Blackmagic Pocket Cinema Camera 4K: Hands-On, Down to the Details

There can’t be much debate on this:  the Blackmagic Pocket Cinema Camera 4K is the most hotly anticipated product of 2018, across the full range of filmmakers, from young indies to shouldercam production boutiques (as I wrote about here previously).  Not just a toy, though costing only $1.3k, it promises to frankly humiliate cameras costing many times more, outperforming them with some key specs that include:  10-bit 4:2:2 UHD-4K internal recording, gigantic touchscreen display, robust audio inputs, portable SSD recording bus, and best of all (eventually), Blackmagic RAW.  Mainly, it means that plenty has been written about it by now:  so, no value here rehashing the spec sheet (like most blogs).  But recently I got the scarce occasion and good luck to get my hands onto one, and to ask a few trailing questions of Bob Caniglia of Blackmagic Design who kindly gave answers the best he could, while the product is heading to market (and I audio-recorded our conversation to ensure accuracy, when I quote him here).

Power, USB-C Connector and Portable SSD Compatibility

As an early adopter of the original Blackmagic Pocket Cinema Camera (should we call that “HD” now?), I struggled with dropped frames using certain SD cards whose demonstrated write speeds were clearly faster than what the camera could even send to the cards.  Blackmagic ended up publishing an “officially supported” list of SD cards, and it basically identified the most expensive one on the market at the time, SanDisk’s Extreme Pro (not just Extreme, not even Plus) capable of writing at up to 95 megabytes per second (MB/s).  It solved the dropped frames, even when shooting at, say, 1080p/30 in ProRes 422 which is way lower at 18.4 MB/s.

Fast forward to now, and we’re in the world of UHS-II SD cards, and CFast 2.0.  The former is mandatory for ProRes at UHD-4K, but you can get away with UHS-I at 1080p.  CFast 2.0 (still an overpriced luxury) is only necessary for CinemaDNG raw which needs up to 272 MB/s (but as discussed later here, soon Blackmagic RAW may become available in a firmware update, with much lower write speeds for dramatically smaller file sizes).  For now, I’m settling with a makeshift 64gb to get started, while I’m grabbing quick shots on the run, using this Transcend UHS-II card at a so-called “V90” speed rating for about $65.  Notably, UHS-II alone does not ensure sufficient write speed.  V90 as a further speed class, does (enlarge the specs at right for a better view).

Bob cautioned that an official list of supported storage media is forthcoming, but another great option to buying expensive new-gen flash cards (with their current low capacities) is to directly connect a portable, bus-powered solid state drive (SSD) to the new BMPCC4K’s USB-C port. There, it’s a safe bet that Samsung’s rather toy-looking shiny blue T5 series will work well, because that’s what Blackmagic is testing with. Personally, I went with the SanDisk Extreme SSD — I like its form factor and weather proofing better. At left, you can see the test results that I ran.  Should be fine!  But Bob cautions, “There are compression schemes that certain [SSD] manufacturers are using to boost their overall capacity, that did not jive with our recording video — because it’s not just data, it’s video, even though it’s captured as data.”  My tests show that there’s a hell of a lot of headroom, but I’ll update here when I’ve confirmed that it works with a sampling of blank media.

Speaking of that USB-C port, a bit of background seems helpful, especially these days since everyone’s getting it wrong!  Something not widely known (because it’s a little embarrassing to the USB standards group), is that USB 3.0 actually got retired — poof, it doesn’t exist anymore!  Now, USB 3.1 Generation 1 is what we used to call USB 3.0, and its maximum throughput is still 5 gigabits per second (Gbps).  The upgrade to that is USB 3.1 Generation 2, doubling the speed to 10 Gbps.  To begin with, most SSDs today can’t even max out Generation 1, and most computers don’t even have a Generation 2 port.  And there’s where it gets confusing:  USB-C, the new rounded connector that (finally!) doesn’t care what’s up and what’s down, just plugs in firmly, compared to Micro-USB and its variants which were dangerously weak.  But just because it’s USB-C, doesn’t mean it’s faster, if it’s carrying USB 3.1 Generation 1 or worse.  That said, my SanDisk Extreme (and the Samsung T5) SSD are compatible with Generation 2 via the USB-C port, so even though the BMPCC4K doesn’t end up surpassing the 5 Gbps mid-point, any computer can offload (or edit from) the SSD at very high speeds using a compatible Generation 2 bus.

And there’s just one more thing about USB-C:  it also doesn’t guarantee any particular power level, even though we’re starting to see USB-C power supplies at 45 watts and even 90 watts for laptop computers.  Thus, the typical USB 5 volts that’s usually in the zone of milliamps (and occasionally an amp or two) isn’t enough to power the BMPCC4K that requires 12-20 volts of DC power.  Bob confirmed later that actually the BMPCC4K’s USB-C port, aside from being a media storage bus, also can accept USB power — but only to charge the battery (and the Canon LP-E6 standard is closer to home at 7.2 volts).  But, get this:  the world is quickly moving to yet another USB standard, called PD (which stands for Power Delivery) and when spec’ed correctly, it handshakes to ascertain which voltage to send, straight on up from 5 to 12 to 20 volts.  I recommend that Blackmagic consider adding USB PD compatibility, using its inherent handshake to ensure that powering the camera directly (not just charging the battery) is enabled whenever provided a proper high-power USB PD source, albeit disabled from a non-PD source.  Until then, for alternative power options, we will need to send 12-20 volts using the BMPCC4K’s proprietary 2-pin LEMO-like connector (for which Blackmagic sells adapters, including a D-Tap and also 2.5mm barrel connector).

Audio XLR and Plug Inputs Mixed with Four-Microphone Array

The original Pocket had a jack for 1/8″ stereo plugs, and its preamp was quite noisy, combining a high noise floor with internal electrical interference.  The 4K upgrade improves upon this dramatically, and the XLR input port is just the tip of the iceberg.  What’s more important, is that an XLR input also usually delivers “balanced” audio using a third conductor, carrying an additional copy of the signal with its polarity reversed until arriving at the destination, thereby accommodating longer cable runs with less interference/noise.  Another implicit feature of most XLR inputs is the delivery of phantom power (usually 48 VDC), so that condenser and other non-dynamic microphones can operate without a dedicated battery.  It’s notable that in order to maintain portability, the Pocket 4K uses a so-called “mini-XLR” port:  it’s very uncommon so far, and it requires extra caution.  For example, my lavalier microphone of choice for years has been the excellent Audio Technica AT899, which terminates in a mini-XLR plug and requires phantom power.  But it’s meant to be plugged into its own proprietary power pack — and if it went straight into the Pocket 4K, it would get fried!  (Same goes for many Sennheiser lav cables meant for wireless packs.)  Meantime, my favorite on-camera stereo mic is the Rode Stereo VideoMic X (that I’ve reviewed here), and its mini-XLR ports gladly accept phantom power as an alternative to a built-in 9VDC battery.  All told, buyers of the Pocket 4K will definitely be scouring for adapters, and there aren’t many.  But to start with, here’s a safe bet for a cable (these are oddly very hard to find), and here’s a shortest version.  Read your specs carefully!

One of the least explained features on the new Pocket 4K is its four-array microphones designed for paired stereo sound capture. It’s not quite Ambisonic audio, but that’s a good place to start for understanding it; another point of reference is a typical Amazon Alexa that uses more than one microphone to isolate unwanted noises and improve clarity.  Bob noted that “Resolve has the tools in the Fairlight section to be able to create [something similar to selective Ambisonic audio], and [the four microphones] give you a better starting point because you have the wider capture, for lack of a better term.”  It won’t quite be necessary to use Resolve’s Fairlight audio tools for processing the recorded stereo audio, but the versatility of honing in the sound from this arrayed metadata, will be a unique function.

And more good news, is that the camera’s operating system is very versatile in letting you mix just about every combination of audio sources onto the camera’s two multiplexed audio channels.  For example, you can designate the recorded left channel to receive the XLR input but the right channel to receive the internal microphones monaurally; you can have the recorded left and right channels receive the 1/8″ stereo input (sorry, no plug-in power), or receive the full microphone array; and so on.  The only limitation is that it’s not capable of a full mixdown — i.e., you can’t record the complete stereo array from the camera microphones, mixed together with the XLR input.  You’ve got to choose one source for each of the two recorded channels.

There is no built-in audio limiter, but there is a “pad” switch to toggle attenuation of the audio level for each of the XLR and camera microphones.  There is also a low-cut filter, which is commonly used to reduce the booming effects of wind noise and hum (though you’re better off adding that in post-production anyway, with the ability to hone in the cutoff frequency).

Bluetooth Remote Control

Unfortunately, Bluetooth is not able to relay a live audio feed of what the camera is capturing, according to Bob.  (I’m keen to that function these days, after writing here about the new Roland R-07.)  It’s solely for remote control, and the iOS app for iPads and iPhones is the same as the one recently launched for the URSA Mini Pro.  (Android is out of luck, which always baffles me given its 88% market share worldwide.  Apple elitism = Darth Vader.)   Bob also noted, “We released the SDK so that people can create their own,” mentioning that there could be a need, different from the iOS Blackmagic app, to prevent anyone on-set from being able to access important/vulnerable functions like the record button.  This is another example of Blackmagic opening up their systems to let the free market determine what’s best for user experiences.  Compare that to the paranoid behavior of Sony, for example, whose “PlayMemories” protocol stubbornly blocks focus control for important situations like gimbal stabilizers (e.g., Ronin S and Crane 2) — Panasonic meanwhile has begun opening up focus control, with its GH5 and GH5S, and Canon always allows it.  If Blackmagic can, I expect they will.

Power Switch (not a button!)

Among the many infuriating product quirks on the original Pocket, was its power switch.  With one quick tap (e.g., a bump while inside a case or, indeed, a pocket), it would turn on and drain the already tiny battery, without you knowing.  And, it would take a long-press to turn it off, which is what you really wanted to prevent it from turning on at the wrong times.  No more here:  it’s a solid boolean switch.  But, Bob warns, there’s no automatic sleep function:  it will burn until the battery’s dead, when it’s on.

Sharpening and Extended Video

The Blackmagic Pocket Cinema Camera 4K inherits two unique features that were introduced after the URSA Mini Pro, into the URSA Broadcast camera.  Both of them, true to the broadcast ready-made notion itself, anticipate the option to “bake into” the recorded footage certain things that you might otherwise have time to fine-tune in post.  First, there’s Sharpening, which speaks for itself:  Default, Medium, High.  (Notably, all camera sensor processing applies some amount of sharpening, distinct from raw, but this function increases sharpening past the norm.)  Second, there’s Extended Video, which has been widely misunderstood by bloggers/vloggers:  it doesn’t extend the dynamic range any farther than the mode named Film, and it doesn’t “extend” the colorspace to Hybrid Log Gamma or REC.2020 HDR.  Rather, it simply captures video in the flatter log curve of Film mode, then on-the-fly converts it back to REC.709 for the final recorded version.  It’s a quick and dirty way to get more dynamic range, and a “film” look, without fiddling if you’re in a pinch for time.  Essentially, it bakes in the official Blackmagic LUT.  To those who claim they lack the time to simply drop a LUT onto log footage, I hope they’ll at least pick Extended Video instead of Video without hesitation (blown highlights be damned).

Dual ISO at 400 and 3200

This is a feature that people are just coming around to appreciate, after its introduction to a mass audience with the Panasonic GH5S.  (Bob insists that the Pocket 4K’s sensor is not the same manufacturing part as the one in the GH5S.)  But there’s still a lot of misunderstanding about what it means, and this will result in worse footage unless it’s used properly.

To begin with, there has always been a sort of golden rule when it comes to ISO:  the lower the better!  Since ISO is basically like a volume control on the sensor, and noise levels (that ugly digital kind of static noise) only increasing with more ISO, you wanted to always start low, then nudge up only as necessary after exhausting the other factors of aperture and shutter speed (the latter usually locked according to the 180-degree shutter rule, and more on that below).  Thinking back again to the original Pocket, over time shooters began to realize that the image quality actually got worse below 800.  This was because the native ISO of that HD sensor was 800 itself, and nudging beneath that value would cripple the dynamic range of your footage — especially accounting for the Film log curve that’s optimized for a native ISO.  Still, that didn’t stop me from oftentimes sinking down below 800, whether because I didn’t have an ND filter handy, or even so, it got just too damn bright with a wide-open aperture for shallow depth-of-field.

The lesson being:  you want to dwell at the native ISO, and go higher (but not lower) only if necessary.  When you add dual ISO into the mix, that principle gets interesting.  The native ISOs of the new Pocket 4K are 400 and 3200.  (Don’t forget that these numbers don’t mean anything in comparison to other camera models — it’s apples to oranges, so the original Pocket’s native 800 doesn’t necessarily “look brighter” than the Pocket 4K’s native 400.)  This means that there are two ISO categories:  200 to 1000, and 1250 to 25,600.  Bob advises that once you start getting close to 1000, you’re better off bumping up to 1250 straightaway.  Isn’t that ironic?  It means that the image will simply look noisier with worse color at 1000 ISO, compared to 1250 ISO which is activating the native 3200 ISO.  As for me, I’ll probably be treating 800 ISO as the ceiling for most of my shoots at the lower native ISO, and then bumping up to 1250 ISO when any more exposure is called for.  These are decisions you’ll make, by the way, from simply selecting the ISO number:  there’s no additional step required to switch between the native ISOs.  It happens because of the number you choose.

Shutter Angle vs. Shutter Speed

One of the frequently violated disciplines of cinematography is the 180-degree shutter rule, which ensures proper motion blur on each frame.  (It’s not edgy or badass to “break the rule,” it’s just amateur.  The strobe-y look of the beach combat in Saving Private Ryan caused by Janusz Kamiński’s high shutter speed represents 0.0000000000000001% of running times throughout cinema history.)  As with the original, the new Pocket 4K lets you simply set the shutter speed to 180 degrees and then let it adjust the actual value depending upon your current frame rate.  But after some digging into the menus, we found that it’s also possible to set things the old fashioned way, using proper fractions (a ratio, so that the denominator is double the frame rate).

HFR and Window Sensor

You’ll notice a little physical button on the back-right that says HFR, for High Frame Rate.  It’s notable that a related function in the menus, Window Sensor, deals with the fact that the 120 frames per second of HFR mode cannot handle a full UHD capture, and it stores the resulting file in 1080p HD resolution.  Most cameras (e.g., my Sony a7 III) give you one choice, to crop the image down so that the framing gets zoomed in.  Here, Blackmagic gives you the option to let the full sensor downscale the image to HD, so that there’s no cropping.  However, for best image quality, you can leave off the Window Sensor option so that there is no downscaling, and the HD capture is happening in the center portion of the sensor:  this results in a further 2x crop factor (i.e., then 4x against the benchmark of a full-frame sensor).

An interesting side note, purely academic, that I discussed with Bob is the way that the total sensor area (in this case, generally Micro Four Thirds) is actually divided into active pixels that you see in your footage, and “ring pixels” around that, used for other things like exposure, white balance, etc.  Remember how Panasonic got cornered by their decision not to include internal image stabilization with their GH5S, in comparison to the GH5?  That’s because the GH5S (in order to acquire more light with larger photosites) maximized the active pixels to such an extent that the sensor couldn’t bob around off-axis to compensate for motion, without revealing inactive pixels.  These are behind-the-scenes trade-offs that camera designers make, adding a layer of complexity to the vanilla concept of “sensor size.”

I learned a few more things, including the fact that the Pocket 4K’s operating system is actually already found in the Hyperdeck Studio Mini, given the same lasting Blackmagic strategy to make all OS’s deploy together and look the same.  I also confirmed that the still picture function is really just a straight frame grab (rather than any higher-resolution, pre-binned image that you’d get from a hybrid camera like Sony’s Alpha series).  Thus you’ll be getting 8.8 megapixels stills, which is simply 4096 x 2160 (4KDCI) resolution.

I really like the form factor, and how it felt when gripped in both hands — but the simple fact is, there’s no eyepiece.  Without that third point of contact, we’re going to be seeing a ton of poorly shot, shaky footage with this thing.  The camera has no internal image stabilization, and for most of its lenses, Panasonic is very skimpy about including Power O.I.S., even though what we used to call “wide-angle” at, say, 25mm is actually telephoto at 50mm after the 2x crop factor.  And of course, Olympus and everyone else completely leaves out optical image stabilization from their MFT lenses.  Self-proclaimed “pro” shooters always harp that ARRI and RED cameras don’t use stabilized lenses either, but this little thing is called a Pocket Cinema Camera.  It’s begging for sloppy operation, without stabilization to compensate for that.  And the camera body’s wide berth will present a balancing challenge to most motorized gimbal stabilizers.  (You’ll still see footage shot on gimbals, but not fully balanced right, and thus visibly struggling if you look closely.)

These things are arriving widely in the next few days.  In the United States, the two biggest dealers are closed for a religious holiday until October 3, even though boxes are sitting in warehouses ready to ship.  But once it arrives with me, I’ll post more results here, including a full menu guide (in the style of my previous one for the a7R III).  If you’re not in line already, pre-orders are backing up into November now, so be sure to lock-in at the usual places:  Adorama, B&H, CVP and Amazon.  And, to hold you over, here’s the direct link to Blackmagic’s lengthy manual for the Pocket 4K.  It’ll keep you plenty busy while waiting…