RESUMEN

Hoarse voice affects the efficiency of communication between people. However, surgical treatment may result in patients with poorer voice quality, and voice repair techniques can only repair vowels. In this paper, we propose a novel multidomain generative adversarial voice conversion method to achieve hoarse-to-normal voice conversion and personalize voices for patients with hoarseness. The proposed method aims to improve the speech quality of hoarse voices through a multidomain generative adversarial network. The proposed method is evaluated on subjective and objective evaluation metrics. According to the findings of the spectrum analysis, the suggested method converts hoarse voice formants more effectively than variational auto-encoder (VAE), Auto-VC (voice conversion), StarGAN-VC (Generative Adversarial Network- Voice Conversion), and CycleVAE. For the word error rate, the suggested method obtains absolute gains of 35.62, 37.97, 45.42, and 50.05 compared to CycleVAE, StarGAN-VC, Auto-VC, and VAE, respectively. The suggested method achieves CycleVAE, VAE, StarGAN-VC, and Auto-VC, respectively, in terms of naturalness by 42.49%, 51.60%, 69.37%, and 77.54%. The suggested method outperforms VAE, CycleVAE, StarGAN-VC, and Auto-VC, respectively, in terms of intelligibility, with absolute gains of 0.87, 0.93, 1.08, and 1.13. In terms of content similarity, the proposed method obtains 43.48%, 75.52%, 76.21%, and 108.62% improvements compared to CycleVAE, StarGAN-VC, Auto-VC, and VAE, respectively. ABX results show that the suggested method can personalize the voice for patients with hoarseness. This study demonstrates the feasibility of voice conversion methods in improving the speech quality of hoarse voices.

RESUMEN

One-shot voice conversion (VC) aims to convert speech from any source speaker to an arbitrary target speaker with only a few seconds of reference speech from the target speaker. This relies heavily on disentangling the speaker's identity and speech content, a task that still remains challenging. Here, we propose a novel approach to learning disentangled speech representation by transfer learning from style-based text-to-speech (TTS) models. With cycle consistent and adversarial training, the style-based TTS models can perform transcription-guided one-shot VC with high fidelity and similarity. By learning an additional mel-spectrogram encoder through a teacher-student knowledge transfer and novel data augmentation scheme, our approach results in disentangled speech representation without needing the input text. The subjective evaluation shows that our approach can significantly outperform the previous state-of-the-art one-shot voice conversion models in both naturalness and similarity.

RESUMEN

Voice-controlled devices are in demand due to their hands-free controls. However, using voice-controlled devices in sensitive scenarios like smartphone applications and financial transactions requires protection against fraudulent attacks referred to as "speech spoofing". The algorithms used in spoof attacks are practically unknown; hence, further analysis and development of spoof-detection models for improving spoof classification are required. A study of the spoofed-speech spectrum suggests that high-frequency features are able to discriminate genuine speech from spoofed speech well. Typically, linear or triangular filter banks are used to obtain high-frequency features. However, a Gaussian filter can extract more global information than a triangular filter. In addition, MFCC features are preferable among other speech features because of their lower covariance. Therefore, in this study, the use of a Gaussian filter is proposed for the extraction of inverted MFCC (iMFCC) features, providing high-frequency features. Complementary features are integrated with iMFCC to strengthen the features that aid in the discrimination of spoof speech. Deep learning has been proven to be efficient in classification applications, but the selection of its hyper-parameters and architecture is crucial and directly affects performance. Therefore, a Bayesian algorithm is used to optimize the BiLSTM network. Thus, in this study, we build a high-frequency-based optimized BiLSTM network to classify the spoofed-speech signal, and we present an extensive investigation using the ASVSpoof 2017 dataset. The optimized BiLSTM model is successfully trained with the least epoch and achieved a 99.58% validation accuracy. The proposed algorithm achieved a 6.58% EER on the evaluation dataset, with a relative improvement of 78% on a baseline spoof-identification system.

Asunto(s)

RESUMEN

Voice conversion (VC) consists of digitally altering the voice of an individual to manipulate part of its content, primarily its identity, while maintaining the rest unchanged. Research in neural VC has accomplished considerable breakthroughs with the capacity to falsify a voice identity using a small amount of data with a highly realistic rendering. This paper goes beyond voice identity manipulation and presents an original neural architecture that allows the manipulation of voice attributes (e.g., gender and age). The proposed architecture is inspired by the fader network, transferring the same ideas to voice manipulation. The information conveyed by the speech signal is disentangled into interpretative voice attributes by means of minimizing adversarial loss to make the encoded information mutually independent while preserving the capacity to generate a speech signal from the disentangled codes. During inference for voice conversion, the disentangled voice attributes can be manipulated and the speech signal can be generated accordingly. For experimental evaluation, the proposed method is applied to the task of voice gender conversion using the freely available VCTK dataset. Quantitative measurements of mutual information between the variables of speaker identity and speaker gender show that the proposed architecture can learn gender-independent representation of speakers. Additional measurements of speaker recognition indicate that speaker identity can be recognized accurately from the gender-independent representation. Finally, a subjective experiment conducted on the task of voice gender manipulation shows that the proposed architecture can convert voice gender with very high efficiency and good naturalness.

RESUMEN

Silent speech interfaces (SSIs) convert non-audio bio-signals, such as articulatory movement, to speech. This technology has the potential to recover the speech ability of individuals who have lost their voice but can still articulate (e.g., laryngectomees). Articulation-to-speech (ATS) synthesis is an algorithm design of SSI that has the advantages of easy-implementation and low-latency, and therefore is becoming more popular. Current ATS studies focus on speaker-dependent (SD) models to avoid large variations of articulatory patterns and acoustic features across speakers. However, these designs are limited by the small data size from individual speakers. Speaker adaptation designs that include multiple speakers' data have the potential to address the issue of limited data size from single speakers; however, few prior studies have investigated their performance in ATS. In this paper, we investigated speaker adaptation on both the input articulation and the output acoustic signals (with or without direct inclusion of data from test speakers) using the publicly available electromagnetic articulatory (EMA) dataset. We used Procrustes matching and voice conversion for articulation and voice adaptation, respectively. The performance of the ATS models was measured objectively by the mel-cepstral distortions (MCDs). The synthetic speech samples were generated and are provided in the supplementary material. The results demonstrated the improvement brought by both Procrustes matching and voice conversion on speaker-independent ATS. With the direct inclusion of target speaker data in the training process, the speaker-adaptive ATS achieved a comparable performance to speaker-dependent ATS. To our knowledge, this is the first study that has demonstrated that speaker-adaptive ATS can achieve a non-statistically different performance to speaker-dependent ATS.

Asunto(s)

RESUMEN

Voice conversion has made great progress in the past few years under the studio-quality test scenario in terms of speech quality and speaker similarity. However, in real applications, test speech from source speaker or target speaker can be corrupted by various environment noises, which seriously degrade the speech quality and speaker similarity. In this paper, we propose a novel encoder-decoder based noise-robust voice conversion framework, which consists of a speaker encoder, a content encoder, a decoder, and two domain adversarial neural networks. Specifically, we integrate disentangling speaker and content representation technique with domain adversarial training technique. Domain adversarial training makes speaker representations and content representations extracted by speaker encoder and content encoder from clean speech and noisy speech in the same space, respectively. In this way, the learned speaker and content representations are noise-invariant. Therefore, the two noise-invariant representations can be taken as input by the decoder to predict the clean converted spectrum. The experimental results demonstrate that our proposed method can synthesize clean converted speech under noisy test scenarios, where the source speech and target speech can be corrupted by seen or unseen noise types during the training process. Additionally, both speech quality and speaker similarity are improved.

Asunto(s)

RESUMEN

BACKGROUND AND OBJECTIVE: Most dysarthric patients encounter communication problems due to unintelligible speech. Currently, there are many voice-driven systems aimed at improving their speech intelligibility; however, the intelligibility performance of these systems are affected by challenging application conditions (e.g., time variance of patient's speech and background noise). To alleviate these problems, we proposed a dysarthria voice conversion (DVC) system for dysarthric patients and investigated the benefits under challenging application conditions. METHOD: A deep learning-based voice conversion system with phonetic posteriorgram (PPG) features, called the DVC-PPG system, was proposed in this study. An objective-evaluation metric of Google automatic speech recognition (Google ASR) system and a listening test were used to demonstrate the speech intelligibility benefits of DVC-PPG under quiet and noisy test conditions; besides, the well-known voice conversion system using mel-spectrogram, DVC-Mels, was used for comparison to verify the benefits of the proposed DVC-PPG system. RESULTS: The objective-evaluation metric of Google ASR showed the average accuracy of two subjects in the duplicate and outside test conditions while the DVC-PPG system provided higher speech recognitions rate (83.2% and 67.5%) than dysarthric speech (36.5% and 26.9%) and DVC-Mels (52.9% and 33.8%) under quiet conditions. However, the DVC-PPG system provided more stable performance than the DVC-Mels under noisy test conditions. In addition, the results of the listening test showed that the speech-intelligibility performance of DVC-PPG was better than those obtained via the dysarthria speech and DVC-Mels under the duplicate and outside conditions, respectively. CONCLUSIONS: The objective-evaluation metric and listening test results showed that the recognition rate of the proposed DVC-PPG system was significantly higher than those obtained via the original dysarthric speech and DVC-Mels system. Therefore, it can be inferred from our study that the DVC-PPG system can improve the ability of dysarthric patients to communicate with people under challenging application conditions.

Asunto(s)

RESUMEN

We investigate a structured sparse spectral transform method for voice conversion (VC) to perform frequency warping and spectral shaping simultaneously on high-dimensional (D) STRAIGHT spectra. Learning a large transform matrix for high-D data often results in an overfit matrix with low sparsity, which leads to muffled speech in VC. We address this problem by using the frequency-warping characteristic of a source-target speaker pair to define a region of support (ROS) in a transform matrix, and further optimize it by nonnegative matrix factorization (NMF) to obtain structured sparse transform. We also investigate structural measures of spectral and temporal covariance and variance at different scales for assessing VC speech quality. Our experiments on ARCTIC dataset of 12 speaker pairs show that embedding the ROS in spectral transforms offers flexibility in tradeoffs between spectral distortion and structure preservation, and the structural measures provide quantitatively reasonable results on converted speech. Our subjective listening tests show that the proposed VC method achieves a mean opinion score of "very good" relative to natural speech, and in comparison with three other VC methods, it is the most preferred one in naturalness and in voice similarity to target speakers.

RESUMEN

In this paper, we propose a hybrid system based on a modified statistical GMM voice conversion algorithm for improving the recognition of esophageal speech. This hybrid system aims to compensate for the distorted information present in the esophageal acoustic features by using a voice conversion method. The esophageal speech is converted into a "target" laryngeal speech using an iterative statistical estimation of a transformation function. We did not apply a speech synthesizer for reconstructing the converted speech signal, given that the converted Mel cepstral vectors are used directly as input of our speech recognition system. Furthermore the feature vectors are linearly transformed by the HLDA (heteroscedastic linear discriminant analysis) method to reduce their size in a smaller space having good discriminative properties. The experimental results demonstrate that our proposed system provides an improvement of the phone recognition accuracy with an absolute increase of 3.40 % when compared with the phone recognition accuracy obtained with neither HLDA nor voice conversion.

RESUMEN

Text-to-speech options on augmentative and alternative communication (AAC) devices are limited. Often, several individuals in a group setting use the same synthetic voice. This lack of customization may limit technology adoption and social integration. This paper describes our efforts to generate personalized synthesis for users with profoundly limited speech motor control. Existing voice banking and voice conversion techniques rely on recordings of clearly articulated speech from the target talker, which cannot be obtained from this population. Our VocaliD approach extracts prosodic properties from the target talker's source function and applies these features to a surrogate talker's database, generating a synthetic voice with the vocal identity of the target talker and the clarity of the surrogate talker. Promising intelligibility results suggest areas of further development for improved personalization.

Asunto(s)