Research focus on audio and speech processing is concerned with fundamental concepts and applications of signal processing techniques and information theory to processing of speech and audio signals.
Audio compression: the objective is to find a compact representation of audio signals (such as music) with as little quality loss as possible. Methods better than mp3 and aac have been developed in the group.
Speech enhancement: the goal is to remove interfering noise and other degrading signals from a speech signal, such that understandability or automatic speech recognition is improved.
Fingerprinting: in order to compare a given piece of music against a large database of songs, one could use crosscorrelation. A fingerprint is a very compact representation of songs that makes the matching process far more efficient. The technology used for instance by MusicDNA is researched in order to exploit fingerprints for more than matching, for instance for quality assessment.
Watermarking: a watermark is a imperceptible modification to a (music) signal in such as way that it can be used for identifying a particular song. The embedded information can be used for instance for tracing content over the Internet. Our focus is on improving the robustness of the scalar costa watermarking scheme (SCS, or DC-QIM).
Meer weten?
Wil je meer informatie over de inhoud van het dossier of in contact worden gebracht met de kennisaanbieder? Neem dan contact met ons op.