TEMPORAL AND SPATIAL GENERATION OF MULTICHANNEL AUDIO SIGNALS Russian patent published in 2010 - IPC G10L19/00 

FIELD: physics, acoustics.

SUBSTANCE: invention relates to coding multichannel audio signals. A selected channel for a multichannel signal, which is presented through frames assembled from discretisation values having high time resolution, can be coded with high quality when parametric presentation of the signal waveform is obtained, which is the signal waveform in a presentation with intermediate resolution of the selected channel. Parametric presentation of the signal waveform includes a sequence of intermediate parametres of the signal waveform, having intermediate time resolution below the high time resolution of discretisation values and higher the low time resolution given by frame recurrence rate. Parametric presentation of the signal waveform with intermediate resolution can be used to generate a reconstructed channel to obtain a channel having a signal envelope closer to the envelope of the selected initial channel. The time scale on which generation takes place is shorter than the time scale of frame by frame processing, that way increasing quality of the reconstructed channel. Conversely, the time scale for generation is higher than the time scale for discretisation values, which considerably reduces amount of data required by parametric presentation of the signal waveform.

EFFECT: improved spatial perception of the reconstructed multichannel signal.

37 cl, 16 dwg