The emergence and success of ‘internet music’ as a viable internet ‘media’, has been driven by major advances in sound data file compression. Data compression can simply defined as ‘storing data in a format that requires less space than usual’.

Songs (unless purely created in the digital domain) will have to be translated into a digital form, before they can be transmitted along the internet. To ‘digitise’ music, the sounds must be ‘digitised’, by sampling the music at discrete intervals using an analogue to digital converter (ADC). The ADC measures a sound wave's amplitude many times per second, it is these numeric values that can be recorded digitally. Digitising music in this way, using a ‘CD quality’ sampling rate i.e. 44.1 Khz at 16 bit resolution in stereo, equates to a 1400 Kbps ADC bit rate. This will result in data files that are approximately 300 Kb in size per 4 minutes (typical song) of digitised audio. Using a 56.6 Kbps modem, a 4 minute song would take 6 minutes of downloading time, an average album would take over an hour. It can easily be appreciated that compression technology would need to be employed to make the transmission of digitised music files viable across the infrastructure of the internet, especially for the modem connected ‘home’ user.

There are a variety of sound data file compression techniques, but only a few have been standardised. By far the most popular compression technique employed in internet music is ‘MP3’. MP3 is the file extension for the compression technique defined by the ‘moving pictures expert group (MPEG) audio layer 3’ standard (see http://www.cselt.it/mpeg/). Layer 3 uses perceptual audio coding and psychoacoustic compression to remove all superfluous information (more specifically, the redundant and irrelevant parts of a sound signal – mainly the components that the human ear doesn't hear). It also adds a MDCT (Modified Discrete Cosine Transform) that implements a filter bank, increasing the frequency resolution 18 times higher than that of the older layer 2 standard.

Because MP3 files are small, they can easily be transferred across the internet. Surprisingly, this great loss of data has no noticeable effect on the quality of the audio file. That’s the simple reason why MP3 has become so popular. MP3 is capable of providing CD-quality sound with approximately one-twelfth of the amount of data storage. As long as the MP3 files are recorded at a rate of 128Kbps or greater, they’ll come very close to matching the audio quality of a CD. Using MP3 compression, at this high quality setting, a typical CD album can be compressed to a file approximately 3.5 MB in size (as compared to an uncompressed size of 42 MB).

Another emerging data compression technology is ‘liquid audio’ (www.liquidaudio.com), this standard uses ‘lossless algorithms’ and is currently being widely accepted as an alternative to MP3 compression as it actively limits, to a greater degree than MP3, the audible component of the data removed during the compression process.

Further links of interest on compression technologies:-

‘The Compression Technology in Multimedia’ at http://www.scit.wlv.ac.uk/~c9581158/main.html

‘Compression pointers’ at http://www.internz.com/compression-pointers.html

‘Mitsuharu Arimura’s Bookmarks on source coding data compression’ at http://www.hn.is.uec.ac.jp/~arimura/compression_links.html#audio