Tuesday, July 6, 2010
So you wanna shoot high-speed HD? Get a Phantom Gold, right? Sure, if you enjoy headaches. Now I don't want to knock Vision Research too much here, but truth be told: the Phantom line of high-speed digital cameras is a pain to work with. I like to compare a Phantom HD Gold to a 2-year old toddler: every thing's just dandy until you take away their ice-cream or forget to perform a black balance every ten seconds.
These cameras can produce stunning, high-resolution images BUT you MUST baby them. Their sensors are extremely susceptible to temperature changes. This is why it is a requirement to constantly check your image and make sure it is clean before every take. If you don't, you'll end up with superbad mojo in post and even worse juju in life. I'm also not a huge fan of the windows operating system that is the foundation of the Phantom's capture software. It works, but then it doesn't. Freezes. Reboot. Not a fan. Remember, Arri never had me boot up Windows XP in order to shoot 150fps on a 435. I really don't see the need for Vision Research to insist on such a weak interface.
OK, enough about the Phantom. Like everything else, it's a tool. Until it fails...
Which brings me to the Weisscam HS-2. As far as I know there are only two HS-2's in LA via Clairmont Camera. Tentative steps...
The Weisscam solves some of the Phantom's shortcomings by running an automatic black reference calibration that constantly, and on the fly, evaluates the dark noise in an image and adjusts for the best image quality. This is a huge advantage as there is no need to interrupt shooting to perform any "Wizard-of-Oz" techie tricks.
The Weisscam's workflow approach is similar to the Phantom's in that the camera is meant to constantly store frames (at a specified frame rate) in an internal RAM module while waiting for the operator to tell it to stop. Once a recording is stopped the preceding footage must be output to either the Weisscam Digital Magazine or via the camera's HD-SDI outputs where it can be captured in real-time (1000fps played back at 24) with a SRW-1, nanoFlash, or other HD recording device. The internal RAM buffer stores image frames as sort of free floating information. None of it is actually recorded onto any media or hard drive. It simply holds your shot there temporarily until you send it out to a recorder or begin recording right over it again. That's why if the camera loses power you've also lost your shot. It was never actually stored anywhere, it was just present for a fleeting moment...
That's precisely why there are two power inputs on the camera body. In case of a battery failure any shot stored on the internal RAM buffer won't vanish into thin air like it would if you have to reboot Windows (hint). Power redundancy is a good thing to have in high-speed digital capture. Does the new Phantom Flex have two power inputs? Yes, I think it does. About time, boys.
And did I mention that the Weisscam comes with this small, super-sweet, touch sensitive LCD, remote bluetooth controller so you can turn the camera on and off effortlessly when you've got it mounted on a 50ft Technocrane? Oh, right, VR also introduced their RCU. About time, boys...
Here's a few observations from the field when using the nanoFlash on high-end, digital cinema cameras like the Sony F35 and Arri D-21.
The nano is a very smart machine. It will automatically detect an incoming HD video signal and display its format at the bottom of the LCD screen. If you are unsure of what a camera is sending you via its HD-SDI out make sure you take a moment to check with the nano.
I've used the nasnoFlash to record directly off a Sony F35's interface box with no problems. But the "monitor out" SDI outputs will NOT pass any Timecode or audio information. Take note because this is important. If you need embedded audio and TC then you must connect the nanoFlash to one of the SDI outs on the interface box. And if you are shooting 4:4:4 then your best bet is to separate the SRW-1 from the camera body, connect it to the SRPC-1 Video Processor (aka: The Toaster!), and use the monitor out SDI output. Just be sure to take note that this signal is set at a default 59.94i frame rate. This means that if you have set the F35 at 23.98PsF you will NOT be getting a true 24P signal coming out of the toaster. If you've enabled your nanoFlash to record PsF but have not added a 3:2 pulldown to the 29.98PsF out you'll end up with motion artifacts.
Fortunately, you can easily add a 3:2 pulldown after the fact to get your 23.98PsF back. But you're best off by either setting a 3:2 pulldown via the SRW-1's video menu or enabling a 3:2 pulldown in the nanoFlash itself.
Also of great interest is the ability for the nano to accept a Timecode Trigger to begin recording. If you are working in a Record Run TC mode this means that the nanoFlash will begin recording when you hit your camera's record button (as soon as the TC advances). If you're working with a single-system (audio + video are one) this is a sleek and efficient process. If you're running a dual-system (audio + video are separate and will be matched later in post) then you're most likely using a Time-of-Day TC mode and you'll have to trigger your nano manually.
Some more tidbits: A 32GB CF card will easily hold more than 50min. worth of 23.98PsF material at 50Mbps Long-GOP (HDCAM SR anyone?). At 100Mbps you'll get about 40min.
You can record any Log Gamma modes with a nanoFlash including S-Log, Panalog, etc. as long as you're sending the nano a 4:2:2 signal. Your D-21 set to Log-C 4:4:4 will have you seeing funny things in post production.
You may need to add the XDCAM plugin to Quicktime Player in order to view nano QTs properly. Check: Sony Support or Calibrated Software or Perian.
Be sure to consult Convergent Designs approved Media page when picking out CF cards to use with your nanoFlash. It never hurts to test something before you push the button and sink the Titanic.
Finally, the nano3D rig is right around the corner. Having two synced nanoFlashes for 3D television and commercial work will be just awesome. Plus I bet you can record 4:4:4...
OK, so this product has been on the market for about a year now. It was aimed at the prosumer video shooters as a way to bypass a camera's internal, and most likely inferior, recording codec by capturing a HD stream via the robust Sony XDCam codec at a very impressive max 280 Mbps.
There are many things to like about the nanoFlash. But the most important is the fact that this little device (measuring approx. 4"x4"x1" and weighing in at under a pound) records your choice of Quicktime or MXF files (and a couple others) directly to solid state media. In this case, pro-level compact flash cards. What this means is that for a mere $3K you can enter the fantastic realm of a simple data-centric workflow and leave all your tapes (and tape decks) behind. Because the nano is built and engineered by an independent company, Convergent Designs, you are not tied into any proprietary recording format or recording media. And because this unit accepts both SDI and HDMI (mini) inputs you can use it to record just about anything that will send out a HD signal.
A few tech specs: the nanoFlash is an 8-bit video recording device that delivers 4:2:2 color sampling at selectable compression bit rates. For max quality you want I-Frame at 280 Mbps. What does this mean for a 4K Digital Intermediate? Capturing at 8-bit instead of 10-bit will reduce your dynamic range. And 4:2:2 won't give you all the color information that a 4:4:4 recording will. But if you've got the cash for a 4K DI then you're probably shooting film anyway.
So what about the small screen? Last time I checked the max bit rate of HDTV broadcasts was a mere 19.4 Mbps. Not to mention that what's sent over the air is usually 4:2:2 at 8-bits. The nanoFlash has the ability to deliver fantastic picture quality for episodics and commercials. 'Nuff said.
Why spend the cash on HDCAM SR when you can Velcro a nano on the side of your F35? Shooting Panalog on a Genesis? The nano can record it. What about RED? Though the R3D workflow has dramatically improved it can still be a pain. So if you're OK with 720p, why not dial in a good lookin' image via the RED Video Menu and capture it with a nanoFlash? And though the current Canon 5D Mark II doesn't output full raster images via it's HDMI out perhaps the next interation will. And then, nano-it!
The truth is this little box can cut your post-production costs dramatically. With new high-end digital cinema cameras hitting the streets everyday the nanoFlash can be your secret weapon. A file-based workflow is inevitable. It holds the promise of immediacy, redundancy, and affordability. Flash memory will get cheaper. CPUs will get faster. And 1s and 0s can already be infinitely copied and stored.
Panasonic had it right when they introduced their P2 workflow. RED got it right when they decided to record directly to CF cards. Everyone is excited about Arri's ALEXA because it can record Apple ProRes files directly to flash memory. And Sony may be late to the game, but they'll be updating their HDCAM SR format to record directly to solid-state media as well.
The thing is, the nanoFlash is already here, it's non-proprietary, it's inexpensive, small, draws almost no power (6.5 watts when recording), and works like a charm. No wonder I own two of them...