The Mel-Frequency Cepstral Coefficients (MFCC) feature extraction method is a leading approach for speech feature extraction and current research aims to identify performance enhancements. One of the recent MFCC implementations is the Delta-Delta MFCC, which improves speaker verification. In this paper, a new MFCC feature extraction method based on distributed Discrete Cosine Transform (DCT-II) is presented. Speaker verification tests are proposed based on three different feature extraction methods including: conventional MFCC, Delta-Delta MFCC and distributed DCT-II based Delta-Delta MFCC with a Gaussian Mixture Model (GMM) classifier.

A novel approach for MFCC feature extraction

A speech coding method in which spectrum parameter representing a spectrum envelope and a pitch parameter representing a pitch are obtained from an input discrete speech signal. A frame interval is divided into subintervals in accordance with the pitch parameter. A sound source signal in one of the subintervals is obtained by obtaining a multipulse with respect to a difference signal obtained by performing prediction on the basis of a past sound source signal. Correction information for correcting at least one of the amplitude and the phase of the sound source signal are obtained and output in other pitch intervals in the frame.

Speech coding/decoding method having an excitation signal

A phonetically based segmentation of speech is performed to classify segments into five classes: onset, unvoiced low-pass voiced, steady-state voiced, and transient voiced. The segment lengths are constrained to an integer multiple of a unit-frame. For each segment class, a distinctive coding scheme based on vector excitation coding (VXC) is used. The maximum bit-rate is 3.6 kb/s, and a moderate coding delay of 45 ms is incurred. Performance is roughly comparable to conventional VXC/CELP (code-excited linear prediction) coding at 4.8 kb/s. >

Phonetically-based vector excitation coding of speech at 3.6 kbps

A speech coding system providing reconstructed voiced speech with a smoothly evolving pitch-cycle waveform. A speech signal is represented by isolating and coding prototype waveforms. Each prototype waveform is an exemplary pitch-cycle of voiced speech. A coded prototype waveform is transmitted at regular intervals to a receiver which synthesizes (or reconstructs) an estimate of the original speech segment based on the prototypes. The estimate of the original speech signal is provided by a prototype interpolation process which provides a smooth time-evolution of pitch-cycle waveforms in the reconstructed speech. Illustratively, a frame of original speech is coded by first filtering the frame with a linear predictive filter. Next a pitch-cycle of the filtered original is identified and extracted as a prototype waveform. The prototype waveform is then represented as a set of Fourier series (frequency domain) coefficients. The pitch-period and Fourier coefficients of the prototype, as well as the parameters of the linear predictive filter, are used to represent a frame of original speech. These parameters are coded by vector and scalar quantization and communicated over a channel to a receiver which uses information representing two consecutive frames to reconstruct the earlier of the two frames based on a continuous prototype waveform interpolation process. Waveform interpolation may be combined with conventional CELP techniques for coding unvoiced portions of the original speech signal.

Prototype waveform speech coding with interpolation of pitch, pitch-period waveforms, and synthesis filter

Methods for waveform interpolation in speech coding

2.4-kb/s speech coding based on pitch-prediction multipulse speech coding is described. An adaptive segmentation procedure to effectively control the renewal rate for synthesis model parameters, a linear time-varying pitch-synthesis filter, a model for multipulse excitation during one pitch period, and a vector quantization for linear predictive coding parameters are introduced. Subjective evaluation demonstrates that pitch-prediction multiphase speech coding can provide much more natural-sounding synthetic speech than single-pulse-excitation speech coding (as with a vocoder) or pitch-interpolation multipulse speech coding at 2.4-kb/s. >

2.4 kbps pitch prediction multi-pulse speech coding

This paper discusses the possibility of an input-dependent orthogonal transform for low bit rate speech coding. An orthogonal transform, based on an autoregressive model of short-time input speech, which is applicable to low bit rate coding, is presented. Its performance is evaluated in comparison with that for conventional transforms, KLT and DCT. The experiments confirm that the input-dependent orthogonal transform improves the average distortion versus average information rate performance over that for the input-independent transform DCT.

S. Ono

Papers

2.4 kbps pitch prediction multi-pulse speech coding

Linear transformation for low bit rate speech coding