The automatic recognition of sound events by computers is an important aspect of emerging applications such as automated surveillance, machine hearing and auditory scene understanding. Recent advances in machine learning, as well as in computational models of the human auditory system, have contributed to advances in this increasingly popular research field. Robust sound event classification, the ability to recognise sounds under real-world noisy conditions, is an especially challenging task. Classification methods translated from the speech recognition domain, using features such as mel-frequency cepstral coefficients, have been shown to perform reasonably well for the sound event classification task, although spectrogram-based or auditory image analysis techniques reportedly achieve superior performance in noise. This paper outlines a sound event classification framework that compares auditory image front end features with spectrogram image-based front end features, using support vector machine and deep neural network classifiers. Performance is evaluated on a standard robust classification task in different levels of corrupting noise, and with several system enhancements, and shown to compare very well with current state-of-the-art classification techniques.

/pdf/robust-sound-event-classification-using-deep-neural-networks-3o8zcrj38q.pdf

Robust sound event classification using deep neural networks

Modern text-to-speech synthesis pipelines typically involve multiple processing stages, each of which is designed or learnt independently from the rest. In this work, we take on the challenging task of learning to synthesise speech from normalised text or phonemes in an end-to-end manner, resulting in models which operate directly on character or phoneme input sequences and produce raw speech audio outputs. Our proposed generator is feed-forward and thus efficient for both training and inference, using a differentiable alignment scheme based on token length prediction. It learns to produce high fidelity audio through a combination of adversarial feedback and prediction losses constraining the generated audio to roughly match the ground truth in terms of its total duration and mel-spectrogram. To allow the model to capture temporal variation in the generated audio, we employ soft dynamic time warping in the spectrogram-based prediction loss. The resulting model achieves a mean opinion score exceeding 4 on a 5 point scale, which is comparable to the state-of-the-art models relying on multi-stage training and additional supervision.

End-to-End Adversarial Text-to-Speech.

Despite the ability to produce human-level speech for in-domain text, attention-based end-to-end text-to-speech (TTS) systems suffer from text alignment failures that increase in frequency for out-of-domain text. We show that these failures can be addressed using simple location-relative attention mechanisms that do away with content-based query/key comparisons. We compare two families of attention mechanisms: location-relative GMM-based mechanisms and additive energy-based mechanisms. We suggest simple modifications to GMM-based attention that allow it to align quickly and consistently during training, and introduce a new location-relative attention mechanism to the additive energy-based family, called Dynamic Convolution Attention (DCA). We compare the various mechanisms in terms of alignment speed and consistency during training, naturalness, and ability to generalize to long utterances, and conclude that GMM attention and DCA can generalize to very long utterances, while preserving naturalness for shorter, in-domain utterances.

Location-Relative Attention Mechanisms for Robust Long-Form Speech Synthesis

While speaker adaptation for end-to-end speech synthesis using speaker embeddings can produce good speaker similarity for speakers seen during training, there remains a gap for zero-shot adaptation to unseen speakers. We investigate multi-speaker modeling for end-to-end text-to-speech synthesis and study the effects of different types of state-of-the-art neural speaker embeddings on speaker similarity for unseen speakers. Learnable dictionary encoding-based speaker embeddings with angular softmax loss can improve equal error rates over x-vectors in a speaker verification task; these embeddings also improve speaker similarity and naturalness for unseen speakers when used for zero-shot adaptation to new speakers in end-to-end speech synthesis.

Zero-Shot Multi-Speaker Text-To-Speech with State-Of-The-Art Neural Speaker Embeddings

This paper presents Non-Attentive Tacotron based on the Tacotron 2 text-to-speech model, replacing the attention mechanism with an explicit duration predictor. This improves robustness significantly as measured by unaligned duration ratio and word deletion rate, two metrics introduced in this paper for large-scale robustness evaluation using a pre-trained speech recognition model. With the use of Gaussian upsampling, Non-Attentive Tacotron achieves a 5-scale mean opinion score for naturalness of 4.41, slightly outperforming Tacotron 2. The duration predictor enables both utterance-wide and per-phoneme control of duration at inference time. When accurate target durations are scarce or unavailable in the training data, we propose a method using a fine-grained variational auto-encoder to train the duration predictor in a semi-supervised or unsupervised manner, with results almost as good as supervised training.

/pdf/non-attentive-tacotron-robust-and-controllable-neural-tts-zrhsmww2tn.pdf

Non-Attentive Tacotron: Robust and controllable neural TTS synthesis including unsupervised duration modeling

Neural TTS has demonstrated strong capabilities to generate human-like speech with high quality and naturalness, while its generalization to out-of-domain texts is still a challenging task, with regard to the design of attention-based sequence-to-sequence acoustic modeling. Various errors occur in those inputs with unseen context, including attention collapse, skipping, repeating, etc., which limits the broader applications. In this paper, we propose a novel stepwise monotonic attention method in sequence-to-sequence acoustic modeling to improve the robustness on out-of-domain inputs. The method utilizes the strict monotonic property in TTS with constraints on monotonic hard attention that the alignments between inputs and outputs sequence must be not only monotonic but allowing no skipping on inputs. Soft attention could be used to evade mismatch between training and inference. The experimental results show that the proposed method could achieve significant improvements in robustness on out-of-domain scenarios for phoneme-based models, without any regression on the in-domain naturalness test.

Robust Sequence-to-Sequence Acoustic Modeling with Stepwise Monotonic Attention for Neural TTS

Neural TTS has shown it can generate high quality synthesized speech. In this paper, we investigate the multi-speaker latent space to improve neural TTS for adapting the system to new speakers with only several minutes of speech or enhancing a premium voice by utilizing the data from other speakers for richer contextual coverage and better generalization. A multi-speaker neural TTS model is built with the embedded speaker information in both spectral and speaker latent space. The experimental results show that, with less than 5 minutes of training data from a new speaker, the new model can achieve an MOS score of 4.16 in naturalness and 4.64 in speaker similarity close to human recordings (4.74). For a well-trained premium voice, we can achieve an MOS score of 4.5 for out-of-domain texts, which is comparable to an MOS of 4.58 for professional recordings, and significantly outperforms single speaker result of 4.28.

Modeling Multi-speaker Latent Space to Improve Neural TTS: Quick Enrolling New Speaker and Enhancing Premium Voice

The shifted delta cepstrum (SDC) is a widely used feature extraction for language recognition (LRE). With a high context width due to incorporation of multiple frames, SDC outperforms traditional delta and acceleration feature vectors. However, it also introduces correlation into the concatenated feature vector, which increases redundancy and may degrade the performance of backend classifiers. In this paper, we first propose a time-frequency cepstral (TFC) feature vector, which is obtained by performing a temporal discrete cosine transform (DCT) on the cepstrum matrix and selecting the transformed elements in a zigzag scan order. Beyond this, we increase discriminability through a heteroscedastic linear discriminant analysis (HLDA) on the full cepstrum matrix. By utilizing block diagonal matrix constraints, the large HLDA problem is then reduced to several smaller HLDA problems, creating a block diagonal HLDA (BDHLDA) algorithm which has much lower computational complexity. The BDHLDA method is finally extended to the GMM domain, using the simpler TFC features during re-estimation to provide significantly improved computation speed. Experiments on NIST 2003 and 2007 LRE evaluation corpora show that TFC is more effective than SDC, and that the GMM-based BDHLDA results in lower equal error rate (EER) and minimum average cost (Cavg) than either TFC or SDC approaches.

/pdf/time-frequency-cepstral-features-and-heteroscedastic-linear-1ftet48nh7.pdf

Time–Frequency Cepstral Features and Heteroscedastic Linear Discriminant Analysis for Language Recognition

This paper presents a speech BERT model to extract embedded prosody information in speech segments for improving the prosody of synthesized speech in neural text-to-speech (TTS). As a pre-trained model, it can learn prosody attributes from a large amount of speech data, which can utilize more data than the original training data used by the target TTS. The embedding is extracted from the previous segment of a fixed length in the proposed BERT. The extracted embedding is then used together with the mel-spectrogram to predict the following segment in the TTS decoder. Experimental results obtained by the Transformer TTS show that the proposed BERT can extract fine-grained, segment-level prosody, which is complementary to utterance-level prosody to improve the final prosody of the TTS speech. The objective distortions measured on a single speaker TTS are reduced between the generated speech and original recordings. Subjective listening tests also show that the proposed approach is favorably preferred over the TTS without the BERT prosody embedding module, for both in-domain and out-of-domain applications. For Microsoft professional, single/multiple speakers and the LJ Speaker in the public database, subjective preference is similarly confirmed with the new BERT prosody embedding. TTS demo audio samples are in https://judy44chen.github.io/TTSSpeechBERT/.

Speech Bert Embedding for Improving Prosody in Neural TTS

In this paper, we describe two approaches for language identification (LID) using support vector machines (SVM) and phonetic n-gram. One is to use the language model scores of phone sequences to do SVM training. The other is to use the n-gram probabilities of those phones to train SVM models. For the second approach, we propose a new effective normalization method. In the experiments of 30 s test for 5 languages, our new normalization method shows a relative reduction of 15.8% in terms of equal error rate (EER) compared with the traditional one. And it makes the system using the second approach reaches an EER of 2.4%, a relative reduction of about 35.5% in comparison with the first one. Details of implementation and experimental results are presented in this paper.

Yan Deng

Papers

Robust Sequence-to-Sequence Acoustic Modeling with Stepwise Monotonic Attention for Neural TTS

Modeling Multi-speaker Latent Space to Improve Neural TTS: Quick Enrolling New Speaker and Enhancing Premium Voice

Time–Frequency Cepstral Features and Heteroscedastic Linear Discriminant Analysis for Language Recognition

Speech Bert Embedding for Improving Prosody in Neural TTS

Automatic language identification using support vector machines and phonetic N-gram