Electrical Engineering And Systems Science

A Spoken dialogue system for an unseen language is referred to as Zero resource speech. It is especially beneficial for developing applications for languages that have low digital resources. Zero resource speech synthesis is the task of building text-to-speech (TTS) models in the absence of transcriptions. In this work, speech is modelled as a sequence of transient and steady-state acoustic units, and a unique set of acoustic units is discovered by iterative training. Using the acoustic unit sequence, TTS models are trained. The main goal of this work is to improve the synthesis quality of zero resource TTS system. Four different systems are proposed. All the systems consist of three stages: unit discovery, followed by unit sequence to spectrogram mapping, and finally spectrogram to speech inversion. Modifications are proposed to the spectrogram mapping stage. These modifications include training the mapping on voice data, using x-vectors to improve the mapping, two-stage learning, and gender-specific modelling. Evaluation of the proposed systems in the Zerospeech 2020 challenge shows that quite good quality synthesis can be achieved.

In this paper, we propose a method of utilizing aligned lyrics as additional information to improve the performance of singing voice separation. We have combined the highway network-based lyrics encoder into Open-unmix separation network and show that the model trained with the aligned lyrics indeed results in a better performance than the model that was not informed. The question now remains whether the increase of performance is actually due to the phonetic contents that lie in the informed aligned lyrics or not. To this end, we investigated the source of performance increase in multifaceted ways by observing the change of performance when incorrect lyrics were given to the model. Experiment results show that the model can use not only just vocal activity information but also the phonetic contents from the aligned lyrics.

This paper investigates different trade-offs between the number of model parameters and enhanced speech qualities by employing several deep tensor-to-vector regression models for speech enhancement. We find that a hybrid architecture, namely CNN-TT, is capable of maintaining a good quality performance with a reduced model parameter size. CNN-TT is composed of several convolutional layers at the bottom for feature extraction to improve speech quality and a tensor-train (TT) output layer on the top to reduce model parameters. We first derive a new upper bound on the generalization power of the convolutional neural network (CNN) based vector-to-vector regression models. Then, we provide experimental evidence on the Edinburgh noisy speech corpus to demonstrate that, in single-channel speech enhancement, CNN outperforms DNN at the expense of a small increment of model sizes. Besides, CNN-TT slightly outperforms the CNN counterpart by utilizing only 32\% of the CNN model parameters. Besides, further performance improvement can be attained if the number of CNN-TT parameters is increased to 44\% of the CNN model size. Finally, our experiments of multi-channel speech enhancement on a simulated noisy WSJ0 corpus demonstrate that our proposed hybrid CNN-TT architecture achieves better results than both DNN and CNN models in terms of better-enhanced speech qualities and smaller parameter sizes.

In this paper, we propose a novel way of addressing text-dependent automatic speaker verification (TD-ASV) by using a shared-encoder with task-specific decoders. An autoregressive predictive coding (APC) encoder is pre-trained in an unsupervised manner using both out-of-domain (LibriSpeech, VoxCeleb) and in-domain (DeepMine) unlabeled datasets to learn generic, high-level feature representation that encapsulates speaker and phonetic content. Two task-specific decoders were trained using labeled datasets to classify speakers (SID) and phrases (PID). Speaker embeddings extracted from the SID decoder were scored using a PLDA. SID and PID systems were fused at the score level. There is a 51.9% relative improvement in minDCF for our system compared to the fully supervised x-vector baseline on the cross-lingual DeepMine dataset. However, the i-vector/HMM method outperformed the proposed APC encoder-decoder system. A fusion of the x-vector/PLDA baseline and the SID/PLDA scores prior to PID fusion further improved performance by 15% indicating complementarity of the proposed approach to the x-vector system. We show that the proposed approach can leverage from large, unlabeled, data-rich domains, and learn speech patterns independent of downstream tasks. Such a system can provide competitive performance in domain-mismatched scenarios where test data is from data-scarce domains.

Deep attractor networks (DANs) perform speech separation with discriminative embeddings and speaker attractors. Compared with methods based on the permutation invariant training (PIT), DANs define a deep embedding space and deliver a more elaborate representation on each time-frequency (T-F) bin. However, it has been observed that the DANs achieve limited improvement on the signal quality if directly deployed in a reverberant environment. Following the success of time-domain separation networks on the clean mixture speech, we propose a time-domain DAN (TD-DAN) with two-streams of convolutional networks, which efficiently perform both dereverberation and separation tasks under the condition of a variable number of speakers. The speaker encoding stream (SES) of the TD-DAN is trained to model the speaker information in the embedding space. The speech decoding stream (SDS) accepts speaker attractors from the SES and learns to estimate early reflections from the spectro-temporal representations. Meanwhile, additional clustering losses are used to bridge the gap between the oracle and the estimated attractors. Experiments were conducted on the Spatialized Multi-Speaker Wall Street Journal (SMS-WSJ) dataset. The early reflection was compared with the anechoic and reverberant signals and then was chosen as the learning targets. The experimental results demonstrated that the TD-DAN achieved scale-invariant source-to-distortion ratio (SI-SDR) gains of 9.79/7.47 dB on the reverberant 2/3-speaker evaluation set, exceeding the baseline DAN and convolutional time-domain audio separation network (Conv-TasNet) by 1.92/0.68 dB and 0.91/0.47 dB, respectively.

Automatic speaker verification (ASV) vendors and corpus providers would both benefit from tools to reliably extrapolate performance metrics for large speaker populations without collecting new speakers. We address false alarm rate extrapolation under a worst-case model whereby an adversary identifies the closest impostor for a given target speaker from a large population. Our models are generative and allow sampling new speakers. The models are formulated in the ASV detection score space to facilitate analysis of arbitrary ASV systems.

In this paper, we present a deep neural network-based online multi-speaker localisation algorithm. Following the W-disjoint orthogonality principle in the spectral domain, each time-frequency (TF) bin is dominated by a single speaker, and hence by a single direction of arrival (DOA). A fully convolutional network is trained with instantaneous spatial features to estimate the DOA for each TF bin. The high resolution classification enables the network to accurately and simultaneously localize and track multiple speakers, both static and dynamic. Elaborated experimental study using both simulated and real-life recordings in static and dynamic scenarios, confirms that the proposed algorithm outperforms both classic and recent deep-learning-based algorithms.

The Fearless Steps Initiative by UTDallas-CRSS led to the digitization, recovery, and diarization of 19,000 hours of original analog audio data, as well as the development of algorithms to extract meaningful information from this multi-channel naturalistic data resource. The 2020 FEARLESS STEPS (FS-2) Challenge is the second annual challenge held for the Speech and Language Technology community to motivate supervised learning algorithm development for multi-party and multi-stream naturalistic audio. In this paper, we present an overview of the challenge sub-tasks, data, performance metrics, and lessons learned from Phase-2 of the Fearless Steps Challenge (FS-2). We present advancements made in FS-2 through extensive community outreach and feedback. We describe innovations in the challenge corpus development, and present revised baseline results. We finally discuss the challenge outcome and general trends in system development across both phases (Phase FS-1 Unsupervised, and Phase FS-2 Supervised) of the challenge, and its continuation into multi-channel challenge tasks for the upcoming Fearless Steps Challenge Phase-3.

The machine recognition of speech spoken at a distance from the microphones, known as far-field automatic speech recognition (ASR), has received a significant increase of attention in science and industry, which caused or was caused by an equally significant improvement in recognition accuracy. Meanwhile it has entered the consumer market with digital home assistants with a spoken language interface being its most prominent application. Speech recorded at a distance is affected by various acoustic distortions and, consequently, quite different processing pipelines have emerged compared to ASR for close-talk speech. A signal enhancement front-end for dereverberation, source separation and acoustic beamforming is employed to clean up the speech, and the back-end ASR engine is robustified by multi-condition training and adaptation. We will also describe the so-called end-to-end approach to ASR, which is a new promising architecture that has recently been extended to the far-field scenario. This tutorial article gives an account of the algorithms used to enable accurate speech recognition from a distance, and it will be seen that, although deep learning has a significant share in the technological breakthroughs, a clever combination with traditional signal processing can lead to surprisingly effective solutions.

The performance of text-to-speech (TTS) systems heavily depends on spectrogram to waveform generation, also known as the speech reconstruction phase. The time required for the same is known as synthesis delay. In this paper, an approach to reduce speech synthesis delay has been proposed. It aims to enhance the TTS systems for real-time applications such as digital assistants, mobile phones, embedded devices, etc. The proposed approach applies Fast Griffin Lim Algorithm (FGLA) instead Griffin Lim algorithm (GLA) as vocoder in the speech synthesis phase. GLA and FGLA are both iterative, but the convergence rate of FGLA is faster than GLA. The proposed approach is tested on LJSpeech, Blizzard and Tatoeba datasets and the results for FGLA are compared against GLA and neural Generative Adversarial Network (GAN) based vocoder. The performance is evaluated based on synthesis delay and speech quality. A 36.58% reduction in speech synthesis delay has been observed. The quality of the output speech has improved, which is advocated by higher Mean opinion scores (MOS) and faster convergence with FGLA as opposed to GLA.