Transference and retrieval of compress voice message over low signal strength in satellite communication

摘要

This paper provides an alternative way of sending voice using satellite communication system through SMS. Voice transferring method over SMS is useful in situations when signal strength is low and due to poor signal strength voice call connection is not possible to initiate or signal is dropped during voice call. So we record voice, apply adaptive Wiener filter to remove background noise, detect voice segments using voice activity detector (VAD) and discard unvoiced segment of voice and after this apply compression. We compare six different types of compression algorithms; Lempel-Ziv-Welch (LZW), Huffman coding, prediction by partial matching (PPM), arithmetic coding (AC), Burrows-Wheeler-transform (BWT) and Lempel-Ziv-Markov chain (LZMA). Each compression scheme achieves an average compression ratio of 16.22, 10.22, 25.33, 14.77, 21 and 18.44 respectively. For experimentation, first we develop an application using Java 2 Platform, Micro Edition (J2ME) language and use Thuraya SG-2520 satellite phone. Proposed algorithm performs approximately 20% better if we use adaptive Wiener filter with PPM compression scheme instead LZW scheme.