"Dusty Horne's Sound and Fury" - Written by Frank Tamburin, Staring Natasha Pring.
We were recently involved in a pretty fun little project over here at HON HQ that we’d like to share. Our friends Frank Tamburin and Natasha Pring, came to us with a play they’d written about the life and work of little know foley artist Dusty Horne. The vision was to tell Dusty’s story on stage, as an immersive multi-media comedy show. The show would feature several clips from the films Dusty herself worked on as foley artist. The fun (and complicated) bit was that the lovely and talented Natasha, playing Dusty, was to perform the foley track live to projected picture - No mean feat. The problem was that the clips Frank and Tash had selected, still had the real Dusty's foley track embedded within the stereo mixdown of the DVDs audio. Some serious editing/repairing was required, and that’s where we came in...
What We Did and How We Did it...
There were about ten clips from a bunch of different films that we were tasked with removing the foley from. Some were straightforward, and others required a bit more surgical precision. The main difficulty was removing enough of the original foley from the audio track, so that if Tash was slightly out, or something went slightly wrong with her live performance, the original foley wouldn’t be noticed by the audience. Having said that though, in all bar a few of the tracks, the foley was mixed with music and dialogue tracks. If too much foley was removed, the dialogue or music became noticeably effected. A bit of a balancing act was required to get the best sounding solution.
90% of all the foley removal was done within RX5 Advanced, with the remainder (the more musical moments) processed in Logic X. A combination of the dialogue de-noiser tool, and the spectral repair tool seemed to work fairly well. We found, interestingly, that running the de-noiser processing twice, at a low level of reduction seemed to provide far fewer nasty digital artefacts than running it once on a more aggressive setting. There must be some sort of nerdy truth as to why that’s the case, but none of us would even pretend to know anything about that!
Here’s a screen grab of the untreated audio for the Viking Women clip. (You can watch the before and after in the video at the end.) You can easily identify the musical elements as they are rich in harmonics. Using the spectral repair tool in the gaps between these harmonics was surprisingly effective at removing the chains/handcuff sounds. At the end, you can see the large brighter thump of the head smack. We tackled this by duplicating the music just prior to this event, using automation to continue the crescendo, and a reverb match to provide a nice decay tail, which seemed to work pretty convincingly.
It was definitely a trial and error process, but really fun to be involved in. I hear from Frank and Tash that their preview went down a storm and that the play is off to Edinburgh Fringe Festival.
A great project. We wish them all the luck!