Lightroom 3 Beta and the Leica M9
I’ve been testing Lightroom 3 for a couple of days, and I confess myself extremely impressed. The implementation of chroma noise processing is worlds better.
Case in point, here is an M9 image from one of the review sites taken at ISO 2500:
The white rectangle denotes a shadow area that I used for testing. Below is a comparison of the LR 2.5 and LR 3 renderings with the same settings for noise processing / sharpening. Other settings are their defaults so perhaps not the same but this is what you get when you open the file.
This is 100% pixels:
Remember: this is the same DNG file. The one that’s _vastly_ better is the Lightroom 3 image, I don’t think I need to say which is which.
Seriously this software just gave the M9 that little extra oompf I wanted to see. I think if that’s what ISO 2500 shadows look like then it’s fine. No, no fine, great! I’ll take this any time.
What can the D700 do with this software wonder? Well ISO 6400 is more usable for one thing. But the lower high ISOs (heh) don’t improve as much honestly.
Cheers.
Nikkor 24-70mm f/2.8G AF-S ED Review
This past July I photographed a friend’s wedding exclusively with my favorite manual-focus prime lenses. The results were as good as everyone had hoped, but it got tedious for me after few hours at such high intensity. After reflecting on the subject I decided that next time I would rent a top-grade normal zoom and relax. Not long after that I was invited to work at GTC to supplement staff photographers – I accepted and immediately placed an order for a rental lens.
The 24-70mm lens is at the center of professional outfits from Nikon, Canon and Sony. It is not as popular among amateurs as the 70-200mm lens, or as exotic in looks and images as the 14-24mm lens, but it is the one that holds the product line together. Contrary to its siblings it is entirely reasonable for one to use only this lens and still be a whole photographer/person.
This review pertains to the latest version of the lens from Nikon that launched the same day as the historic D3 camera. The optical design shows very little resemblance to its predecessor after the first three elements. The number of elements and groups is the same however, and in this incarnation the surface most prone to flare (presumably) has been coated with Nikon’s “Nano” coating. In an interview on Nikon’s PR website their engineers relate this technology to T* in spirit, more on this later.
Operation
At first blush the 24-70mm looks and feels like a shorter version of the 70-200mm VR. There are some differences: this lens changes dimensions while zooming, and the focus ring is a lot shorter for example. The presentation is the same however, in that this is a solidly built lens but not at all a light or compact one – I find it too long for taste anyway. Balance on the D700 is close to ideal but the combined weight is not, even though this lens has ample amounts of plastics in it (sigh). I can use this combination comfortably and nimbly for hours, but I get punished the next day with sore arms.
This lens doesn’t have VR but I think it should. Its raison-d’être is photojournalism, which is to capture people in their usually dark and cave-like contexts. For this task a maximum aperture of f/2.8 without VR barely meets minimum requirements. Of course this lens launched on the same day as the D3 camera and I can see how Nikon thought that it gave them a free pass.
The zooming action on the 24-70mm is cleverly designed to work with its deep flower-shaped hood but it is not always so smooth in action. The lens extends significantly while zooming at the wide end so that the lens can “peek out” and avoid hard vignetting. This extension combined with the weight of the moving train means that the force required to zoom varies a lot with the angle to Earth’s gravity. When perfectly horizontal the motion is very fluid, but aim at an angle and your zooming fingers are holding the weight of the lens.
As with most lenses these days, manual focus is an option but a half-baked implementation relegates that to a backup purpose only. The focus ring is too light, doesn’t stop at infinity and there is no depth-of-field scale to support range focusing (which is an real option at 24mm). At least with this lens you can expect some depth of field to hide errors. Auto-focus is snappy but otherwise unremarkable.
If it’s not clear already, I was quite disappointed with the physical implementation of this lens.
Drawing
The 24-70mm produces sharp and contrasty images under basically all conditions. This is the short and simple summary of its drawing properties. I was not disappointed here.
Let me clearly answer some obvious questions now: no it is not as sharp as modern prime lenses, and no it is not as contrasty as Zeiss T* lenses. It is quite simply a very good lens that overcomes an important handicap in its construction – that it is a zoom lens. I might even say that while it is not as good as my favorite lenses, it is basically “good enough” for everyone (me included).
Overall the drawing at f/2.8 is a bit soft, particularly in the corners, yielding good but not great results. At this aperture fringing is evident in many pictures where focus wasn’t spot-on and a few where it actually was. At f/4 the lens improves considerably and by f/5.6 the remaining aberrations are no longer field relevant. Detail reaches almost fully to the far corner at the smaller apertures, with only the last few pixels of the frame taking on a mild blur on FX sensors. I shot this lens almost exclusively at f/5.6 – f/8 for my own personal work and was thrilled with the results.
Bokeh is good, being almost neutral at most settings of the lens. I was quite pleased by this. Chromatic aberrations are under control, and fringing artefacts aren’t a problem except at full aperture. In my limited usage the lens appeared to be very strong against veiling flare, this is both an excellent performance and exactly what I expected from Nano coating (as it relates to T*). If all of Nikon’s new lenses used the Nano coatings the line-up would be pleasant to use indeed.
Conclusion
Simply put the Nikkor 24-70mm is a good lens. There are better, faster, smaller and lighter lenses out there, but none that are all of these at once and cover this whole focal range too. In this respect the 24-70mm is a truly excellent zoom lens which deserves its high reputation.
Would I recommend this lens to other photographers? Yes, but I am never going to buy one myself. The size of it scares me away and I enjoy better quality already in smaller primes. I might rent it again, but I’m not even sure of that.
Once again I wish I could recommend an alternative. My most desired wish from Nikon is to see a 24-??mm f/4 VR lens. Usually people will chime in “24-100mm f/4 VR” but personally I would be thrilled to see “24-70mm f/4 VR” if it were of a reasonable size and weight. This is what I would like to recommend.
Dust settling after M9 launch
It’s been about 3 weeks since the launch of M9. I love launch events where people get cameras in their hands the same day. We have DNGs available, lots of first-hand accounts…
Observation #0: M9 is awesome. Except where specifically noted below, I think one can assume the camera is great. If you don’t push the envelope on it then it’s just plain great.
Observation #1: KAF-18500’s noise is similar to KAF-10500’s and, adjusting for print or display density, it is comparable to contemporary APS-C cameras. The best supporting evidence is found on Sean Reid’s site.
At first I thought I was seeing terrible results compared to the D700 but I changed my mind after comparing the grain of hand-processed DNGs and NEFs on my own PC. The noise that’s there also looks like it can be tamed with noise processing plug-ins for LR/PS like DFine 2.0. Hence you still need to stay close to windows during daylight, get faster lenses if possible and brace yourself (literally).
My only remaining issue with M9’s sensitivity (indirectly) is that the M9 lacks a convenient way to fix bad lighting conditions with controlled light because both its flash units are unacceptably awkward. I’m convinced that the M9 with a small flash (see: SB-400) can deliver as good quality as I can extract from the D700.
Observation #2: KAF-18500’s detail levels compare favorably to 24MP cameras but in exchange the risk of moiré is real. The best supporting evidence is found on Erwin Puts’ site.
It would probably have been more convenient to use a 24MP sensor with an AA filter. At this point Adobe Photoshop probably can’t be avoided anymore – and I have been avoiding it – so either learn to deal with moiré by hand and apply yourself on batches, or write a Photoshop macro for that, or get CaptureOne’s moiré removal plug-in.
There exist AA filters that screw into lens like ordinary filters, but so far they haven’t been made in sizes less than Series-8. Still, this is a very interesting avenue to explore. Maybe the company that makes them can be compelled to make smaller sizes to cater to the small rangefinder market.
Observation #3: M9’s system board writes its large files to SD flash memory at a slowish rate. It takes 11 seconds to write a 37MB file, at a rate of around 3.4MB/s, and faster cards aren’t really supported by the system’s hardware or firmware.
Compared to, say a D90, the M9 writes 3x larger files at about 0.5x the byte transfer rate – so if you think continuous shooting on the D90 is compromised, you haven’t experienced an M9 yet. Apparently under some conditions you can get your M9 system board to freeze if you keep hitting it with more frames while it’s busy writing, like the M8, but these reports aren’t widespread. There isn’t a solution for this unless a firmware update adds lossless compression or delivers the faster rates expected on faster cards.
This is fodder for a DSP-vs-ASIC (e.g. Maestro, Expeed, Digic, Bionz) debate that will lead us eventually to the M10. In the mean time the mantra seems to be “one shot is all you need, but 2 or 3 is supported for your convenience”. :^)
Conclusion from the above: The ugly sides of the M9 are now known. I don’t see a deal-breaker, but the effervescent enthusiasm phase has passed.
A tale of two 18MP cameras
In this first half of September 2009 we’re looking at two important new cameras, each vying for the position of standard-bearer for their categories. Both are built around high-tech 18MP sensors, but that’s where the similarities end because the two cameras could not be more different. One is a fast-action do-it-all reflex camera and the other is a more reflexive camera that will do all that you can make it do.
The 7D
Could we be looking at the end of the road for the xxD series as the standard advanced-amateur camera? The series surely isn’t dead but I’m not sure if the advanced-amateur will be the target from here on. There is one clear message from Canon though: if you were considering buying a D300s, please consider buying a 7D instead.
This said, the rest of the message from Canon is hard to hear because there is little 21st century communication coming out of Canon. There isn’t a 30-page PDF loaded with professional images shot by this year’s coolest photogs. There isn’t a flash website to take you through the features and the philosophy of the camera. There isn’t an interactive learning center where you can poke at a fake 7D or watch people shoot a 7D through its paces.
At face value the 7D has (perhaps simply is) more of everything. Lots of pixels per frame, lots of frames per second, good HD video. There isn’t much to say because really it looks great at everything.
The only problem the 7D is facing is lens starvation. The meaning of starvation is either that you have no lenses to choose from, or the lenses don’t feed your sensor to satiety. I looked at the image samples from Dpreview, Rob Galbraith and others, and few of them have managed pictures to show off the 18MP sensor. The Dpreview samples simply aren’t good and they were (most/all?) using Canon’s topoftheline EF-S 17-55 lens. Rob Galbraith’s were all flat and muddy. The resolution chart from Imagine-Resource is the only one I found that was detailed. These soft images aren’t caused by the AA filter either, because the resolution charts manage to generate moiré.
A 7D owner who’s planning on actually getting 18MP images is going to need really excellent lenses. I mean forget 90% of the Canon lens portfolio because they’re not good enough. This is going to be a very serious undertaking for those who are minded to harness the resolution that’s included. You will need primes, and they need to be modern, like from the last 2-3 years tops. Look at Zeiss ZE’s if you think you can use them.
The M9
Forget the last 3 years, this is the first digital M camera. The M8? It’s going the way of the M5. The M series goes M4 to M6, M7 to M9, and so say we all.
No I am not putting a yet-unannounced product on a high pedestal. The M9 is probably still going to have variously-important flaws.
It probably doesn’t make any concessions for eyeglasses. It probably doesn’t eclipse the M8 for memory interface performance. Weather resistance probably isn’t on a higher level than the M8’s. It probably is still susceptible to accidental button presses. The battery and SD card are probably still hidden under an awkward metal plate. There probably is nothing surprising included either, so for example your dream of image stabilization probably probably isn’t being realized.
We’ll find out the details over the coming weeks, I’m sure David Farkas and Erwin Puts will have lots of meaty details for us to read, among others. I’m looking forward to the Dpreview preview.
All of these probable flaws are irrelevant however, because more likely than not the M9 is now a good enough digital platform to be viable in the long term. The M8’s imaging pipeline plagued it with infrared contamination or cyan drift (take your pick), took away your high-speed wide-angle lens options and gave you few reasons to want to use it over a high-end Canon or Nikon camera. These issues we can expect have all been fixed with the M9 and that’s why it’s not an M8.3.
This is all that we really wanted: an M camera that does what it needs to, with little fanfare, and delivers richly detailed images that meet the current state of the art. Compared to the 7D, this camera will have no problem finding the right lenses to match. From day one the M9 will meet with the best lens portfolio there is, ready to crank out images to Leica’s signature.
It’s interesting to note that the M9 probably still leaves open plenty of ways to compete against it. A slightly larger body (à la Zeiss Ikon) with a larger viewfinder, weather seals and a well-positioned CF card slot at a lower price could compete successfully against the camera with a red dot. Any kind of body with a motion-stabilized Sony backlit sensor could compete too, at any price. A Contax G style autofocus rangefinder is also an easy competition option. Combine all three ideas and maybe you even have an M9 killer (oops, let’s hope not)(oh no, wait… let’s hope so).
I don’t see any threat to the M9 from micro-4/3. It’s not even close. The M9 is all about wide-angle journalism and micro-4/3, fitting in coat pockets. Except in fantasy the two systems do not meet.
In closing
The 7D is a nice camera, one which I think will feel most at home with an array of Zeiss lenses. I think many amateurs will have a blast with this camera, even if most of them will never know what 18MP really looks like. Either way they will love their camera and their pictures, and this is all that matters. Good job Canon!
The M9 should restore Leica’s place in the world, assuming that the price is right. Since the M8 makes no apparition in the new M system brochure (i.e. it is not shown as a low-end substitute) I will presume the plan is a strict replacement at a similar price. I can’t wait to get mine when my number comes up. Welcome back Leica!
For me personally, there is no discussion to have over the 7D and M9. I have no stake in Canon and don’t see the 7D as a historic milestone worthy of a cross-system leap. The M9 however might well set my path for me – I think it shows the way I have to go, I feel compelled by it. Unless some disastrous flaw is revealed in the M9’s sensor then I will eventually get one even if I have to skip all other of life’s luxuries.
I want to believe
Having viewed the little Leica “pre-mercial” a half-dozen times… they’ve got me hooked.
“For all photographers, novice and professionals alike”.
I want to believe. So desperately.
Sony A850 ties with Olympus E-P1 and Pentax K7 for cameras of the year?
[This isn't much a posting, I apologize.]
A 24×36mm stabilized sensor with 25 million pixels for less than $2000 by next year. The camera is small and well built, its viewfinder is spacious and bright. What does that say about the high end of the APS-C camera market? Does it still have room for high profit margins?
Starting next year it’s going to be difficult to price an APS-C camera where the D300s is today. We’re supposedly 4 days away from the unveiling of the Canon 7D, a high-end APS-C camera if rumors are true. Will the 7D be an outlier in its genus already on the day it is launched? Meanwhile the internet price of the D700 has fallen in price to almost $1000 less than what I paid for it in Massachusetts – unable to hold a premium price.
When I wrote the title above I had a moment of pause. Neither Nikon or Canon really deserve the label of camera of the year. They need some real innovation and soon. A high quality imager does not necessarily imply a large and heavy camera system with a premium price.
Side note
I was heartbroken when I read the rumor about Zeiss abandoning their digital rangefinder project because they felt it wouldn’t sell to pay for itself. I don’t quite believe it though. Maybe I’m in denial.
Current rumors
Looks like I just tempted fate, because we’ve got lots of rumors of digital rangefinders this month.
My 2 cents is that the Zeiss rumor is quite believable. The key technical aspects look to me like they’ve been solved. They have been partners with Sony for a very long time. Sony has an excellent full-frame sensor that will get the backlit-silicon treatment (Exmoor-R) soon. Sony also has nothing to lose from adding a rangefinder to the line-up, and much street-cred to gain.
There’s two basic ways this can play out:
1) Zeiss develops an Ikon Digital camera with Cosina’s body, and a sensor from Sony.
2) Sony develops a Hexar D-RF camera with its own sensor, and body parts from Zeiss/Cosina.
It all depends who you see as the driving force. A company like Sony has a marked advantage with the overall implementation of an electronics-rich product like this, but do they have the will to do it? I think Zeiss has the will right now, it’s time, but do they have the electronics know-how and the support network? It can be a partnership but someone’s got to be on top and I’m not sure who it is.
My bet is that we’ll see this Zeiss/Sony rangefinder camera at Photokina 2010. We might get a sneak peek before that if Solms makes noise about its own plans.
The Leica rumor – that 9/9/9 story – doesn’t sound right to me. It’s not impossible we’ll get an announcement about the M9’s development, but I don’t expect a tangible M9 camera. Ditto for a “Digital CL” or whatever an entry-level M-camera might be called.
The only way I see an M9 showing up on 9/9/9 is if they aborted a sensor upgrade plan for the M8.2 and re-branded that into an M9. This is something that might have gotten shelved to work on the S2, now being rolled-out to make extra money this coming Christmas season. There was some hint of that upgrade idea long ago if you recall Mr. Lee’s comments.
That could take the shape of an integrated IR filter (a technology which they admit they have now) and/or a boost to a slightly higher resolution and lower noise sensor still with the same 1.33X crop factor. The problem with a resolution change is that it would almost guarantee you’d need to swap every other piece of electronics to match data rates, which pushes further in time for me.
Maybe I’m underestimating Leica’s engineering freedom and manpower. I don’t know. I’m being pessimistic maybe.
Humble Beginnings
I stepped out of the box and into the unknown…
Got the first few rolls developed and scanned at a 1-hour place. The scans received a heavy dose of automagical consumerist processing, sadly. They’ll have to do until I can get myself co-located with a Nikon scanner.
Mystery lens revealed!
What is it?…
It’s the 24mm Elmarit-M ASPH!
It’s not actually much bigger thanthe 35mm Biogon.
I’ll have much more to say later, but for now I’ve got some people to go shoot with this.
