Digital Representation of Audio

Abstract: This report focuses on lossless compression of audio, which is usually built on predictioncoding. The aim of the report is to summarize five lossless compression algorithms, presentand compare their methods of compression, and propose what a future format could look like.Information is gathered and presented in the form integrative literature review, with snowballsampling starting from a convenience sample. The results are presented in a timeline and atable analyzing the formats’ methods of compression. The study finds that older formats werebuilt to minimize the usage of the limited processing power and still deliver high compressionrate. Newer formats use heavier, but more effective compression algorithms and predictionmethods. The results show that prediction coding via the LPC (Linear Predictive Coding)method will probably not be the preferred method in the future as the compression rate-gaindiminishes with higher orders of prediction. The prediction presented in this study is that in thefuture lossless formats will use predictive coding, where sine-functions will work as predictors.The study also predicts that compression of PCM audio will not be interesting for long, astechnological advancements will render the compression redundant when storage capacity andbandwidth grows. Instead, it predicts that the representation of analog signals in a digitalenvironment will be the new frontier in audio research and proposes the BEA format, whereanalog signals are stored as Sine functions.