The last shots were with the sun over the hills with Danielle left in shadow. Lucky for us, we were shooting the aerial photography on a Red Epic. This meant I could change the shadow color temperature to 9300K in post and just make the shot look overcast instead of blue. I also had to do a bit of compositing in the background so the sunny background fit with the flatly lit foreground.
The ground photography was done with a prototype Panasonic GH4 camera; one of only five available in the U.S. currently. We couldn't get the new underbelly grip with SDI outputs because there's only one in the U.S. right now! Needless to say, the camera behaved like you would think a prototype camera would behave. However, we did make it through the shoot since the hiccups were minor and I shot most of the 4k footage to an internal SD card. That allowed a bit of cropping in post. We used the over cranking feature on one shot to get Danielle's hand digging into the dirt at 96fps. I directed her to "smack the dirt like an ex-boyfriend." That seemed to work. Hmmm...
At first we were a bit panicked when the camera arrived without media. The Panasonic rep told us that the camera required hard to find UHS-3 SD cards, which are incredibly expensive. On Friday, after my inbox was full of desperate messages on ideas of how to get these elite SD cards by Sunday, we found out that the requirement is actually UHS-1, class 3 or higher cards. That lowered the pressure a bit and we were back on track to shoot in 4k. Testing proved we were ready to go.
My experience with the new auto-focus was hit and miss. It was fast, as promised, however the focus tracking would often get lost and focus out to infinity with quick camera moves. This *may* have been related to the prototype firmware we were using, but this behavior is pretty much what you expect from contrast auto-focus. That's why Sony and Canon have moved to hybrid contrast/phase detect auto-focus systems. Why weren't we using a radio follow focus system when we had one available? Well, there wasn't time to properly rig and test such a system. That, and I didn't have a way of remotely monitoring the video in HD back to my AC to pull focus. We had what we had and I just had to make it work.
Just to add to my proclamation earlier this year to "fail harder" I was shooting on a prototype Aerial Mob handheld stabilization system - pretty much the same design as a MoVI. I had never really used one of these stabilization systems before. People think it's as easy as handheld, but it's not. Just like a Steadicam, the stabilizer has a certain kinetic time constant that you have to have a feel for. The stabilizer has to be programmed with certain fast/slow response characteristics. For certain moves the response was too slow and others it was too fast. Each time the stabilizer gets "re-programmed" it takes 10 minutes, so we opted to change some of the shots to fit the stabilizer rather than waste precious sunshine. The MoVI is no different according to others I've talked to who have used it.
What I found out last week is that Aerial Mob will be sharing some space at the Tiffen booth at NAB late this week and they will be taking our video as a demo piece! Makes me wish I could go this year. NAB is the pinnacle show for film makers, so that's a big deal for everyone involved.
The audio mix was done at my house with Danielle speaking into a Pearlman TM-1 large diaphragm condenser mic. Tom licensed the music from somewhere and stitched it together. After a 3 hour mix sessions I got it to sound clear as a bell.
Hope you enjoy the video and find it inspirational.