NU 2014-099

Inventors

Mark Cartwright

Bryan Pardo*

Short Description

A novel and intuitive way to retrieve and/or create desired audio concepts using vocal input and automated computer analysis.

Background

Audio synthesizer programming is difficult for all but the most experienced users, and almost impossible for novices. The industry has created some products to simplify the complexity of synthesizer programming by providing presents but choosing a sound from thousands of often poorly named presents is also a daunting task. Some research projects allow the user to provide a recording of the exact sound they want as input and then use the system to optimize the parameters to match that input. However, this can also be difficult since users typically don't have a recording of the exact sound they want. Northwestern's technology, SynthAssist, offers users the ability to program audio synthesizers in a more natural, intuitive way. It allows users to express themselves musically in ways that were not previously possible, and broadens the market of potential users/buyers by making audio synthesizer programming more intuitive.

Abstract

Programming an audio synthesizer can be a difficult task for many. The current software-based synthesizers have complex controls and unfriendly interfaces, which limit the productivity of both novice and experienced users. To overcome this barrier, Northwestern researchers developed a novel audio synthesizer system called SynthAssist to help users efficiently find their desired sound in the audio database. In contrast to existing technologies which require users to express their desired sound using confusing synthesis parameters, Northwestern's SynthAssist allows a more intuitive way by allowing a user to express their desired sound by either vocalizing it or providing a similar sound as input. This synthesis programming process proceeds through the iterative automated computer analysis based on advanced algorithms and the evaluative feedback from the users. Therefore, with a friendly user interface and a natural approach, SynthAssist makes audio synthesizers more accessible and lets users focus on high-level goals instead of low-level controls.

Stage of Development

Functional Prototype with a library of 9000 synthesized sounds. Demo of SynthAssist

Applications

• Programming an audio synthesizer using vocal imitation
• Searching through a database of 'preset' synthesizer sounds using vocal imitation
• Searching through a database of sound effects using vocal imitation

Advantages

• User-friendly interface
• Intuitive communication way through vocal imitation
• Efficient database search based on vocal imitation
• No complex controls

Publications

Mark Cartwright M and Pardo B (2014) 'SynthAssist: Querying an Audio Synthesizer by Vocal Imitation', NIME'14.

IP Status

Issued US Patent No. 9,390,695