Fujinon has just announced two new lightweight cine zooms for E-mount cameras, the MK18-55mm T2.9 and the MK50-135mm T2.9. The 18-55mm will cost US$3799 when it ships at the end of the month; the longer 50-135mm ships in July for a similar amount of money, though the exact price hasn’t been finalized yet.
Fujinon made a prototype 18-55mm available for testing last week, and Portland’s Koerner Camera kindly provided a Sony FS7 to put the lens on.
Fujinon sings the praises of this lens for the following reasons:
- Parfocal – Holds focus throughout a zoom. Few still-camera zooms actually do this.
- T2.9 – Fast aperture for shallow depth of field.
- 4K optical performance edge to edge.
- Lack of breathing (change of image size when focusing).
- Professional Cinema Feel in a small, lightweight affordable design:
- 200 degrees lens rotation
- Manual iris
- Proper professional cine gearing
- Built-in Macro feature
- Back-focusable
- Lens tracks to the center: no lateral image drifting while zooming.
Design
The MK18-55mm T2.9 is a large-sensor, cine-style zoom that shares DNA with Fujinon’s Cabrio lenses: large-sensor zooms with broadcast-derived features like a macro mode and field-adjustable back focus, though the MK18-55 lacks the Cabrio’s handgrip. It’s a fully manual, fully mechanical 3x lens designed for APS-C / Super35mm coverage, with a 28.5mm diameter image circle.
The lens is about 206mm / 8.1 inches long and 87mm / 3.4 inches in diameter.
A broadcast-style clamp-on rubber lens hood with a press-in lens cap adds a bit to the length and girth.
The front of the lens is threaded for 82mm filters and has an 85mm outside diameter.
The lens weighs less than it looks like it should—1 kg / 2.2 pounds—due to plastic bodywork. The fixed and rotating elements of the body, including the green ring at the rear, are plastic, though the lens mount flange is metal.
Plastic construction frequently gets the stinkeye, often unfairly: a well-engineered plastic structure can be just as precise and dimensionally stable as a machined metal design, and it’s considerably lighter. Whether it’s quite as robust when used to hammer nails as, say, a Zeiss lightweight zoom or a Cabrio is unknown—I didn’t perform that test—but arguably a $5 hammer is a better tool for that sort of task in any case.
The focus ring spins 200º and is calibrated in both meters and feet: meters on the front of the ring, feet on the trailing edge. The ring focuses a bit past infinity, and down to 0.85mm / 2’9” on the near end. It has a wide 0.8 pitch cine gear, but no rubberized or textured grip for finger-driven operation.
The zoom ring turns through 90º and is also geared for cine accessories. There’s an M3 threaded socket for a zoom lever. The iris ring has 60º of travel and is linearly calibrated and cine-geared. It runs from T2.9 to T22, and like a broadcast lens, it has a Closed setting past T22, so you can fully block the light without having to separately cap the lens.
Aft of the iris ring there’s another broadcast-derived feature: a flange back (back focus) setting ring with a macro button.
For those not familiar with the process, it’s quite simple to set back focus on a broadcast zoom. Set up a focusing target at medium distance from the camera; say, ten feet or so. Open the lens to its widest aperture, zoom in, and focus. Then zoom out, leaving the focus ring alone, and refocus at full wide angle by adjusting the flange-back ring. Zoom in again and refocus with the focus ring; zoom out and refocus with flange-back. Repeat until the zoom stays in focus at both ends of its range; it shouldn’t take more than two or three passes.
ENG (electronic news gathering) cameras aren’t built to the same dimensional standards as high-end cine cameras, so their flange focal distances tend to be more variable, especially after a few hard knocks in daily use. (In tube-camera days, that distance would also change as the camera heated up and as ambient temperature varied!) Having a field-adjustable back focus is essential on ENG cameras, especially with zooms running 15x and higher.
Modern E-mount cameras don’t have the same thermal issues as older tube cameras, but even so, there’s no guarantee that your FS7 and a6500 have precisely the same flange focal depth, or that either one is precisely correct. Being able to quickly adjust the lens to match the camera is a boon, but it’s also a responsibility: it’s very easy to screw up the lens by misadjusting it—or by forgetting to properly lock the flange-back ring in place once you’ve adjusted it (ask me how I know, grin).
There’s also a MACRO switch: slide it back against against its locking spring, and you can essentially override the back-focus lock and focus the lens closer than its focus ring allows. At 18mm, you can focus as closely as 0.32m / 12.5”, or about 3.5” from the front element of the lens, moving out to around 0.64m / 25” at the 55mm focal length. (That minimum distance is a bit closer than what’s specified in the manual; it’s what I measured on the FS7 with both the focus and back-focus rings at their nearest settings.)
The far side of the lens has vestigial witness marks ahead and aft of the focus ring, but there are no “dumb side” scales for iris, focus, or zoom. The lens is designed to be driven from the operator’s side only.
As mentioned, there’s an M3 tapped socket for a zoom lever. There was supposed to be a lever in the box, but it had gone walkabout during this lens’s earlier peregrinations. All I can do is show you a picture from the manual:
The lower half of the picture also shows the use of a lens support post. The fixed barrel between the focus and zoom rings has several M3 tapped sockets, and there’s a three-part support post supplied with the lens that threads into the bottom-most hole.
The support mounts to a user-supplied cine lens bracket (like this one), though the 3/8” threaded hole on the post may need a 1/4”-to-3/8” bushing to mate with the brackets I found in a quick ‘net search. The main post is 25mm / 1” tall, with two 18mm / 0.7” extenders to be used as needed; lens brackets typically have a height-adjustable screw for fine-tuning.
Here’s the post in action, holding up the lens for photography:
While the 18-55mm is quite light, a support is still very helpful, especially if you’re driving the lens with a follow-focus or with motors. Aside from the FS7 Mk II with its positive-lock mechanism, E-mount cameras use spring-loaded bayonet mounts, and the force of a follow-focus or motor can easily rock an unsupported lens in its mount, wobbling the picture in the process.
Handling and Operation
The MK18-55 is a cine lens and works like one. All three control rings are geared and turn in the “correct” directions. Control rings rotate in place with no longitudinal travel. The front of the lens is fixed, neither rotating with focus changes nor moving in and out as the lens is focused and zoomed. The 18-55 and its companion 50-135mm are the same size and weight, have the same front diameter, and the same control placement, so you’ll be able to swap between ‘em without having to reposition your matte box, follow-focus, and/or other lens accessories. Similar weights and sizes also mean you can swap these lenses without having to radically rebalance your Steadicam or gimbal.
All three control rings are well damped and turn with consistent drag. They may not have quite the silky, buttery-smooth feel of the rings on a Cabrio, Zeiss, or Cooke zoom (nor do they cost as much!), but they’re as good as or better than those on any other lens I’ve used near this price point. Side-loading the focus ring with a follow-focus elicited a faint rasping sound at one end of the focusing range, but it was of the mild “there’s stuff moving around in there” variety, not a nasty “something’s out of alignment” sort of noise. It may be a prototype-only issue, as a similar noise was with the prototype Veydras I tested.
All three rings are cine-geared but are otherwise unprovided with gripping surfaces, so it’s well-nigh impossible to tell ‘em apart by feel. When I ran around ENG-style with the 18-55mm (no rods, no follow-focus), I often grabbed the iris ring instead of zoom, or zoom instead of focus. Yes, part of this is “reviewer works with unfamiliar lens for a freakin’ hour and hasn’t learned where the rings are? Well, duh!”—guilty as charged. But I don’t have quite the same problem when there are textural differences between rings. Simple solution: attach the stubby zoom lever, and it becomes readily apparent when you’re ahead of it, behind it, or on it. I used the support post in lieu of the missing zoom lever, and my ring-fumbling problems were solved.
The back-focus and MACRO settings worked as advertised. Fujinon warned me that the back-focus ring was a bit loose on this prototype, but I found it no looser than those of many broadcast zooms I’ve used.
The sliding MACRO lock switch needs to be pulled back against spring tension before the macro ring can be turned, and that spring is stiff enough that you’re unlikely to engage macro mode by accident. No complaints there.
The lens can be tilted or shaken without any squeaking, creaking, or clunking noises: everything inside and out is firmly attached and well secured. It feels solid, like a potato… or a professional lens.
Aside from my whingeing about the undifferentiated control rings and the lack of dumb-side markings, there’s nothing I can find fault with in the way the lens handles or operates. It just works.
Performance
The MK lenses are fully mechanical zooms with no electronics of any sort. They are old-school designs in that the lens must stand on its own: it can’t depend on any cleverness in the camera to correct for geometric distortion or shading errors or chromatic aberrations. By the same token, zooming and focusing mechanisms are cam- and helicoid-driven; there aren’t any linear motors with bizarre servo maps allowing otherwise impossible feats of element positioning. By making the MK a fully mechanical lens—a completely honest lens, if you will—Fujinon has set themselves a difficult task.
Now, the 18-55mm I had was a prototype zoom, and as Roger Cicala says, “measurebating zooms is a fool’s errand.” Nonetheless, let’s do some measurebating. Shipping zooms may not be precisely like the prototype and one should always expect some sample-to-sample variations, especially in zooms, but it’s reasonable to expect that this lens is generally indicative of the performance you’ll likely see in the shipping product. With those caveats in mind, let’s look at some frame grabs from the FS7’s 3840×2160 images.
Resolution
Fujinon claims “4K optical performance edge to edge.” I shot a DSC Labs MegaTrumpet chart at 18, 25, 36, 45, and 55mm and didn’t find any reason to disagree. The corners are slightly softer than the center of the image (no surprise there), but even at the edges, I saw plenty of aliasing on the 2000 TVl/ph ends of the resolution wedges—a clear indication that the lens was happily passing more high-frequency detail than the 4K sensor could deal with.
Interestingly there was very little difference with changes in aperture, down to around T11 where diffraction starts kicking in (that’s not a lens defect, that’s just physics). Normally a wide-open lens is noticeably softer than when it’s stopped down a stop or two. While the 18-55 does crispen up slightly at T4 and T5.6, T2.9 is still sharper than I would have expected. Here’s the progression at 36mm from T2.9 on down to T22 (!). Stills are 1:1 extracts from UHD frames, shot in SLog3 and manually graded for consistent exposure and overall contrast, without a LUT. No sharpening applied in camera or in post. The final couple of samples show more noise only because I pulled ‘em up from the deep shadows (ah, the joy of log!).
Here’s a worst-case corner image: 18mm at T4, this time pulled from the log footage using an SLog3-to-LC709A LUT for reasons explained later.
18mm corners were the softest of all the focal lengths.
Normally I’d expect corners to show the most change with changing aperture, but here again there was remarkable consistency as the iris was varied.
Chromatic Aberration
Have another look at that 18mm corner. There’s some radial blue/yellow CA; that’s a bad as it gets (this image was run through a proper LUT so the color and contrast are as they would be in a properly-processed real world image). Remember, test charts are far less forgiving than real-world shots. In outdoor shooting I sometimes saw slight blue fringing on branches against the sky in the corners of wide-angle images, but I had to look for it. Generally speaking I was surprised at how clean the lens was in this regard; I have primes that show more CA than this zoom does.
Parfocal Performance
One of the hardest things to obtain in a zoom is true parfocal performance, where the focus stay fixed as the focal length is changed. The MK18-55 does very well in this regard: I ran several test at different focal distances, refocusing by eye (using the FS7’s 8x focus magnification as well as the 4x focus mag in an HDMI-connected UHD PIX-E5 monitor) at different focal lengths and seeing what the lens scale said. In every case, the focus stayed perfectly consistent, showing no drift at all as far as I could see.
Yes, many electronically-controlled zooms also appear to be parfocal, but that’s an artifact of zoom-coupled servo focus control. Many such zooms, zoomed quickly, will lose focus briefly as the servo tries to keep up. The MK18-55 stays nailed in place no matter how fast you zoom.
Breathing
Fujinon touts the MK’s lack of breathing: a change in image size as focus is pulled.
It’s not perfect, but it’s darned good for a fully mechanical lens. All my mechanical primes show more breathing at comparable focal lengths than this zoom does.
Geometric Distortion
The MK18-55 shows noticeable barrel distortion fully wide, rapidly ramping to noticeable pincushioning around 30mm and staying fairly constant from there to 55mm. That rapid transition from barrel to pincushion distortion can be distracting during fast zooms.
If shooting straight lines is important, keep the lens at 22mm or so where the pictures are rectilinear.
Tracking
The prototype 18-55mm tracked perfectly: I could center something in the image with the FS7’s 8x magnification, and as I zoomed, that centered item stayed exactly in the center through the entire zoom.
Vignetting and Portholing
Wide open at 18mm the lens shows a gradual falloff of illumination from center to corners. At 55mm the falloff happens closer to the center of the image. If you zoom in quickly you can see a “porthole” effect as the bright center shrinks down. Stop down to T5.6 and the illumination flattens out considerably; at T8 it’s even edge-to-edge.
The corner darkening appears to be a bit over a stop at T2.9: visible when shooting a test chart or a white wall, though it’s far less noticeable in most real-world imagery.
Bokeh
Bokeh—the appearance of out-of-focus blur—is the downfall of many a zoom, but not the MK18-55. While pools of defocused light show a very slight edge emphasis, they’re round, smooth and unbothered otherwise; there’s none of the “onion-ringing” or other texturing artifacts that plague many zooms’ bokehs. Mild cat’s-eye squishing at image edges is pleasingly present.
The iris has nine blades and the blades have rounded edges, so that smoothly rounded bokeh is maintained as the lens is stopped down. Bokeh is consistent whether the out of focus area is ahead of or behind the focal distance, and whether the lens is zoomed in or out.
Color
Compared to the other lenses I had handy—two Canon EF-L zooms, a Sigma prime, and a Veydra prime, the Fujinon renders a warm yellow image, as seen in the iris shot above and the test-chart pix below. Only one of the Canon zooms is shown since both were essentially identical in terms of color.
Flare
The MK18-55 is well-behaved; I had to work at it to get any real flare. With the sun in the shot the lens throws off a fairly restrained series of rainbow radial flares and purplish iris images; zoomed in it’s possible to generate some truly beautiful edge flaring with the sun just out of frame, if you get focal length and iris just right:
But that’s the worst of it (or the best of it, depending on your point of view); the lens is quite clean, contrasty, and flare-resistant for a zoom. In particular, veiling flare other than what’s shown above doesn’t seem to be a problem.
I happen to like the way the 18-55 behaves; there’s enough character there to work with when you want it, but it’s not so prevalent as to interfere with normal uses.
The supplied lens hood doesn’t appear to affect this performance; it’s more of a rain hood and collision bumper than a stray-light controller.
Conclusions
At $3800, the MK18-55 is a highly affordable cine-style zoom for E-Mount cameras, offering the predictability and repeatability of a fully mechanical zoom in a compact and lightweight package. Sharpness is superb, even at T2.9; the lens tracks precisely and appears to be perfectly parfocal (or near enough as makes no difference, which practically speaking is the same thing); chromatic aberration is minimal; bokeh is pleasing.
True, the lens shows a fair amount of geometric distortion, and vignetting at wider apertures; it wouldn’t be my first choice for architectural or copy-stand work. How much these matter in your day-to-day shooting isn’t a question I can answer for you. Shooting bookshelves in a library? Possibly problematic. Shooting people? Probably not so much.
An obvious comparison would be to the new Sony SEL-P18110G F/4 OSS servo zoom: the Sony lens is a 6x zoom with better geometry (possibly due to in-camera electronic correction), optical stabilization, autofocus, and power zoom, for slightly less money. It’s a really sweet lens.
Why, then, would you even consider the Fujinon? Yes, it’s about a stop faster, and has nicer bokeh, but it really comes down to one thing: the Fujinon is a true cine lens, and the Sony isn’t.
A cine lens is entirely and completely repeatable and predictable. Focus and zoom and iris stay nailed in place even if you cycle the camera’s power. You can throw a follow-focus on the lens (and a follow-zoom and a follow-iris, or fling standard cine motors at it and use a FIZ controller) and pull focus to marks, smoothly and accurately, every time. You can zoom as slow or as fast as you want, and the image stays perfectly centered and in focus at all times.
If that’s a set of characteristics that excites you (not everyone wants or needs what a cine lens can do, but those who do, really do), the real comparison isn’t so much to the Sony zooms as to the new, comparably-priced Sigma cine-style zooms, the 18-35mm T2 and the 50-100mm T2. Those lenses (which I haven’t tested) offer a wider aperture, but have shorter zoom ranges, aren’t quite parfocal, have differing sizes and weights, and lack user-adjustable back-focus and macro settings. Different strengths, different weaknesses: what’s most important to you will determine which lens set is most enticing.
Pros
- 3x zoom with constant T2.9 aperture
- Parfocal, perfectly tracking zoom
- Cine gearing for focus, zoom, iris
- Smooth, consistent focus / zoom / iris controls
- Metric and Imperial distance scales on the same lens
- Front does not rotate, extend, or contract
- Sharp, high-res images even at T2.9
- Minimal breathing and chromatic aberration
- Pleasing bokeh
- Minimal flare
- Field-adjustable back focus with macro setting
- Support post and zoom lever in the box
- Same dimensions, weight, and control placement as MK50-135mm T2.9
Cons
- Barrel distortion when wide; pincushion distortion in latter half of zoom range
- Considerable edge darkening (vignette) at wide aperture
Cautions
- No electronic anything: no autofocus, no power zoom, no stabilization, no lens data sent to the camera
- No “dumb side” focus / zoom / iris markings
- No finger-friendly gripping surfaces on any of the control rings; no way to tell ‘em apart by feel
Bottom line
If you need that level of ultimate control that only a true cine lens provides and you rock an E-mount camera, the MK18-55mm T2.9 lens and its longer brother the MK50-135mm T2.9 are worth serious consideration.
Where and when
The MK18-55mm T2.9 should be available at the end of the month; the MK50-135mm T2.9 in July. Fujinon tells me that they’ll be available in the usual places—Band Pro, Abel Cine—as well as at B&H, Adorama, many local camera shops, and perhaps even Amazon.
Disclosure: Fujinon contacted me about doing a review and sent me a lens for 48 hours. They arranged for the folks at Koerner Camera in Portland to lend me an FS7 to try it on, too. However, I had to give both the lens and the camera back when I was done, darn it, and aside from the loans, no one has offered me any compensation or other blandishments to write this review. There is no material relationship between me and Koerner or Fujinon.