In this paper, we describe KT-Speech-Crawler: an approach for automatic dataset construction for speech recognition by crawling YouTube videos. We outline several filtering and post-processing steps, which extract samples that can be used for training end-to-end neural speech recognition systems. In our experiments, we demonstrate that a single-core version of the crawler can obtain around 150 hours of transcribed speech within a day, containing an estimated 3.5% word error rate in the transcriptions. Automatically collected samples contain reading and spontaneous speech recorded in various conditions including background noise and music, distant microphone recordings, and a variety of accents and reverberation. When training a deep neural network on speech recognition, we observed around 40\% word error rate reduction on the Wall Street Journal dataset by integrating 200 hours of the collected samples into the training set. The demo (http://emnlp-demo.lakomkin.me/) and the crawler code (https://github.com/EgorLakomkin/KTSpeechCrawler) are publicly available.
Mel Frequency Cepstral Coefficients (MFCCs) are the most popularly used speech features in most speech and speaker recognition applications. In this paper, we study the effect of resampling a speech signal on these speech features. We first derive a relationship between the MFCC param- eters of the resampled speech and the MFCC parameters of the original speech. We propose six methods of calculating the MFCC parameters of downsampled speech by transforming the Mel filter bank used to com- pute MFCC of the original speech. We then experimentally compute the MFCC parameters of the down sampled speech using the proposed meth- ods and compute the Pearson coefficient between the MFCC parameters of the downsampled speech and that of the original speech to identify the most effective choice of Mel-filter band that enables the computed MFCC of the resampled speech to be as close as possible to the original speech sample MFCC.
Speech-based inputs have been gaining significant momentum with the popularity of smartphones and tablets in our daily lives, since voice is the most easiest and efficient way for human-computer interaction. This paper works towards designing more effective speech-based interfaces to query the structured data in relational databases. We first identify a new task named Speech-to-SQL, which aims to understand the information conveyed by human speech and directly translate it into structured query language (SQL) statements. A naive solution to this problem can work in a cascaded manner, that is, an automatic speech recognition (ASR) component followed by a text-to-SQL component. However, it requires a high-quality ASR system and also suffers from the error compounding problem between the two components, resulting in limited performance. To handle these challenges, we further propose a novel end-to-end neural architecture named SpeechSQLNet to directly translate human speech into SQL queries without an external ASR step. SpeechSQLNet has the advantage of making full use of the rich linguistic information presented in speech. To the best of our knowledge, this is the first attempt to directly synthesize SQL based on arbitrary natural language questions, rather than a natural language-based version of SQL or its variants with a limited SQL grammar. To validate the effectiveness of the proposed problem and model, we further construct a dataset named SpeechQL, by piggybacking the widely-used text-to-SQL datasets. Extensive experimental evaluations on this dataset show that SpeechSQLNet can directly synthesize high-quality SQL queries from human speech, outperforming various competitive counterparts as well as the cascaded methods in terms of exact match accuracies.
This paper presents a transfer learning method in speech emotion recognition based on a Time-Delay Neural Network (TDNN) architecture. A major challenge in the current speech-based emotion detection research is data scarcity. The proposed method resolves this problem by applying transfer learning techniques in order to leverage data from the automatic speech recognition (ASR) task for which ample data is available. Our experiments also show the advantage of speaker-class adaptation modeling techniques by adopting identity-vector (i-vector) based features in addition to standard Mel-Frequency Cepstral Coefficient (MFCC) features.[1] We show the transfer learning models significantly outperform the other methods without pretraining on ASR. The experiments performed on the publicly available IEMOCAP dataset which provides 12 hours of motional speech data. The transfer learning was initialized by using the Ted-Lium v.2 speech dataset providing 207 hours of audio with the corresponding transcripts. We achieve the highest significantly higher accuracy when compared to state-of-the-art, using five-fold cross validation. Using only speech, we obtain an accuracy 71.7% for anger, excitement, sadness, and neutrality emotion content.
In this paper, we explore an improved framework to train a monoaural neural enhancement model for robust speech recognition. The designed training framework extends the existing mixture invariant training criterion to exploit both unpaired clean speech and real noisy data. It is found that the unpaired clean speech is crucial to improve quality of separated speech from real noisy speech. The proposed method also performs remixing of processed and unprocessed signals to alleviate the processing artifacts. Experiments on the single-channel CHiME-3 real test sets show that the proposed method improves significantly in terms of speech recognition performance over the enhancement system trained either on the mismatched simulated data in a supervised fashion or on the matched real data in an unsupervised fashion. Between 16% and 39% relative WER reduction has been achieved by the proposed system compared to the unprocessed signal using end-to-end and hybrid acoustic models without retraining on distorted data.
Thanks to improvements in machine learning techniques including deep learning, a free large-scale speech corpus that can be shared between academic institutions and commercial companies has an important role. However, such a corpus for Japanese speech synthesis does not exist. In this paper, we designed a novel Japanese speech corpus, named the "JSUT corpus," that is aimed at achieving end-to-end speech synthesis. The corpus consists of 10 hours of reading-style speech data and its transcription and covers all of the main pronunciations of daily-use Japanese characters. In this paper, we describe how we designed and analyzed the corpus. The corpus is freely available online.
Training a text-to-speech (TTS) model requires a large scale text labeled speech corpus, which is troublesome to collect. In this paper, we propose a transfer learning framework for TTS that utilizes a large amount of unlabeled speech dataset for pre-training. By leveraging wav2vec2.0 representation, unlabeled speech can highly improve performance, especially in the lack of labeled speech. We also extend the proposed method to zero-shot multi-speaker TTS (ZS-TTS). The experimental results verify the effectiveness of the proposed method in terms of naturalness, intelligibility, and speaker generalization. We highlight that the single speaker TTS model fine-tuned on the only 10 minutes of labeled dataset outperforms the other baselines, and the ZS-TTS model fine-tuned on the only 30 minutes of single speaker dataset can generate the voice of the arbitrary speaker, by pre-training on unlabeled multi-speaker speech corpus.
We present a state-of-the-art speech recognition system developed using end-to-end deep learning. Our architecture is significantly simpler than traditional speech systems, which rely on laboriously engineered processing pipelines; these traditional systems also tend to perform poorly when used in noisy environments. In contrast, our system does not need hand-designed components to model background noise, reverberation, or speaker variation, but instead directly learns a function that is robust to such effects. We do not need a phoneme dictionary, nor even the concept of a "phoneme." Key to our approach is a well-optimized RNN training system that uses multiple GPUs, as well as a set of novel data synthesis techniques that allow us to efficiently obtain a large amount of varied data for training. Our system, called Deep Speech, outperforms previously published results on the widely studied Switchboard Hub5'00, achieving 16.0% error on the full test set. Deep Speech also handles challenging noisy environments better than widely used, state-of-the-art commercial speech systems.
Vocoders received renewed attention as main components in statistical parametric text-to-speech (TTS) synthesis and speech transformation systems. Even though there are vocoding techniques give almost accepted synthesized speech, their high computational complexity and irregular structures are still considered challenging concerns, which yield a variety of voice quality degradation. Therefore, this paper presents new techniques in a continuous vocoder, that is all features are continuous and presents a flexible speech synthesis system. First, a new continuous noise masking based on the phase distortion is proposed to eliminate the perceptual impact of the residual noise and letting an accurate reconstruction of noise characteristics. Second, we addressed the need of neural sequence to sequence modeling approach for the task of TTS based on recurrent networks. Bidirectional long short-term memory (LSTM) and gated recurrent unit (GRU) are studied and applied to model continuous parameters for more natural-sounding like a human. The evaluation results proved that the proposed model achieves the state-of-the-art performance of the speech synthesis compared with the other traditional methods.