This weekend I caught up with my co-director Dan Fryer, the one man audio editing machine, to see how close we were to unleashing The Last Bookshop. We had been hoping to announce the film’s completion in February 2012, but somehow the calendar is now saying March.
Dan fixed me with the shell shocked expression of a man who had just spent several hours trawling through web forums searching for solutions to complicated IT problems.
He explained that The Last Bookshop is indeed, ostensibly, finished, but that the data is sitting in two halves: the audio half and the visual half. The audio half sounds brilliant, and is finally finished after the mammoth edit and polish mentioned in my last post. The visual half also looks brilliant, and was finished last year. The problem came when Dan put these two halves together…
They don’t fit.
Things start off fine, but gradually the audio and the visuals become less and less synchronised until you start to feel you are watching the dubbed Kung Fu version of The Last Bookshop. Seeing as Alfred Hoffman doesn’t actually have any ninja fight sequences in the film (more’s the pity) this is very definitely an issue.
So, how has this happened?
The Last Bookshop was shot at 24 Frames Per Second (FPS) which Dan faithfully assures me is actually 23.98 FPS – the kind of pedantry I enjoy.
The film has been edited in Final Cut Pro, while the audio has been edited in Adobe Audition – to get the best out of both packages. In their separate halves, the audio and video match the right length, and all is fine and dandy. However, when the finished audio is exported from Audition into Final Cut Pro to be reunited with the visuals something odd happens. Final Cut Pro seemingly assigns our audio a Time Code rate equivalent to 25 FPS. Which is especially odd because audio doesn’t have a frame rate. Audio doesn’t have frames. Nevertheless, the resultant discrepancy knocks the audio and visuals increasingly out of synch
How do we solve this? Altering things in the ‘Easy Setup’ area hasn’t helped. Dan’s not sure. Maybe we can process the visuals to match the new rate of the audio? Or perhaps we should cut the film into 30-second chunks and painstakingly match the audio up a step at a time, until the whole film is synched up.
That sounds like either madness or genius. But then I’m not overly technical. When I put this to Dan he said, “I’m not really sure I know what I’m talking about either. It’s all very mathematical and annoying.”
The Bakery: refreshingly honest about our collective technical incompetence.