Panasonic HVX200 Workflow

We shot our first feature, The Red Machine, using the Panasonic HVX200 and P2 cards, and despite hearing some dire warnings about the folly of using new technology, we had absolutely no problems with the camera, and we found the whole experience very fluid and easy.

Since finishing the movie, we’ve had lots of questions from other filmmakers about our basic workflow, so while this blog will continue to explore the whole process on an ongoing basis, here’s the quick-and-easy overview of how we’ve been using the HVX200 and its footage:

1) We shoot in 24PN mode, recording to P2 cards. The resolution is 720P, not 1080, because 720 is the native size of the camera’s chip; if you shoot in 1080, the camera has to upconvert the footage, and it’s doing so using a relatively cheap converter. We feel that if we ever need a 1080 version of the movie, we’d rather upconvert just the footage that we end up using, and we’d rather do so at a facility that has a much better machine for doing the conversion.

Each time the camera runs, it creates a new clip in the MXF format, which is a QuickTime movie in a wrapper that also contains metadata about the clip.

2) When the P2 cards are full, we pull them out of the camera and put them into the PCMCIA slot of an Apple PowerBook. (We use the last generation PowerBook, right before the MacBooks came out, because that has the right size PCMCIA slot, but there’s now a PCMCIA adaptor from Duel Systems that makes it possible to use a Mac Book.)

To use a PowerBook for downloads like this, you need a utility program that’s on a CD that comes with the HVX200. (You can also download the program from the HVX200 resource page on the Panasonic website.)

3) We download files onto 100-GB G-RAID-mini bus-powered jump drives (powered through the FireWire 800 cable that connects them to the computer), which are made by G-Tech. The instant we download a card, we immediately make a clone of the footage on a second jump drive, so that from the time the P2 cards leave the camera, there are always two copies of every clip. (One caveat about the G-RAID-mini drives: although we had no issues with the drives during the feature, there seems to be something very problematic with the FireWire 800 ports on them, because since we’ve wrapped, we’ve had to return both of our drives to the manufacturer for replacement after the 800 ports completely stopped working.)

4) I was told by the boys at Plaster City Post that it’s best not to unwrap the MXF files on the laptop, so we do that on our G5 desktop machine back at our home studio. If we want to look at the footage on the laptop while we’re still out in the field, we use the program P2 Log from Imagine Products. This worked out to be extremely useful for us on the feature, because it made it so easy for us to check continuity.

5) Back in our home studio, we copy the original MXF files to a drive. Then in Final Cut Pro (which has to be at least version 5.0.4 to handle P2 footage), we unwrap the MXF files using the command File>Import>Panasonic P2. This generates new QuickTime movies that are placed into the Capture Scratch folder on our main editing drive (currently a 1.5 TB RAID from OWC). We then copy those unwrapped QuickTimes to a different back-up drive. Once we have those three copies — the MXF backup, the unwrapped QuickTimes in the Capture Scratch folder on our main editing drive and the QuickTime backups — we erase the clips on the jump drives.

The new QuickTime clips created by Final Cut Pro still have the same names given to them by the camera — nonsensical six or seven-digit strings of letters and numbers that are useless for identifying the clips in the Final Cut browser. To get around that, we give each file a new name in the Shot/Take column of the browser, but we keep the original clip name as the official Name, so that when we want to import a clip into, say, After Effects, it’s easy to find out how the clip is named on the drive.

We’re also totally anal about organizing the browser, with everything grouped by scene number, then subgrouped by shot:

I saw a quote from Verna Fields, one of the great doyennes of film editing, who used to teach at USC and was a mentor to Steven Spielberg, George Lucas and Peter Bogdanovich:

“The most useful thing I taught I can teach you all in about three minutes. In those days, everybody had the feeling that the old methods hampered them. When I remarked about coding the film and logging, some young filmmakers said, “Cut by numbers like they do in the studios? I’m not cutting by the numbers; I’m cutting by what’s on the film.” I was able to convince the students that in order to achieve real freedom in editing you have to be able to lay your hands on the exact piece of film you want when you want it. It’s really dull tedious, painful work, but you must log in and code your film. If you can ‘t find that close-up, if it’s in the bottom of the barrel somewhere, you have to cut to a second choice; you’re not really free.”

While that may seem unrelated to this kind of HVX200/P2 workflow, it actually isn’t: when you’re going to wipe an entire day’s footage off the cards onto which you captured it, organization becomes absolutely critical — which just shows how little things really change, no matter how much they may seem to…