🏆 2015 Winter/Spring URP (Excellence Award)


Wansoo Kim

New audio manipulation interface using non-verbal voice

This research suggests a new type of audio manipulation interface called Sound Parrot that uses non-verbal voice as input.

The sound designer just have to express the sound that he/she picked as the original sound using non-verbal voice and express the target sound that he/she wants to make. The Sound Parrot interface will than analyze the expressions to suggest what sound manipulation tool to use and will give some direction about the settings of these sound manipulation tools. The sound designer can tweak these values to make the desired target sound.

This research verifies the tendency to explain the use of non-verbal voice when expressing manipulated sound. Non-verbal voice were used 48.75% on the experiment to express the use of sound manipulation tools using any means necessary.

The result of the same experiment was also used to find common characteristics of non-verbal voice expressions. These were analyzed using the factors used in speech processing. The result of this was used to produce a sound and was compared to the target sound.

The results were promising for some sound manipulation algorithms such as reverbs, filters, and delays but some algorithms such as pitch shift were less promising.

Department of Industrial Design, KAIST
© SketchLab All Rights Reserved.