso and same shutter time,s o relevant)……
OM-D 12.4
Leica M9 11.7
Leica M240 13
Canon 11.7
Eh the rest of the measurements for the M240 anyone:
Low light range 1860 (commemerating the Münich soccer team)
Color depth 24 bits
Total score 84…..nice.
Greets, Ed
Thank you George, we are certainly not spoilt for choice. For what is still a relatively young technology the pace of improvement is staggering
I’d recommend M9 users just to relax about the DxO verdict. If you just love to pick your camera and a beautiful lens, feel it in your hand and walk out into real life to shoot with passion – that’s what really counts IMO. George’s post shows again that not only camera technology counts. I still use a Canon DSLR (5D3, need its great AF system for wildlife tele shooting, its mediocre DxO verdict doesn’t bother me), and I have two comments to this post:
(1) Canon now has some extremely sharp lenses in its line…
(2) one of my fav lenses is the opposite of what sharpness freaks really want: Canon’s legendary 85/1.2. I use it of course for portrait, but even sometimes for street:
http://www.dpreview.com/galleries/6385346624/photos/2464553/frankfurt-hopper-style
Unfortunately I really cannot afford Leica’s Noctilux lenses. So I think I’ll stick with Canon (if only because of the 85/1.2) for a while…
So I wish all you lucky Leica M9 owners: don’t let your passion been torn down by any lab reviews. And Steve, keep your great site going!
They can’t relax. Haven’t you see the price of the M9?
Well, I know why I still (have to) stick with a more cheapo brand, even, as a German, I am quite happy about Leica’s economic recovery as a legend… But looking at Leica price policy I really ask myself if they aim more and more for rich gadget collectors, and less for real photographers.
“Hopper style”: great image!
Michiel, thanks for this nice compliment 🙂
M9 and Canon are undoubtedly great cameras, but I think this comparison is a bit unfair for the Olympus E-M5 because due to smaller sensor size the diffraction starts to reduce the image quality at f8. Therefore if direct comparisons between cameras with different sensor sizes are made, this should be noted. The m43 lenses are typically sharpest around f4, which also corresponds to f8 in full frame system at equivalent focal length, in terms of DOF. Also enlarging the Olympus image by 125% of course somewhat reduces the image quality .
This is just a side note, but I think it should be noticed if direct image quality comparisons between systems are made.
(Btw, it is bit inaccurate to refer Olympus E-M5 as ‘OM-D’, as seems to be the common habit these day in many websites. It’s the same thing than to refer Canon 5D mark III as Canon EOS. OM-D is a group name for a line of cameras, not a name for an individual model. Although at the moment there is only one model in the OM-D line 😉
So what you’re saying is, comparing same size images, a smaller sensor produces less quality.
Which, of course, is true. Hail full frame.
No, what I’m saying is that different systems produce best quality at different settings due their nature (sensor size here). I’m not claiming that E-M5 would produce better quality images than the full frame cameras here, but that it would perform relatively better if the diffraction effect would have been compensated by choosing a wider aperture with m43 than with full frame. And if similar crop is used for image quality comparison, then the better choice is to compress larger images than to stretch the smaller images to avoid “artificial” pixels in the pictures.
I’m owning a full frame camera and a m43 camera, so this has nothing to do with personal preferences, but just about testing methods here.
It’s also a bit unfair on the EM5 to use that Panasonic zoom, which isn’t a bad lens but certainly not the best that could have been used in that situation. To equate to the Leica’s 90mm, the Olympus 45mm (set at f4, of course), would have been a close match. To equate to the Canon zoom, the 4/3 14-35/f2 zoom (again set at f4) would have been a close match but the 4/3 12-60/f2.8-4 zoom would have been good too. Sorry, George, the Lumix zoom is not regarded as the best 4/3 zoom – though I’m sure you meant micro 4/3!
Anyhow, it’s a good comparison and shows that Leica do know what they’re doing.
Discussions about the merits of various testing or comparison methods can fill entire libraries.
I personally find the “enlarge to same size, compare” method useful. More so than downsampling, upsampling or whatever to compensate for differences in sensor size or megapixel count, because it comes closest to what we do: produce images of a certain size, not dictated by the size of the sensor in question.
How much you can “blow up” an image (to quote Michelangelo Antonioni) is depending on a few things (one of which is the sensor pixel count). I have seen unsharp images of a D800e (backfocus damn you Nikon) that could produce nice prints at 8 x 10 inch but not much bigger. I have seen Ellen von Unwerth photo’s shot with a Canon 5D (I saw a video of the actual shoot) and blown up to 180 x 120 cm. And a freind of mine made 70 x 70 inch shots using a Panasonic GF1….(Gianni Galassi), and the graphic nature of his shots lets him get away with it (as does the blazing mediteranian light he uses in his shots).
So the simple equation, more Mpixels is bigger print size does not work…in this case physics can be beaten by computers skils. If the image is there a 16 Mpixel micro 4/3 can perform just as well as a 16 Mpixel Nikon D4 or a 36 Mpixel Nikon D800 or a Leica M240 (if the image is not there). But the image has to be there. Personally I would not buy a D800e since I think for FF 24 Mpixels is the sweet spot….as for Micro 4/3 it is 16 Mpixels….more would be possible (sacrificing low light performance, but I wouldn’t care since the sun shines for free :-)) but that would mean I’m not able to stop down sufficiently to have enough DOF (I like DOF as a landscape shooter).
The physical size of the sensor however does not limit the way you can enlarge you’re shots (no matter what some people like to tell you) only the pixel count does (if you are shooting not difraction limited of course) and whether your lens is sharp enough to outresolve the sensor (Leica shooters have no worries in this department) and is correctly focussed (more of an issue to Leica shooters if they are a bit forgetfull (personal experience) and are accustomed to AF Nikons).
Greets, Ed.
Thanks Ed, very helpful. I’ll stick with the 12Mp of my lowly and obsolete D700 then.
Comparing an OM-D and Leica M9 or any FF camera at the same aperture is a test protocol error…..
A professional tester (10 years experience under my belt) uses so called “use cases”….a use case is a real life situation which your system has to solve.
Now any micro 4/3 has a DOF 2 as large as a FF camera. So in order to take the same picture (and a photographic image is gouverned by 2 things being shutter time and aperture) a micro 4/3 system can use an aperture twice as large a FF camera. And in my book I take pictures I don’t stare at the sensor.
Having established that we can then conclude that in order to have the same DOF as my OM-D the Leica (or any other FF camera) has to be stopped down 2 stops in order to get the same DOF……so my OM-D 2.8 compares to a Canon’s (or any other FF camera) 5.6….now we will have to use the same shutter time since else our FF freinds would come home with montion blurred pictures (be it from the subject or the fotographer) so lets asume a 1/250 of a second at f 2.8 for the OM-D (or any other micro 4/3 camera) and lets asume 1/250 of a second at f 5.6 for the FF…..now what happens to the ISO…..well in order to compensate the 2 stops the FF will have to double the ISO twice…..so ISO 800 for the OM-D is ISO 3200 for the FF and then the world looks a lot more equal doesn’t it.
So lets nicely forget all that, small sensor, big sensor hassle…..it is not the sensor that takes a picture it is the camera and the photographer. If I want bokeh I’ll shoot a FF for the rest, I can asure you, the difference isn’t that big in capable hands and the capability of those hands is much more decisive in any case.
Greets, Ed.
Hi Ed! I would think (not having any testing experience myself) thatbtesting procedures are governed by what they are supposed to establish. What you describe is a procedure where DoF is equalized; what I described was general image quality, looking at same size images from the same distance.
But I agree, in the end tests are certainly not the decisive factor. They’ve never changed my opinion on any of the equipment I used; they usually just confirmed what I already knew.
Yeps Michiel,
That is the point…..you can’t do a general image quality test if you don’t equalise the image…and that can only be done by equalising DOF and shutter time. Now don’t be offended Michiel…..even the lads at DxO don’t get this right and sensor wise you are 100 in the right. But I don’t shoot with a sensor (neither do the lads at DxO by the way, either they use a lens or project an image on the sensor), I shoot with a lens and that has an apperture and I know full well that in order to get the same image (DOF and shuttertime) I have to stop down the aperture of the FF Nikon F3 to 5.6 in order to have the same DOF as the OM-D……
So you can’t compare image quality if you don’t make the same image as JVV has already said so well. In the above test the DOF in the OM-D image is twice as large as the DOF in the Canon and Leica image. No wonder then that the resulting image quality suffers. If the OM-D was set to aperture 4 it would have given the same image…..and ISO or shuttertime could have been reduced and both would have significantly improved the performance of the system.
Greets, Ed.
I don’t get your point. The depth of field is irrelevant in all of these shots. I am a few miles from this spot so everything will be in focus regardless of the camera and aperture. I picked f8 to get the star effect and show that difference (I love the effect with a Leica lens dislike it with most other lenses) and also because it is typically an optimum aperture for any lens. If I have to crop and enlarge a shot there will be a difference in resolution and clarity between the larger and smaller sensors. There is no way changing the aperture will eliminate that difference. Your point is understandable if depth of field and bokeh are important qualities of the shot but that still has nothing to do with the sensor or these photos.
That’s actually what I wanted to say too. With infinity shots you’ve got the distracting DoF/aperture thing out of the way, and can concentrate on “general image quality” comparison, which in this case would result in the observations you gave George.
No you don’t Michiel…….it’s an issue of the micro 4/3 versus FF test. And that is true at any focal length etc. It plain an simple optics….physics, not art. A true honnest comparisan of both systems can only be done at different apertures….
Greets, Ed.
“I don’t get your point”
Not to be helped George I have to admit that the fact that DOF is irralevent to your photo is true, but that does not make it irralevent to your test. Physics are hard to master George and even harder to work around and physics of optics is one of the hardest (engraved in stone so to speak).
At aperture 8 the OM-D has twice the DOF of your FF bodies, you do agree about that right (else you have to study some optics which I don’t think you have to do). Now wether that is relevant for the shot, I care the proverbial rodents derriere about, the fact is that it does (now as I told you even DxO does not get this so please don’t see this as an insult) influence shutter time (and of course DOF). The fact of the matter is to give the same image as a FF at 8 the OM-D has to be stopped down to 4. Now you shot at base ISO (well done) so you could have cut the shutter time by 4. Now chroma noise is an issue with any sensor at longer shutter speeds (also simple signal to noise physics). Now I don’t have all these nice camera’s lying around, so I wonder what youre test would have looked like when you would have taken my objections/additions into account…..
By the way I fully agree with the fact that an M9 sensor is more then up to the job if used by a skilled photographer (as a 12.000 dollar camera and lens should be). But that photographer has to work a little harder in order to get the same results as a Canon or a Nikon shooter that can up the ISO more easily and freely.
Greets, Ed
Amen Ed, good job explaining the reality.
Another test error is to compare primes (the excellent Leica 90 mm) against zooms at night. A typical zoom contains about twice as much groups and elements as a fixed lens (the Elmarit 90 contains 4 elements in 4 groups according to Ken (the Ken)). The Pana for instance contains 14 elements in 9 groups. Now the more elements and the more groups a lens contains the more chances there are for internal reflections and so coma etc. occurs more easely….
Now the 45 Olympus contains 9 elements in 8 groups (the Pana/Leice 14 elements in 9 groups)…..so the Olympus should be the one to use in this test. Not completely fair but okay…..since no other lens exists for the system with fiewer elements.
And for the Canon, well there is a 1.2 85 that could be used……since well price is no option right but the 1.8 would also be better siuted for this kind of test.
Now if you want to reestablish the Leica M9 sensor…..well I can asure you I will sell my OM-D for any M9. Not because of the sensor, but because of the system. Leica glass is brilliant and completely in tune with their sensor……and to me the MM is the best Leica yet….(even beating the M240), to me it is a stroke of pure genious…..ridiculed at first (also by me) but boy thus that camera make sense.
Greets, Ed.
Interesting, thanks for the explanation. I completely agree on the MM.
Ed, I appreciate your patience in explaining these points. I particularly agree about the effect of testing with different lenses. I have done these tests with primes such as the Canon 10mm f2.8 macro and the big surprise was how good the new 24-70 zoom performs. Anyway, this test did compare apples and oranges, and maybe even fish and bicycles, and is invalid as a pure comparison for all the reasons you and others have mentioned. My idea was not to get a pure test of the sensor but to test which camera would work best or be good enough if asked to take that shot from that place in those conditions. I also wanted to see how close the E-M5 was getting to FF image quality and up to a point I found it close and much better than other 4/3 cameras I have tried. A fourth photo I did not include was taken with a Lumix GH1 I have used for several years. It was a mess at this enlargement. Anyway, thanks for the feedback.
Ah, GH1 and an OM-D are totally different camera’s, no comparison there….but the OM-D/GH3 generation has closed the gap to FF sensor type camera’s considerably (I just jumped from a GF1 to an OM-D and have to say, so that’s what clouds look :-)) sometimes overtaking DX camera’s.
I agree that in LL situations where you have to shoot wide open I would like to have a FF camera when given the choice but in most surcomstances the OM-D delivers and somehow even manages to amaze (just shot a night shot with a Nikon manual lens at 5.6 and 3200 iso and 1/6 of second and it is sharp…..usable even for larger prints and I shot in by hand using IBIS).
George if you had shot the OM-D at 4 (instead of at 8) you would have found it even more close to the call…..I have compared base ISO shots (and 99% of my shots are base ISO) of my OM-D with base ISO shots from a D4 (both are 16 Mpixel machines) and I could not tell which was which (pixelpeeping till my eyes hurt). Try it yourself at the Imaging Resource comparometer ™…..
Inch for inch I think Olympus has made an amazing camera and value for money wise and weight for quality wise. And both count heavely in my book.
“My idea was not to get a pure test of the sensor but to test which camera would work best or be good enough if asked to take that shot from that place in those conditions.”
I applaud that George but by not allowing for the difference in DOF you more or less unwillingly did just that (as all the rest of the testers do), test the sensor and not the system. Maybe I’m a bit more sensitised to DOF issues because I use medium and large format (analog) camera’s as well, and for instance using a 6×9 you rarely use apertures bigger then 16 and 22 or 32 are common enough. But with my OM-D 5.6 or 8 is as far as stop down (the limit so to speak) and using a kitt lens at 5.6 max at 45mm that means not much room to play with (one real disadvantage of the micro 4/3 standard).
Greets, Ed.
Ed, I tried shooting this again with comparable DOF and primes, as you suggested. You may be correct about the comparable DOF effect even when it is not a focus issue. I shot all three cameras and this time the shots were essentially equal in quality. I only hesitate to say you were right because I got a better shot even at smaller apertures on the E-M5. The absence of haze may be a factor. Regardless, the Leica and Canon were both very good and the E-M5 was just as good, which is very impressive. The only advantage was to the Leica because of its great star effect. The Canon, BTW, was excellent with the 100mm IS macro but the 85 f1.8 was not as sharp as the other lenses including the Lumix 14-35 zoom. So many variables, so little time.
Good points. I will refer to the Olympus as the E-M5. I still put the blame on Olympus for coming up with such an awkward name to start with.
These are an unfair comparison for the E-M5 in some respects. It is a smaller sensor and is enlarged more than the full frame cameras in the photos above. I decided to do that because to me the bottom line is the photos a camera produces and what struck me in doing these shots was the fact that the size of the sensor still matters, perhaps not as much as in the past but it is still noticeable.
George and Ed,
There is more valuable information in this comment thread than in most articles about comparing sensors/systems. 🙂 You are both very intelligent and I appreciate you sharing your knowledge so publicly and clearly. It was quite a fulfilling conclusion to hear that all three images were almost equal when tested again. It’s sad to think how many people might pass up this great info because they don’t read the comments haha.
The photography community thrives on great conversations like this one, and what an appropriate site for it to be hosted. So thank you, Steve, for the environment you’ve created here, and thank you, George and Ed, for sharing.
I think the moral of the story is, if you like what the M9 does, keep it. Whatever DxO says is irrelevant if you like the results the camera gives you. That’s not to say that DxO’s numbers don’t refer to real things. They do. But at the end of the day it doesn’t matter if you like the pictures you’re getting.
As a Leica M9-P and M Monochrom owner / user, I too read the original DxO report with some interest. I took the time to go to their website to try to understand their methodology. What I found most interesting is that they test lenses as well. I checked to see what they had to say about Leica Lenses. They had nothing to say because they had not tested any. As George correctly points out and illustrates, it is the end result, the photograph, made by a photographer with a (camera) sensor and a lens, that is the real determinant. Leica lenses are unequaled in 35mm photography. Further, the equipment is just a tool. It is no different than an artist’s pallet, brushes and canvas; in the right hands magic can happen, in the wrong hands, just another snap shot. So, to paraphrase a line from the old movie, Risky Business, “Leica, There is No Substitute!”
“Leica, There is no substitute”
According to the quality of shots above, there indeed appears to be several substitutes.
Actually no they aren’t digital rangefinders which is the main reason to shoot Leica otherwise get a cheap FUJI or NEX
The word “cheap” makes your whole posting worthless imo. This is not a Forbes magazine of camera owners, this is about photography and tools.
Hmmm, and here I thought it was all about the results.
DxO…and i quote
“DxO is not measuring sensor performance directly. It is measuring raw files constructed by the manufacturer. It then measures some quantitative characteristics that purportedly are proxies for image quality, which are tenuous at best. It then provides an arbitrarily weighted overall score. That score has little to do with the sensor itself notwithstanding its promotion of it as a “sensor score”.
Ultimately, it is a score on how a manufacturer constructs its raw file. Therefore, NR applied across the whole range of iso will produce a very good noise score. This is irrespective of data integrity or whether there are false detail. That is why there are wide ranging scores for the exact same sensor. Acuity, micro contrast, data integrity, detail, resolution are deemed irrelevant for IQ according to DxO.
Because of DxO’s constant self promotion, idiots out there think the IQ from a D800 is better than IQ from Phase One and Hassie.”
So don’t let DxO drag you in or down…means nothing in “Real World Photography”.
@Bob: not better, but close (D800E) to MF, and more or less equal to S2, using of course the right glass. Better fine tune your focusing and exposure skills, and get those murals printed!
“If I were shooting this for sale” – that sums it all. To me even OM-D looks fine here.
I would love to buy Leica and Notilux to be honest, but it is more like an expensive watch for a enthusiast like me. I might just settle with XE-1, just waiting for Zeiss X mount lenses to be released.
I was never worried about the sensor quality (cracking apart) of the M9. I’ve seen too many good and very good images made with M9’s.
Reports like DxO’s, whether they produce excellent results or less than excellent results should always be viewed in connection with the images (yes Steve, real world ones of course).
I find Lloyd Chambers’ reports very useful, but also only as a “companion” to what you really see in an image.
You illustrate what’s often overseen: it’s the whole system that counts. The M9 really rocks with the Peter Karbe designs!
Interesting comparison…wonder how the RX1 would do…
It would do fine but the RX1’s real magic is from f2-4ish not f11. It loves to be shot wide open and is a shallow DOF rock star 🙂