A hybrid speech codec for VoIP telephony applications is presented employing combined parametric and perceptual coding techniques. The signal is divided into voiced signal components that are encoded using the perceptual algorithm, unvoiced components that are encoded parametrically and transients that are not encoded with a lossy method. The codec architecture where voiced part of the CELP residual signal is perceptually encoded and transmitted to the decoder along with the CELP main bit stream is also examined. Various methods for transient detection in the speech signal are discussed. The results of experiments revealing the improved subjective quality of the transmitted speech are also presented.
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member and would like to subscribe to the E-Library then Join the AES!
This paper costs $33 for non-members and is free for AES members and E-Library subscribers.