Pattern Matching Techniques for Replacing Missing Sections of Audio Streamed across Wireless Networks

摘要

Streaming media on the Internet can be unreliable. Services such as audio-on-demand drastically increase the loads on networks; therefore, new, robust, and highly efficient coding algorithms are necessary. One method overlooked to date, which can work alongside existing audio compression schemes, is that which takes into account the semantics and natural repetition of music. Similarity detection within polyphonic audio has presented problematic challenges within the field of music information retrieval. One approach to deal with bursty errors is to use self-similarity to replace missing segments. Many existing systems exist based on packet loss and replacement on a network level, but none attempt repairs of large dropouts of 5 seconds or more. Music exhibits standard structures that can be used as a forward error correction (FEC) mechanism. FEC is an area that addresses the issue of packet loss with the onus of repair placed as much as possible on the listener's device. We have developed a server-client-based framework (SoFI) for automatic detection and replacement of large packet losses on wireless networks when receiving time-dependent streamed audio. Whenever dropouts occur, SoFI swaps audio presented to the listener between a live stream and previous sections of the audio stored locally. Objective and subjective evaluations of SoFI where subjects were presented with other simulated approaches to audio repair together with simulations of replacements including varying lengths of time in the repair give positive results.