High image quality, shallow depth of field and smooth focus pulls are now possible from native Micro Four Thirds lenses, courtesy of the Voigtlander Nokton f/0.95 lenses. And with SLR Magic and Zhongyi Optics with its Mitakon line rapidly following suit (and both SLR Magic and Voigtlander announcing 10mm and 10.5mm lenses coming), concerns over MFT lenses for shallow depth of field or precise focus pull for video are a thing of the past – as long as you don’t insist on native auto-focus.
It was the test I’d been waiting to do since I’d first read about them: the trio of f/0.95 Voigtlander Nokton lenses (17.5mm, 25mm and 42.5mm providing 35mm, 50mm and 85mm full frame equivalent fields of view, respectively) mounted on the Panasonic GH4.
Matt Bruce and the team over at lenslends.com made the gear available to us and at my request threw in the Canon 24mm T1.5 Cine Prime as well, so I could try it out on one of my Rebel SL1’s for comparison.
Yep. As some of you may recall, I traded in my 5D Mk II and some of my L glass at the end of last year for a pair of tiny Rebel SL1’s and three EF-S STM lenses: the 10-18, 18-55 and 55-250. I thought – and still think – the SL1 gives away nothing to the 5D Mk II when it comes to video except for low light performance (you can go as shallow as you want using native Canon glass on the SL1, believe me) and a fuller implementation of Magic Lantern.
So, like, um, having five of them for the price of a single 5D3.
But that doesn’t make the SL1 great for video, just better than the 5D2 – and more than adequate for most situations (live event coverage and low light being notable exceptions, my aging eye sight being another).
Which doesn’t make it perfect, just like every other camera (some people I hold in high regard don’t like the very sharpness of the GH4, for example, that makes my eyes water with desire).
And oh, that Canon glass.
Of course, the Metabones Speed Booster and Speed Booster Ultra allow you to fit Canon lenses to the GH4, but those lenses are big, heavy, expensive, and not optimal for manual focus unless you go into the Cine line – and then you have the price of the Metabones itself.
Thus enter the Noktons, which– with wide maximum aperture (at f/0.95, equivalent to f/1.9 on 35mm full frame), weighty manual focus and manual aperture control ring – remind me of nothing so much as the old Canon FL series manual lenses I grew up with and loved as a kid.
Spoiler alert: I love these Noktons.
I hope that a site like DxoMark or dpreview decides to subject these lenses to the very thorough and consistent testing for which they are known (both the Voigtlander and SLR Magic lenses are currently previewed only on DxoMark, and are not even covered on dpreview), but in the meantime, we did our own relatively informal and imperfect testing and are happy to share our results with you.
Our Testing Approach
We mixed real world and static testing over the course of a few days – but no charts nor studio shots.
No pixel peeping (well, a little)!
Test 1: Time Lapse of a Great Cause
We had volunteered to shoot a time lapse for one of our favorite social impact organizations – North Light Community Center in Philadelphia – the day they and more than 100 volunteers built their new playground. We arrived with the GH4 and 17.5mm Nokton in hand – among other gear — at 7:00AM.
Setup with the GH4’s built-in intervalometer was a breeze, and over the course of almost 8 hours, it fired like clockwork every 60 seconds. When it was over, we still had plenty of battery life left – in sharp contrast to the SL1 which required an external intervalometer and whose battery we had to change midway through (never mind the SL1’s absence of many other features).
Other than setting the camera to “vivid” and aperture priority (the Noktons don’t send EXIF information to the GH4, but setting the camera this way meant it would handle changing light conditions by automatically varying shutter speed and ISO), we just let it run at f/5.6.
No need for cages, intervalometers or external viewfinders or monitors – the GH4 with the 17.5 Nokton was about as self-contained as you can get and did a great job.
At a duration of just 1 frame for each of the still shots on the Final Cut Pro X timeline, you can see the result here.
North Light KaBoom Playground Timelapse
To my eye, the images just “popped” – and as for corner sharpness, check out the tether ball in the upper right corner of one of the JPEG stills used in the timelapse.
Of course, this was not a test of sharpness wide open – the reason these Noktons are so interesting — so for tests 2 and 3 we found what we thought was a pretty perfect test subject for both edge to edge sharpness and focus pull.
Tests 2 and 3: Focus Pull and Time Lapse of A Fading Middle Class
We settled on the garage door of a 1954 split level home that badly needed repainting as our subject and conducted two additional tests.
First, we positioned a bicycle 6 – 10 feet in front of the garage door with the brake lever in the upper left corner of the frame (you know the rap on most high speed lenses and their inability to stay sharp all the way to the corner). The intention was to do a manual focus pull wide open at ISO 200 from the brake lever to the garage. In keeping with the notion of small and light, I opted to focus by hand and a twist of the focus ring rather than employing the Edelkrone FocusONE PRO follow focus in my kit.
We mounted the 17.5mm Nokton on the GH4 and the 24mm Canon Cine Prime on the SL1 (both yielding approximately the same field of view as a full frame 35mm lens, both with wide maximum apertures and both meant for manual focus). We used a Zacuto EVF and Pro-Finder on the SL1 to give us the focus assist that comes built-in with the GH4.
Take a look at the footage in Part 2 yourself and draw your own conclusions. I’m not going to give it away here (not yet)!
Voigtlander Nokton Lens Review Part 2: Focus Pull and Sharpness in motion
Then we shot still images (all RAW) using the GH4 with all three Noktons, and the Rebel SL1 with the 24mm T1.5 Cine Prime, the “plastic fantastic” EF 50mm f/1.8 II and the EF-S 18-55f/3.5-5.6 IS STM – from maximum to minimum aperture, one frame each — and threw them into a time lapse..
No doubt about it, my eyes require as large and high resolution a monitor as possible for focusing these days, and I apologize for both some rotational skew and just a touch of imperfect focus at various points along the way that I only noticed once I had FCPX up on a 27″monitor.
Again, take a look at the footage for Part 3 yourself.
Voigtlander Nokton Lens Review, Part 3: Stills time lapse
So what’s the overall verdict?
Results & Conclusion
In a nutshell, the GH4/Nokton combination rocked, and for what I do at the level I do it, I simply don’t need or want more (OK, I’d LIKE more dynamic range and better high ISO performance, but these are not deal breakers for me and I only want them because I know they can be had for a price, though frankly I’m not willing to pay). It’s a setup that is small, light, fast,sharp, functional, relatively inexpensive – and has a quality feel that makes it a joy for both video and still photography.
And the Noktons? Did I write I love them? Yes, I did.
I was surprised not by the fact that they were sharp (though I was pleasantly surprised by corner sharpness), but by how well they controlled lateral chromatic aberration (LCA) especially in comparison to the Canons – even the Cine Prime (especially the prime, as it is designed for full frame coverage and thus had a theoretical built-in advantage over the Nokton, and yet had by far the worst CLA of the bunch — see below).
And I was delighted by how well the GH4/Nokton handled manual focus pulls.
It’s icing on the cake that you can buy all three Noktons for $1,000 less than the price of a single Canon Cine Prime 24mm (which sells for $4,290 on its own) and buy a set of gear rings for pocket change.
The Cine Prime is lovely and actually somewhere between awe-inspiring and intimidating, but unless you’re mating it to a killer full frame camera – well beyond any of the Canon crop-sensor bodies including the 7D2 – it doesn’t begin to make sense to me optically, never mind the Rube Goldberg ergonomics you see in the accompanying photos to make it usable on a Canon crop-sensor body..
Big surprise, huh?
A Touch of Pixel Peeping
At full screen view on a 27”Apple Thunderbolt display, if I looked closely enough (zooming in as far as I could) I saw optical tradeoffs among all of the lenses — and they were surprising, as you'll have already surmised from my comments above.
Maybe they shouldn’t have been.
The first conclusion I drew was: lenses designed more recently – all else being equal – are technically superior at any given price point, and sometimes without regard to price point. Even at maximum scale, the Noktons to my eye were generally (though not always) as sharp or sharper, more contrasty and better-corrected for lateral chromatic aberration than the older Canon designs of the Cine Prime 24 T1.5 and 50 f/1.8 II.
In direct pixel-peeping comparisons, I liked what came through the 17.5mm Nokton on the GH4 better than I did what came through the 24mm T1.5 Cine prime. And the 42.5mm Nokton at pixel-peeping magnification blew away the Canon 50mm f/1.8 II (then again, it ought to – at $999, it’s 8 times the price of the Canon). There were color cast differences but these are (more than sharpness, I believe) a matter of personal taste, easily adjusted in post.
Might the optical results be different when comparing the Canon lenses with a Canon DSLR other than the SL1? Sure, I suppose so. After all, the older lenses were designed for 35mm full frame film, not crop-sensor bodies. And do Canon bodies now correct for CLA? Yes, they do.
But do I expect big differences between the SL1 sensor and any other crop sensor Canon DSLR? Not really. A quick check on DxOMark shows inconsequential differences in the sensor ratings of the Canon SL1 (100D), Canon 7D and Canon 70D (we'll have to wait for a test of the 7D2)
And it was not our objective to test full frame optics on full frame bodies against crop-sensor optics on crop sensor bodies.
Still, perhaps the biggest surprise of all – and the biggest proof statement that “newer is better” was (setting aside for the moment a legitimate assertion made by some – myself included when it comes to the Canon 50mm f/1.4 — that it is the imperfections in lenses which give them “character”) the little Canon kit lens.
While the EF-S 18-55 f/3.5-5.6 IS STM couldn’t compete with the others for shallow depth of field with its maximum aperture of f/3.5 (equivalent depth of field to 5.6 on a full frame body), it controlled LCA at 23mm better than the Cine 24mm – and at both the center and the edge, again to my eye, was approximately as sharp at the magnifications I was viewing them — as both the Cine and the Nokton 17.5.
On the other hand, forget about trying to focus that kit lens manually. No way. The 50/1.8 was not much better in this regard, and in fact none of the STM lenses — or, current autofocus designs in general — are first, second or third choices for manual focus pulls.
It must be written, however,that if you DON’T need ultimate depth of field (you might be surprised just how often your depth of field can be too shallow); low light performance (as in: there’s enough light for autofocus to work and without the need to dial up the ISO to the point of it being unusable); and slow, precise focus pulls – this EF-S lens has to be one of the bargains of the decade, especially when you consider the value of anywhere between 2 and 4 stops attributable to its image stabilization.
Could I be completely wrong about all of this?
It’s possible. I’d love to see what other people come up with in their own testing.
In The End
What I saw, really, are artistic, logistical and financial choices.
Canon lenses designed for full frame sensors allow you to take advantage of superior low light performance and less viewing magnification to achieve the same size image. Add in post-processing, and it has been proven over and over again that you can get stunning results.
The Noktons are peerless on the GH4, and with the GH4’s video capabilities and combined price point, they are a stellar package for 99% of all shooting situations I’m likely to encounter. They will be a logistical and financial no-brainer for many.
The little Rebel SL1 with the 18-55 EF-S IS STM kit lens costs less than a single Nokton, and if you want to shoot kids or events with people in anything other than low light, this combo is a staggering price/performance value.
Your mileage may vary.
Except for those five fiddly, customizable function buttons arrayed on the top and back of the GH4 that actually have imprinted on them nothing other than Fn1 through Fn5. I want a camera, not a computer keyboard circa 1990!
If I needed low light capabilities I’d look at renting an A7s, but frankly, if we’re talking dawn or dusk (which is where a good chunk of that low light capability matters), 8-bit banding may become an issue until (if) Sony offers a firmware upgrade to 10-bit out. Until then, maybe I’d look at spending more to rent a Sony FS7, or something like the Sony FS700 or Canon C100 with an Atomos or Odyssey recorder – i.e., a dedicated camcorder with a minimum 10-bit 4:2:2 out (and, in the case of the FS7 or FS700, 4K to match the down-converted 1080p of the GH4).
If I needed a telephoto, I’d use a Metabones adapter (no need for the Speed Booster if I want maximum magnification) and Canon lens on the GH4.
In the end, it can be said that we are spoiled for choice, but we’d really love to hear from YOU.
(cover photo credit: snap from Hugh Brownstone)