THE STORY
Last August I was working on the production of the film Caress, directed by Sujay Utkarsh, just outside of Boston. I was the sound mixer/boom op on the last day of shooting as well as the sound designer, Foley artist and rerecording mixer. Caress is a 20-minute Psychological Horror short that follows the demise of a young couple during quarantine.
A main theme throughout the film is whether or not the lead characters are being possessed by a malicious entity or whether they are being driven mad by other means. We all knew that sound design was going to play a huge role in this film. None of us knew how large a role it would actually play...
About 1-month out from the final cut of the film, I get a frantic call from the director. "Brandon, do you have backup sound files from when you were the sound mixer?" I answered of course, "YES! let me get those over to you right now." Apparently while he was trying to transport the movie file from one computer to another the audio files got corrupted...
Now that's not a huge problem, because we had our files backed up and we sent them directly over to Sujay. But like I said earlier, I was not there on the other days of filming... and the other sound recordist had since deleted his production takes.
So there we were, a full film and only half of its dialog.
We discussed our options and came to a consensus. We would sound design the beginning dialog. Yep, 7 minutes of original sound designed conversation. We're not talking bringing in voice actors and throwing effects on them.... no. We're talking completely designed synth voices, that react and change as dexterously as the human voice. Think of this: you're listening to someone talk to you under water. The voice is muffled, but you can still hear their inflection. This was the thought here except it needed to sound bigger, scarier, and more...extra-dimensional. The sound design would carry the tone and inflection, and the subtitles would carry the direct message.
Sure, losing the first 7 minutes of dialog was not ideal but, what happened was we were forced to get creative! This production flub was harnessed and used to increase the creative scope past that of the original ideation.
THE SOUND
I started by opening up iZotope's Iris synth plugin and messing around with different samples. The cool thing about Iris is you can program four different samples into one instance of the Iris plugin. Iris also has an awesome collection of samples to choose from. I found that the most voice-like sounds happened when I used a sample from the "Pulse" menu.
So that's what I started with. A continuous pulse sound and then I added one or two more samples in to give each of the character's voices their own distinct color.
After this I went in and recorded midi for each of the vocal lines. The midi here is actually triggering the speed that Iris plays through each of the samples, not the pitch. I used higher notes to convey agitation, and lower notes to convey calm.
I was trying to get the synths to act like real voices, which meant there had to be some EQ control. When a person opens their mouth they are acting like an EQ opening up from low pass (a closed mouth hum) to fully open (singing). I set an envelope to open the 'cutoff' when the synth was triggered. I also used a slow LFO on the 'coarse' and 'fine' pots to give some variation in pitch. I set each of these knobs going in opposite directions. So, if the LFO of the 'coarse knob' (semitone motion) was moving downwards in pitch, the motion of the 'fine knob' (microtone motion) was moving upwards, creating a sort of push and pull, and keeping everything a little more organic in feeling. Less robotic.
To get the synths to trigger, I recorded each of the lines of dialog with my own voice, and set the track to open up a gate on the synth track. Then I muted my vocal line so none of the signal was audible. It was just acting as a trigger for the gate.
The end result was a conversation between the two main characters!
The last step was bouncing the audio and bringing it into the main session with the main mix. Once in there I added some color to the male vocal part because it just didn't seem to have the same impact as the female. I did this by adding a SansAmp PSA-1 distortion plugin on an aux channel and mixing the sound in with the original male vocal part. Then I wrote in Drive automation to increase the harmonic texture as the conversation ebbed and flowed.
The last thing I did was add my favorite reverb onto the vocal tracks! After all, they are supposed to be having a conversation in some demonic language possibly in another dimension! Who knows how sound travels there?
For this I added Native Instruments' Raum to each of the vocal channels. It added the perfect amount space and glue to the scene.
Currently the film is in the final stages of post production and is set to premier in early Fall of 2021. Definitely keep an ear out for those synth voices in the beginning. Now you know how they came to be!
Pictured below is the Main Mix file. This is where we compiled all of our sound designed material and made the finishing touches.
