| 摘要 |
<p>The invention applies to a binaural audio playback system. According to the invention, prior to a step of interpolating the transfer functions of the head of a user of said system, the transfer functions obtained by measurements using a geometric representation of a model of the head of the user are decomposed, and subsequent to the interpolation step, the interpolated function is recomposed with said geometric representation. Advantageously, it can be possible to apply the decomposition and recomposition steps to only the phase or the amplitude of the transfer functions. Advantageously, the head model can be constructed by a simulation step or be chosen from an a priori list ofmodels. The invention therefore makes it possible to personalise the transfer functions on the basis of the features of a listener. The invention makes it possible to significantly improve the quality of the results of the interpolation and therefore the audio rendering of the playback system.</p> |
