Digital Audio - Compression of audio data

Leggi questa pagina in Italiano Lire cette page en Franšais
CDROM Multimedia Audio Course Enjoying this Course?
Download the full version!

Lossy compression algorithms are based on the knowledge of the characteristics of the data to be compressed. Depending on the kind of data at hand, it is possible to pin-point recurring patterns which can be used to carry out the compression. To illustrate this let's imagine that we need to compress a video. Without a doubt we'd make the most of the fact that in a sequence of photograms, contiguous images differ very slightly from one another (that's unless the shot-perspective changes of course). So, with a hypothetical compression algorithm we'd store only the data that has changed from one image to the next. When it comes to compressing audio signal data, we need to take advantage of different characteristics. Most compression algorithms for audio signals are based on the masking principle, namely, on the fact that if a certain frequency of the signal at hand has a sufficiently high amplitude, it effectively masks the adjacent frequencies when they have a lower amplitude [frequency masking ] .

The following diagram illustrates a signal with its various frequencies and a human ear's hearing threshold rate in relation to the frequencies with higher amplitudes:

Digital Audio - Signal to be compressed

Signal to be compressed

The following diagram shows what results when masked frequencies are eliminated by the compression. As we can see, the amount of information to store has considerably diminished.

Digital Audio - Compressed signal

Compressed signal

The most common algorithms that operate on the aforementioned criteria are the ATRAC and MP3 (usually used for transferring audio files on the Internet). The first of the two will be described in the section on Mini-disc [The MiniDisc ] . With MP3s, compression is carried out by using an algorithm called MPEG[35 ] 1 Layer III[36 ] which operates using the masking principle. It allows 12:1 compression-ratio values.



[35 ] MPEG is the acronym for Moving Picture Expert Group. This is a work-group that operates under the direction of ISO (International Standard Organization) and IEC (International Electro-Technical Commission). The group's aim is to create image and sound compression standards for the web which occupy as little bandwidth as possible.

[36 ] Don't mistake the MP3 (MPEG 1 Layer III) with the MPEG3 which is a video signal compression system.