Tagungsband - DAGA 2024
Beitrag
On the Use of Dereverberation Algorithms in Binaural Cue Adaptation
Zusammenfassung:
Binaural Cue Adaptation (BCA) is a signal processing technique for modifying the position of sound sources in a binaural audio signal. In combination with a head-tracking device, it can be used to compensate a listener’s head movements, thus enabling a more immersive reproduction of binaural recordings.Previous publications have assumed an input signal containing a directional component and a temporally uncorrelated ambient component. To modify only those components containing directional information, BCA algorithms have decomposed binaural recordings using a primary-ambient decomposition (PAD). However, this signal model is not well suited to reverberant environments. Therefore, this contribution extends the signal model to include a temporally correlated reverberation component. Building upon this signal model, we propose a BCA algorithm that incorporates the weighted predicted error (WPE) dereverberation algorithm into the PAD. The results of a MUSHRA listening test show that the proposed approach outperforms the baseline in reverberant environments in terms of the perceived quality of the reverberation and the overall signal.