Welcome to the web presence of MusicTrace
 


Top News:

Airplay charts compiled by MusicTrace now available on radiocharts.com
radiocharts.com

Audio Fingerprinting

General

Audio fingerprinting technology is capable of recognizing audio signals on the basis of particular characteristics. These characteristics are unique for each audio signal and, by analogy with human fingerprints, are also referred to as fingerprints. A music recording or commercial can only be recognized if these characteristics were previously registered and then entered in a special database.

In addition to acquisition of these characteristics, no further processing of the audio signal is necessary. When audio fingerprinting technology is implemented, the audio signal itself is not modified, in particular no additional information are incorporated into the audio signal. Recognition of the title is performed exclusively on the basis of content, i.e., only using the characteristics derived from the audio track.

The audio fingerprinting technique implemented by MusicTrace identifies audio material using a robust comparison method and previously registered contents. This means that the title that is to be identified is recognized even if the audio signal has undergone a certain degree of modification. Such modifications may include, for example, linear disruption such as level changes or bandwidth limitation such as those encountered in the case of radio broadcasting. Other modifications include non-linear disturbances, for example encoding as MP3 format. Finally, the system can even recognize pieces of audio material that are incomplete.

In addition, the system can also distinguish between various versions of a particular music recording. This aspect is not merely restricted to distinguishing between the normal and live version of a recording; the system is, for example, also capable of distinguishing various concert recordings of a particular piece of classical music.

The decisive factor for implementation of an audio fingerprinting process is selection of the characteristics that are to be investigated. Distinguishing between a very large number of titles (over 1 million) is only possible if suitable characteristics are selected. At the same time, selection of characteristics has an impact on the size of the fingerprint of a given title and thus directly affects the time required to identify a title.

Top top         Back back