A mixed excitation LPC vocoder model for low bit rate speech coding

By Alan McCree, T. Barnwell. 1995

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

WORLD: A Vocoder-Based High-Quality Speech Synthesis System for Real-Time Applications (2016, background): “Mixed excitation [27] and aperiodicity [28] have usually been used to synthesize natural speech.”
Neural Source-Filter Waveform Models for Statistical Parametric Speech Synthesis (2019, methodology): “Digital filters have been used to merge the noise and periodic signals in the classical speech modeling methods [47], [48].”
Sequence to Sequence Neural Speech Synthesis with Prosody Modification Capabilities (2019, background): “Here f (c) is derived from the common “peakiness” metric [15], i.”
Vocaine the vocoder and applications in speech synthesis (2015, methodology): “A comparison was made against a MATLAB implementation of STRAIGHT and optimized embedded production mixed-excitation implementations with LSP (Line Spectrum Pairs) and MCEP (Mel-Cepstrum) parameterizations [21].”
Effective Spectral and Excitation Modeling Techniques for LSTM-RNN-Based Speech Synthesis Systems (2017, background): “By separating the whole frequency band into several fixed sub-bands, the excitation signal of each sub-band is represented by either PoN or band aperiodicities (BAP) [9]–[11].”
Simultaneous Modeling of Spectrum, Pitch and Duration in HMM-Based Speech Synthesis (1999, unknown): “例えば,本システムにおいて音源は単純に, 有声区間に関してはパルス列,無声区間に関しては白色雑音を用いているが,音源情報も HMM の枠組みでモデル化し,音声符号化の技術である MELP [15] などで用いられている手法と同様の音源生成手法を導入することにより,音質は更に向上すると考えられる (連続値,離散値を含む音源情報はMSD-HMMによりモデル化可能であることに注意する).”
Robust Speaker-Adaptive HMM-Based Text-to-Speech Synthesis (2009, background): “To reduce buzziness, mixed or multi-band excitation techniques [15]–[17] have been integrated into the basic systems to replace the simple pulse or noise excitation and have been evaluated [18]–[21].”
Unified Source-Filter GAN: Unified Source-filter Network Based On Factorization of Quasi-Periodic Parallel WaveGAN (2021, methodology): “LPCNet [23] adopts a WaveRNN [19]-like architecture to generate residual signals while a linear filtering process is applied to generate speech waveforms as in the conventional linear predictive coding (LPC) vocoder [31, 32].”
An HMM-based speech synthesis system applied to English (2002, methodology): “Although synthesized speech has a typical quality of “vocoded speech,” it has been shown in [15] that the mixed excitation model based on MELP speech coder [20] and postfiltering can improve the speech quality significantly.”
Introduction to Digital Speech Processing (2007, background): “Several new parameters of the excitation must be estimated at analysis time and coded for transmission, but these add only slightly to either the analysis computation or the bit rate [80].”