Wednesday, 4/6/2022

Loading...

Media is loading
 

University

Shawnee State University

Major

Digital Simulation and Gaming Engineering Technology

Student Type

Undergraduate Student

Presentation Types

Poster Group Presentation

Keywords:

Machine learning, voice recognition, voice generation, computer-generated music

Abstract

By representing audio information as a many-dimensional vector, it is possible to derive the characteristic eigenvectors of this audio through data manipulation techniques. From these “eigenwaves”, sounds can be identified or created. The identification of audio waves using eigenwaves has the potential to be useful in many practical applications ranging from human voice recognition to the creation of authentic-sounding computer-generated voices. Some limitations of the proposed method also will be described.

Human Subjects

no

Faculty Mentor Name

R. Duane Skaggs

Faculty Mentor Title

Associate Professor of Digital Simulation and Gaming

Faculty Mentor Academic Department

Engineering Technologies

Share

COinS
 

Eigenwaves

By representing audio information as a many-dimensional vector, it is possible to derive the characteristic eigenvectors of this audio through data manipulation techniques. From these “eigenwaves”, sounds can be identified or created. The identification of audio waves using eigenwaves has the potential to be useful in many practical applications ranging from human voice recognition to the creation of authentic-sounding computer-generated voices. Some limitations of the proposed method also will be described.