Sunday, April 08, 2012
Seeds: The process of making a music video Part 3...
Testing the LomoKino was a fun little excersize to figure out what the camera was capable of. The first roll was a bit of a disaster in post. The colour was way off and each frame had to be corrected by hand. The shakiness of the movie was due my own inability to perfectly select the individual frames. The camera test was basically a success. It revealed a hitch in the process though. Around 150 frames per roll meant that I would have to individually scan each one. The colour could be taken care of in the scanner menu under "Color Restoration". This worked for the other rolls.
Wet-scanning on the Epson Perfection V750 Pro has always been a time consuming process, but this was a bit more than I bargained for. By the end of the 4-roll test, I’d become a bit better at selecting the frames, but there’s still a bit of a shudder in the final films. This could be taken care of in post, however, by motion tracking stable objects in the frame. This would also be ridiculously time consuming, not to mention expensive, since it wouldn't likely be me doing it.
By the end of my camera test I was able to get 2 rolls a day finished. Realizing that the Seeds video may have as many as 200 rolls to process, the pre-edit post was becoming a bit daunting. Around 30,000 frames that need to be individually scanned adjusted and cleaned up, would make my job a bit more than I had bargained for.
My discussions with Kyle have been centering around a performance of the song. The 6fps of the frame rate of the LomoKino would prove to be inadequate for a good performance. I’m considering a way to get a 24fps output by speeding up the footage by a factor of 4. This could work if the song was slowed down by the same factor of 4. Kyle could mime the slow version of the song at 6 fps in 7 second increments. The downside of this being that the final sped up shots would be 2 seconds long. That’s a lot of cuts for a relatively slow song.
Also, the 2:45 minute song becomes 11:00 minutes (15,840 frames) 106 rolls of 35mm film to shoot the song ONCE. For a 2:1 ratio that becomes 212 rolls of film, or 31,680 frames. If I got fast enough to scan 3 rolls a day, that’s 71 days of scanning before I can eve edit anything. YIKES!
I need to test this “synch” idea. It works in my head, but that rarely translates well into reality.
I emailed Mike Raso of The Film Photography Podcast about his LomoKino workflow. He’s been scanning strips of film and selecting the frames in photoshop. His films are a bit steadier with this workflow, many not as time consuming as my process. I have another test roll left to scan and might be trying this workflow instead.
Great news from LomoLab UK. Jack Quick, the LomoLab UK Manager said that he was really interested in working with me on this project. They also process and scan! The crux of the conversation centered around the size of the quicktime files. If the files are too small, the quality of the final film won’t be a high enough quality. The price of the service was a bit of a shock at £12.00 per roll (£2,544.00 for 212 rolls), it skyrockets the budget. There’s also the time factor, 5-6 weeks for the number of rolls I need processed. I figure it's either me or the LomoLab that's going to spend a month and a half with this, so it’s worth considering.