Accelerating audio matching on multicore machines

The multiple audio sequences matching problem can be regarded as a pattern identification problem with inputs of multiple highly fragmented audio sequences. Singular Software develops a product which employs a model-based alignment algorithm to match audio sequences on a common time line to solve this problem. This product relies on computationally intensive mathematical operations such as FFT and maximum log likelihood calculation for different models, which limit the performance. This project aims to explore the feasibility and the tradeoffs involved with accelerating audio sequence matching algorithms on multicore processors. This project will explore acceleration using ‘traditional’ multicore CPUs and massively-parallel Graphics Processing Units (GPUs)

Faculty Supervisor:

Matei Ripeanu

Student:

Partner:

Singular Software

Discipline:

Computer science

Sector:

Information and cultural industries

University:

The University of British Columbia

Program: