Milestone I

Non-real-time scoring, sound effects, etc:
The uses of non-real-time sounds will be (generally, and non-mutually-exclusively):

• Scoring purely aesthetic sounds serving the same role as scoring does in film i.e. to build tension or anticipation, to segue between scenes, to be a non-verbal aesthetic cue of the state of the performance.  These will have a significant overlap with the sonic environment sounds but will be intended to be regarded in their own right as significant.  I will use almost exclusively techniques I developed in 461 and 462 for this so I can spend as much time as possible thinking about how to use the instruments versus what instruments or what kind of instruments I would like to use.
• Sonic environment sounds to present an aesthetic context in which dialogue takes place.  These sounds should compliment the mood and pacing set by the actors' actions and speech and will avoid the ~250-1500 Hz band region as to avoid masking their voices.  The same techniques as in scoring will be used along with some found sound.
• Environmental sound effects to give verisimilitude to different settings.  These will mostly be sounds taken from the BBC sound library and other recordings.  Will include among others

1. A party scene: party music, people talking and drinking, etc.
2. The opening scene taking place somewhere studious such as a coffee shop: people holding subdued conversation, perhaps rain or espresso machine etc

• Explicit sound effects to portray events such as the surgeries and riot in the last scene, etc.  For these I will be using found sounds recorded during the workshopping phase as well as sounds taken from the BBC sound library.  The surgeries will be highly stylized so verisimilitude will not be the highest priority; that will be to portray the surgeries as acts of violence on the body and psyche.  The riot will use sounds of angry crowds as well as shouted lines recorded during the workshopping phase.
• The doctor's voice the doctor will be a character portrayed only through sound.  His voice will be altered slightly to foreshadow changes in Jack's voice.  The idea here is to make the doctor sound one-dimensional, as if one aspect (the excited, assuring, 'salesman' aspect) of his personality has taken him over.  I plan to use FFT with the phase vocoder to bring out and shape the higher bandwidth of his voice to make it sound sickly sweet.  Resonz filters will be used to suggest a trite 'happy' chord or interval in his voice, this should be subtle.

Real-time processing - The Executive's voice, Jack's voice & The Voice (which emerges in Jack's mind):
In each monologue after the first surgery, dynamic live processing will be used to portray the emergent characteristics of his psyche that are due to the surgeries.  The two techniques I think will be most effective here are non-linear waveshaping and convolution arranged in parallel formation as to add together the preserved characteristics of the spoken voice as much as possible for the purpose of intelligibility.  The sound source I would like to convolve with is a recording of a friend of mine producing a very noisy tone using an extended trumpet technique.

After the first surgery when only certain thoughts and phrases are being altered, I would like to give the actor as much control as possible over how his voice is altered.  So while I will have the ability to 'tweak' the sounds and adjust the levels, the live processing algorithm will use specific thresholds in amplitude or PV magnitude to trigger a processing of different kinds and degrees.  Hopefully, the result will be a system in which the actor can manage the pacing of the monologue as he sees fit and I will have a minimal role in effecting his voice during the performance.

After the second surgery when Jack is dialoguing with the Voice, while all the lines will be delivered by the actor on stage, I will manually switch the processing on and off (or between different processes) to indicate as much as possible that Jack's older 'self' is in dialogue with an emerged new 'self'.

After the last surgery will perhaps be easiest when Jack's speaking voice is persistently processed indicating that his old self has succumbed to the dominant 'enhanced' psyche.  The executive's voice will be similarly altered with moderately different parameters.

NOTE: I have begun exploring the ProcModR class for real-time implementation as it seems most appropriate for these three distinct uses.