@inproceedings{liberatore2018icassp,

title = {Voice Conversion through Residual Warping in a Sparse, Anchor-Based Representation of Speech},

author = {C Liberatore and G Zhao and R Gutierrez-Osuna},

url = {https://psi.engr.tamu.edu/wp-content/uploads/2018/03/liberatore-icassp2018.pdf},

year  = {2018},

date = {2018-04-15},

booktitle = {Proc. ICASSP},

abstract = {In previous work we presented a Sparse, Anchor-Based Representation of speech (SABR) that uses phonemic “anchors” to represent an utterance with a set of sparse non-negative weights. SABR is speaker-independent: combining weights from a source speaker with anchors from a target speaker can be used for voice conversion. Here, we present an extension of the original SABR that significantly improves voice conversion synthesis. Namely, we take the residual signal from the SABR decomposition of the source speaker’s utterance, and warp it to the target speaker’s space using a weighted warping function learned from pairs of source-target

anchors. Using subjective and objective evaluations, we examine the performance of adding the warped residual (SABR+Res) to the original synthesis (SABR). Specifically, listeners rated SABR+Res with an average mean opinion score (MOS) of 3.6, a significant improvement compared to 2.2 MOS for SABR alone (𝑝 < 0.01) and

2.5 MOS for a baseline GMM method (𝑝 < 0.01). In an XAB speaker identity test, listeners correctly identified the identity of SABR+Res (81%) and SABR (84%) as requently as a GMM method (82%) (𝑝 = 0.70, 𝑝 = 0.35). These results indicate that adding the warped residual can dramatically improve synthesis while retaining the 

 desirable independent qualities of SABR models.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{liberatore2016interspeech,

title = {Generating Gestural Scores from Acoustics Through a Sparse Anchor-Based Representation of Speech},

author = {C Liberatore and R Gutierrez-Osuna},

url = {https://psi.engr.tamu.edu/wp-content/uploads/2018/01/liberatore2016interspeech.pdf},

year  = {2016},

date = {2016-09-08},

booktitle = {Proc. Interspeech},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{liberatore2015interspeech,

title = {SABR: Sparse, Anchor-Based Representation of the Speech Signal},

author = {C Liberatore and S Aryal and Z Wang and S Polsley and R Gutierrez-Osuna},

url = {https://psi.engr.tamu.edu/wp-content/uploads/2018/01/liberatore2015interspeech.pdf},

year  = {2015},

date = {2015-09-06},

booktitle = {Proc. Interspeech 2015},

pages = {608-612},

abstract = {We present SABR (Sparse, Anchor-Based Representation), an analysis technique to decompose the speech signal into 

speaker-dependent and speaker-independent components. Given a collection of utterances for a particular 

speaker, SABR uses the centroid for each phoneme as an acoustic “anchor,” then applies Lasso regularization to 

the method on a speaker-independent phoneme recognition task and a voice conversion task. Using a linear classifier, 

SABR weights achieve significantly higher phoneme recognition rates than Mel frequency Cepstral coefficients. SABR weights can also be used directly to perform accent conversion without the need to train a speakerto-speaker regression model.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{liberatore2015icassp,

title = {Joint Optimization of Anatomical and Gestural Parameters in a Physical Vocal Tract Model},

author = {C Liberatore and R Gutierrez-Osuna},

url = {https://psi.engr.tamu.edu/wp-content/uploads/2018/01/liberatore2015icassp.pdf},

year  = {2015},

date = {2015-04-19},

booktitle = {ICASSP},

organization = {IEEE},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}