Speech representation and transformation using adaptive interpolation of weighted spectrum: vocoder revisited

By Hideki Kawahara. 1997

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, methodology): “TANDEM-STRAIGHT uses the periodic response directly, while Legacy-STRAIGHT manipulates the group delay to avoid buzzy timbre [31].”
ExcitNet Vocoder: A Neural Excitation Model for Parametric Speech Synthesis Systems (2018, methodology): “More recent WaveNet vocoders utilize acoustic parameters directly extracted from the speech such as mel-filterbank energy, mel-generalized cepstrum, BAP, and F0 [13]–[18].”
Restructuring speech representations using a pitch-adaptive time-frequency smoothing and an instantaneous-frequency-based F0 extraction: Possible role of a repetitive structure in sounds (1999, methodology): “In our previous papers [16, 15], a two-dimensional smoothing procedure was prpoosed to implement the basic idea.”
The STC text-to-speech system for Blizzard Challenge 2019 (2019, methodology): “A vocoder may be an iterative algorithm such as Grifﬁn-Lim [10] algorithm, an algorithm built on expert knowledge about speech signal (e.g. STRAIGHT [11], WORLD [12]), or a deep neural network such as WaveNet [13, 3], WaveRNN [14], FFTNet [15], LPCNet [16], ClariNet [17], WaveGlow [18].”
Aperiodicity extraction and control using mixed mode excitation and group delay manipulation for a high quality speech analysis, modification and synthesis system STRAIGHT (2001, methodology): “The F0 estimation method of STRAIGHT assumes that the signal has the following nearly harmonic structure. x(t) = N∑ k=1 ak(t) cos (∫ t 0 (kω0(τ) + ωk(τ))dτ + φk(0) ) , (2) where ak(t) represents a slowly changing instantaneous amplitude. ωk(τ) also represents slowly changing perturbation of the k-th”
Effective Spectral and Excitation Modeling Techniques for LSTM-RNN-Based Speech Synthesis Systems (2017, methodology): “To verify vocoding performance, additional DLSTM-based systems using the STRAIGHT (DLSTM-STRAIGHT) and WORLD (DLSTMWORLD) vocoders were also included [10], [11].”
Speech and melody recognition in binaurally combined acoustic and electric hearing. (2005, methodology): “In the present study, average fundamental frequency was 108 Hz for the tar 136 Hz for the male masker, and 219 Hz for the fem masker ~measured by theSTRAIGHT program, courtesy of Kawahara, 1997!.”