Fundamental frequency (F0) and energy of speech represented by pitch and gain contours are two fundamental concepts of excitation signals in model-based speech coding systems. To compress these parameters in very low-rate speech coding where the simplest form of excitation is used, only differential pitch and gain trackers are used at the expense of a perceivable distortion. To alleviate this problem, a new technique is proposed which effectively compresses the pitch and the gain information using temporal decomposition (TD) based on the relationship between vocal tract characteristics, fundamental frequency, and speech energy.
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.