Get our free extension to see links to code for papers anywhere online!

Chrome logo Add to Chrome

Firefox logo Add to Firefox
Picture for Tomohiko Nakamura

Tomohiko Nakamura

SelfRemaster: Self-Supervised Speech Restoration with Analysis-by-Synthesis Approach Using Channel Modeling


Mar 24, 2022
Takaaki Saeki, Shinnosuke Takamichi, Tomohiko Nakamura, Naoko Tanji, Hiroshi Saruwatari

Figure 1 for SelfRemaster: Self-Supervised Speech Restoration with Analysis-by-Synthesis Approach Using Channel Modeling
Figure 2 for SelfRemaster: Self-Supervised Speech Restoration with Analysis-by-Synthesis Approach Using Channel Modeling
Figure 3 for SelfRemaster: Self-Supervised Speech Restoration with Analysis-by-Synthesis Approach Using Channel Modeling
Figure 4 for SelfRemaster: Self-Supervised Speech Restoration with Analysis-by-Synthesis Approach Using Channel Modeling
* Submitted to INTERSPEECH 2022 
  Access Paper or Ask Questions

Differentiable Digital Signal Processing Mixture Model for Synthesis Parameter Extraction from Mixture of Harmonic Sounds


Feb 01, 2022
Masaya Kawamura, Tomohiko Nakamura, Daichi Kitamura, Hiroshi Saruwatari, Yu Takahashi, Kazunobu Kondo

Figure 1 for Differentiable Digital Signal Processing Mixture Model for Synthesis Parameter Extraction from Mixture of Harmonic Sounds
Figure 2 for Differentiable Digital Signal Processing Mixture Model for Synthesis Parameter Extraction from Mixture of Harmonic Sounds
Figure 3 for Differentiable Digital Signal Processing Mixture Model for Synthesis Parameter Extraction from Mixture of Harmonic Sounds
* 5 pages, 2 figures, to appear in 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2022) 
  Access Paper or Ask Questions

Speech Enhancement by Noise Self-Supervised Rank-Constrained Spatial Covariance Matrix Estimation via Independent Deeply Learned Matrix Analysis


Sep 10, 2021
Sota Misawa, Norihiro Takamune, Tomohiko Nakamura, Daichi Kitamura, Hiroshi Saruwatari, Masakazu Une, Shoji Makino

Figure 1 for Speech Enhancement by Noise Self-Supervised Rank-Constrained Spatial Covariance Matrix Estimation via Independent Deeply Learned Matrix Analysis
Figure 2 for Speech Enhancement by Noise Self-Supervised Rank-Constrained Spatial Covariance Matrix Estimation via Independent Deeply Learned Matrix Analysis
Figure 3 for Speech Enhancement by Noise Self-Supervised Rank-Constrained Spatial Covariance Matrix Estimation via Independent Deeply Learned Matrix Analysis
Figure 4 for Speech Enhancement by Noise Self-Supervised Rank-Constrained Spatial Covariance Matrix Estimation via Independent Deeply Learned Matrix Analysis
* accepted for APSIPA2021 
  Access Paper or Ask Questions

Multichannel Audio Source Separation with Independent Deeply Learned Matrix Analysis Using Product of Source Models


Sep 02, 2021
Takuya Hasumi, Tomohiko Nakamura, Norihiro Takamune, Hiroshi Saruwatari, Daichi Kitamura, Yu Takahashi, Kazunobu Kondo

Figure 1 for Multichannel Audio Source Separation with Independent Deeply Learned Matrix Analysis Using Product of Source Models
Figure 2 for Multichannel Audio Source Separation with Independent Deeply Learned Matrix Analysis Using Product of Source Models
Figure 3 for Multichannel Audio Source Separation with Independent Deeply Learned Matrix Analysis Using Product of Source Models
Figure 4 for Multichannel Audio Source Separation with Independent Deeply Learned Matrix Analysis Using Product of Source Models
* 8 pages, 5 figures, accepted for Asia-Pacific Signal and Information Processing Association Annual Summit and Conference 2021 (APSIPA ASC 2021) 
  Access Paper or Ask Questions

Prior Distribution Design for Music Bleeding-Sound Reduction Based on Nonnegative Matrix Factorization


Sep 01, 2021
Yusaku Mizobuchi, Daichi Kitamura, Tomohiko Nakamura, Hiroshi Saruwatari, Yu Takahashi, Kazunobu Kondo

Figure 1 for Prior Distribution Design for Music Bleeding-Sound Reduction Based on Nonnegative Matrix Factorization
Figure 2 for Prior Distribution Design for Music Bleeding-Sound Reduction Based on Nonnegative Matrix Factorization
Figure 3 for Prior Distribution Design for Music Bleeding-Sound Reduction Based on Nonnegative Matrix Factorization
Figure 4 for Prior Distribution Design for Music Bleeding-Sound Reduction Based on Nonnegative Matrix Factorization
* Accepted and will be presented at APSIPA2021 
  Access Paper or Ask Questions

Independent Deeply Learned Tensor Analysis for Determined Audio Source Separation


Jun 10, 2021
Naoki Narisawa, Rintaro Ikeshita, Norihiro Takamune, Daichi Kitamura, Tomohiko Nakamura, Hiroshi Saruwatari, Tomohiro Nakatani

Figure 1 for Independent Deeply Learned Tensor Analysis for Determined Audio Source Separation
Figure 2 for Independent Deeply Learned Tensor Analysis for Determined Audio Source Separation
Figure 3 for Independent Deeply Learned Tensor Analysis for Determined Audio Source Separation
* 5 pages, 2 figures, accepted for European Signal Processing Conference 2021 (EUSIPCO 2021) 
  Access Paper or Ask Questions

Empirical Bayesian Independent Deeply Learned Matrix Analysis For Multichannel Audio Source Separation


Jun 07, 2021
Takuya Hasumi, Tomohiko Nakamura, Norihiro Takamune, Hiroshi Saruwatari, Daichi Kitamura, Yu Takahashi, Kazunobu Kondo

Figure 1 for Empirical Bayesian Independent Deeply Learned Matrix Analysis For Multichannel Audio Source Separation
Figure 2 for Empirical Bayesian Independent Deeply Learned Matrix Analysis For Multichannel Audio Source Separation
Figure 3 for Empirical Bayesian Independent Deeply Learned Matrix Analysis For Multichannel Audio Source Separation
Figure 4 for Empirical Bayesian Independent Deeply Learned Matrix Analysis For Multichannel Audio Source Separation
* 5 pages, 4 figures, accepted for European Signal Processing Conference 2021 (EUSIPCO 2021) 
  Access Paper or Ask Questions

Sampling-Frequency-Independent Audio Source Separation Using Convolution Layer Based on Impulse Invariant Method


May 10, 2021
Koichi Saito, Tomohiko Nakamura, Kohei Yatabe, Yuma Koizumi, Hiroshi Saruwatari

Figure 1 for Sampling-Frequency-Independent Audio Source Separation Using Convolution Layer Based on Impulse Invariant Method
Figure 2 for Sampling-Frequency-Independent Audio Source Separation Using Convolution Layer Based on Impulse Invariant Method
Figure 3 for Sampling-Frequency-Independent Audio Source Separation Using Convolution Layer Based on Impulse Invariant Method
Figure 4 for Sampling-Frequency-Independent Audio Source Separation Using Convolution Layer Based on Impulse Invariant Method
* 5 pages, 3 figures, accepted for European Signal Processing Conference 2021 (EUSIPCO 2021) 
  Access Paper or Ask Questions

Time-Domain Audio Source Separation Based on Wave-U-Net Combined with Discrete Wavelet Transform


Jan 28, 2020
Tomohiko Nakamura, Hiroshi Saruwatari

Figure 1 for Time-Domain Audio Source Separation Based on Wave-U-Net Combined with Discrete Wavelet Transform
Figure 2 for Time-Domain Audio Source Separation Based on Wave-U-Net Combined with Discrete Wavelet Transform
Figure 3 for Time-Domain Audio Source Separation Based on Wave-U-Net Combined with Discrete Wavelet Transform
Figure 4 for Time-Domain Audio Source Separation Based on Wave-U-Net Combined with Discrete Wavelet Transform
* 5 pages, to appear in IEEE International Conference on Acoustics, Speech, and Signal Processing 2020 (ICASSP 2020) 
  Access Paper or Ask Questions

Real-Time Audio-to-Score Alignment of Music Performances Containing Errors and Arbitrary Repeats and Skips


Dec 24, 2015
Tomohiko Nakamura, Eita Nakamura, Shigeki Sagayama

Figure 1 for Real-Time Audio-to-Score Alignment of Music Performances Containing Errors and Arbitrary Repeats and Skips
Figure 2 for Real-Time Audio-to-Score Alignment of Music Performances Containing Errors and Arbitrary Repeats and Skips
Figure 3 for Real-Time Audio-to-Score Alignment of Music Performances Containing Errors and Arbitrary Repeats and Skips
Figure 4 for Real-Time Audio-to-Score Alignment of Music Performances Containing Errors and Arbitrary Repeats and Skips
* 12 pages, 8 figures, version accepted in IEEE/ACM Transactions on Audio, Speech, and Language Processing 
  Access Paper or Ask Questions

Outer-Product Hidden Markov Model and Polyphonic MIDI Score Following


Apr 08, 2014
Eita Nakamura, Tomohiko Nakamura, Yasuyuki Saito, Nobutaka Ono, Shigeki Sagayama

Figure 1 for Outer-Product Hidden Markov Model and Polyphonic MIDI Score Following
Figure 2 for Outer-Product Hidden Markov Model and Polyphonic MIDI Score Following
Figure 3 for Outer-Product Hidden Markov Model and Polyphonic MIDI Score Following
Figure 4 for Outer-Product Hidden Markov Model and Polyphonic MIDI Score Following
* Journal of New Music Research, Vol. 43, No. 2 (2014) 183-201 
 * 42 pages, 8 figures, version submitted to JNMR. To appear in Journal of New Music Research (2014) 
  Access Paper or Ask Questions