Title :
Spectral envelope extension of narrowband speech using neural networks
Author :
Bulu, Gürhan ; Bingöl, A. Semih
Author_Institution :
Elektrik ve Elektron. Muhendisligi Bolumu, Hacettepe Univ., Ankara, Turkey
Abstract :
This paper describes a Neural Network (NN) based method that extends the bandwidth of narrowband speech to obtain the spectral envelope of the wideband speech. Line Spectral Frequencies (LSF) which are obtained from Linear Prediction Coefficients (LPC) are used to represent the spectral envelope. Spectral distortion measure is used to test the performance of the method.
Keywords :
neural nets; spectral analysis; speech processing; line spectral frequencies; linear prediction coefficients; narrowband speech; neural network based method; spectral distortion measure; spectral envelope extension; Bandwidth; Conferences; Markov processes; Speech; Speech coding; Speech processing;
Conference_Titel :
Signal Processing and Communications Applications (SIU), 2011 IEEE 19th Conference on
Conference_Location :
Antalya
Print_ISBN :
978-1-4577-0462-8
Electronic_ISBN :
978-1-4577-0461-1
DOI :
10.1109/SIU.2011.5929849
