Research And Realization Of Parametric Stereo Coding

To store and transmit stereo audio signals efficiently, they should be compressed by stereo codec. Compared with traditional stereo coding, parametric stereo coding ensures higher coding gain with same quality.Parametric stereo coding is based on spatial hearing. The original stereo signal, which is represented by mono signal and spatial parameters in encoder, can be synthesized in decoder. The basic information of the rebuilt sound field is derived from mono signal, and the property of spatial images is controlled by spatial parameters. Hence it's very important to choose the right parameters to encode in parametric stereo coding.With thoroughly research of spatial hearing, proper parameters are presented in this thesis. A model of parametric stereo coding is introduced using the selected parameters. All kinds of possible parameters are included in the model. Those so-called new parameters can actually be obtained by the parameters in the model. It suits well for the existing parametric stereo coding algorithms.The model is realized, and the details of realization are given. After typical algorithm is analyzed, some possible methods to improve are listed. The scalable bit stream described in the thesis enables bit-rate variable.To overcome the shortcomings of existing parametric stereo coder, some new parameters are presented. A coder using these parameters is designed and it's based on FFT. MUSHRA tests show that the coder has almost the same performance as aacPlus v2, and it's score is even higher when the coherence between channels is low.The work of this thesis can also be dedicated to multi-channel audio coding, where the number of channels is greater than 2.