Analysis of FLAC Music Pieces Recovery

Abstract

In this paper, we propose a method of recovering the music piece of FLAC file downloaded by BitTorrent. It was decoded the split FLAC file via getting the complete frames from piece of FLAC file and adding the temporary header to decode. The percentage of successfully decoding is all above 90%. From the partial audio signal we extract the musical feature and calculate the cross-correlation with the features in database. The feature in database is extract from the complete music file. The Musical feature is detected by chromagram to get the key strength. The percentage of audio identification is growing tendency with the increase of piece size. It experiment shows probability of identify audio with getting the part of FLAC file distributed by BitTorrent.