On September 8th John Martin Goddard successfully defended his Thesis Research, titled "Voice Commands to Control Recording Sessions".
Marty has a long interest in music, music production, and engineering. This work focuses on those topics, e.g., testing the feasibility
of nearly hands-free control of the recording process so users can focus on making music rather than the process of recording.
He has conducted a number of experiments to demonstrate the feasibility of his ideas with a variety of voices, showing that his work
is practical.
Marty developed an effective experimental apparatus that supports voice activated commands, using the open source Audacity
multi-track audio editor. That was driven by his SayPlay software embedding of the Microsoft Speech Recognition engine
and his novel grammatical processing to allow a flexible, musician friendly, environment for managing the editing of production audio tracks.
His presentation materials discussed his ideas and provided a nicely crafted summary of the several types of experiments he conducted
to evaluate verbal command processing methods. Some compelling ideas turned out to be less effective than others for reasons he explored
and documented in this work. He developed other methods that worked very well in practice, over a variety of users and learning
environments.
You will find a nice summary of this work in his defense slides and all the details in his thesis document, below:
He has conducted a number of experiments to demonstrate the feasibility of his ideas with a variety of voices, showing that his work
is practical.
Marty developed an effective experimental apparatus that supports voice activated commands, using the open source Audacity
multi-track audio editor. That was driven by his SayPlay software embedding of the Microsoft Speech Recognition engine
and his novel grammatical processing to allow a flexible, musician friendly, environment for managing the editing of production audio tracks.
His presentation materials discussed his ideas and provided a nicely crafted summary of the several types of experiments he conducted
to evaluate verbal command processing methods. Some compelling ideas turned out to be less effective than others for reasons he explored
and documented in this work. He developed other methods that worked very well in practice, over a variety of users and learning
environments.
You will find a nice summary of this work in his defense slides and all the details in his thesis document, below:
Presentation - pdf,
Background and Design - pdf,
Synopsis of Experiments - pdf,
Techniques for Naming - pdf
Thesis Document
Thesis Document
Presentation Scenes
In the first photo below, Marty is explaining the flow of processing from Speech Recognition, through the SayPlay command grammar processing to Audacity. In the second photo he is describing the structure of his grammar processing and describing the ways in which he was able to give users flexibility in how they construct commands. Then, a final bow.