Mel Log Spectrum Approximation (MLSA) filter for speech synthesis

By Satoshi Imai, K. Sumita, C. Furuichi. 1983

How researchers see this paper

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

1. Xu Tan
2. Heiga Zen
3. Tomoki Toda
4. Junichi Yamagishi
5. Keiichi Tokuda
6. Tao Qin
7. Xin Eric Wang
8. Frank K. Soong

Citations in the corpus

• A Survey on Neural Speech Synthesis (2021, methodology): “, hidden Markov model (HMM) based) are trained with the paired linguistic features and parameters (acoustic features), where the acoustic features include fundamental frequency, spectrum or cepstrum [79, 349], etc, and are extracted from the speech through vocoder analysis [129, 152, 232].”
• Statistical Parametric Speech Synthesis (2007, methodology): “, mel log spectrum approximation (MLSA) filter (Imai et al., 1983)).”
• Fftnet: A Real-Time Speaker-Dependent Neural Vocoder (2018, background): “It is also worth noting that FFT has similar quality to the baseline method MLSA (insignificant p-value) due to noisy artifacts; this implies the training and synthesis techniques are essential to make FFTNet work well.”
• Integration of Spectral Feature Extraction and Modeling for HMM-Based Speech Synthesis (2014, methodology): “Mel-log spectrum approximation (MLSA) filter [27] is generally used to re-synthesize speech from the mel-cepstral coefficients.”
• Vocaine the vocoder and applications in speech synthesis (2015, unknown): “Attempts to make it faster may replace FFTs with an log-spectrum filter [19] or a variable lattice filter [20] and mixed-excitation [21].”
• An adaptive algorithm for mel-cepstral analysis of speech (1992, methodology): “The adaptive analysis system is implemented with an IIR adaptive filter which has the structure of the MLSA filter [ 4 ] and whose stability is guaranteed.”
• An investigation of multi-speaker training for wavenet vocoder (2017, methodology): “[22] Satoshi Imai, Kazuo Sumita, and Chieko Furuichi, “Mel log spectrum approximation (MLSA) filter for speech synthesis,” Electronics and Communications in Japan (Part I: Communications), vol. 66, no. 2, pp. 10–18, 1983.”