Multilevel data compression techniques for transmission of audio over networks

In multimedia transmissions and activities, often more emphasis is placed on image and video, when compared to audio data. This is due to the fact that images and real-time video consume a lot more bandwidth than audio. However, with the advent of multimedia presentations on the network, and computer based multimedia instructions, there are several audio frames that need to be transmitted for a single image frame. In addition, if audio synchronization fails, the multimedia presentation will be erroneous. In teleconferencing and other similar endeavours, audio plays a very important role. As such, there is a definite need for audio compression methodologies, since several audio frames are needed for a single video or image frame, and this bandwidth needs to be optimized. Of the several methodologies available, differential encoding, analysis/synthesis schemes, and transform coding techniques are often used. In transform coding, the discrete cosine transform (DCT) and Fourier transform are more popular, although the discrete sine transform (DST) and wavelet transforms are gaining more and more acceptance. In order to obtain more compression and further optimize the audio bandwidth, the authors have investigated the performance of the audio compression methodologies, especially when they are nested for multilevel compression realizations. The authors also studied the effect of noise and signal quality reduction as a function of nested compressions. It was observed that up to five levels of nesting would produce a comparable compression ratio, with acceptable audio reproduction.