摘要 |
<p>The invention relates to a method for hierarchical encoding of an audio-digital signal, comprising: - a core encoding step (400, 500) producing N scalar indices (I B (n) n=0,..., N -1) for quantifying a set of N samples (x(n) n=0,..., N-1) of an input signal; - a step of obtaining (480, 580) masking filter coefficients; - at least one rank k enhancement encoding step producing N scalar indices (J k (n) n=0,..., N-1) for quantification (Q k enh ), such that said enhancement encoding step comprises: - a step of determining (300, 600) enhancement levels for forming the vectors of an enhancement dictionary (Dic Tj B+k (n) j = 0,..., 2 N-1 ; n = 0,..., N -1) from at least the N core quantification indices (I B (n) n = 0,..., N- 1) and the indices for enhancing the levels of lower ranks (J 1 (n),..., J k-1 (n) n = 0,..., N -1); - a step of searching (310, 340, 610), in the enhancement dictionary producing the N indices (J k (n) n=0, N -1), for the rank k enhancement code by minimising the energy of a filtered error signal (d jw b+k (n)) obtained using the coefficients of the masking filter. The invention also relates to the encoder implementing the method.</p> |
申请人 |
FRANCE TELECOM;KOVESI, BALAZS;RAGOT, STEPHANE;LE GUYADER, ALAIN |
发明人 |
KOVESI, BALAZS;RAGOT, STEPHANE;LE GUYADER, ALAIN |