| 摘要 |
The present invention relates to encoding of an acoustic source signal (x) such that a corresponding signal ({circumflex over (z)}) reconstructed on basis of the encoded information has a perceptually high sound quality. The acoustic source signal (x) is encoded into at least one basic coded signal (P) by a coding arrangement in a transmitter (300). The at least one basic coded signal (P) thus represents perceptually significant characteristics of the acoustic signal (x). The coding arrangement in the transmitter (300) may include at least one spectral smoothing unit (305a), which receives at least one of the signal components on which the basic coded signal P is based and generates in response thereto a corresponding smoothed signal component. At least one enhanced coded signal P(E) is then produced from i.a. the corresponding smoothed signal component. The at least one enhanced coded signal P(E) thus constitutes a perceptually improved representation of the acoustic source signal (x), which is transmitted to a receiver (310) via a transmission medium (306). The receiver (310) receives at least one estimate ({circumflex over (P)}E) of the transmitted signal(s) (P(E)). A spectral smoothing unit (305b) in the receiver (310) produces, on basis of a primary spectrum (Ŷ) decoded from the at least one received estimate ({circumflex over (P)}E), a smoothed primary decoded spectrum (ŶE). Finally, this signal (ŶE) is used to generate an enhanced estimate ({circumflex over (z)}) of the acoustic source signal (x). The at least one smoothed signal component respective the smoothed primary decoded spectrum (ŶE) are produced by reducing a variation in coefficient values of the corresponding non-smoothed spectra representing frequency information above a threshold value (fT).
|
