Voice communication between air traffic controllers (ATCos) and pilots is critical for ensuring safe and efficient air traffic control (ATC). This task requires high levels of awareness from ATCos and can be tedious and error-prone. Recent attempts have been made to integrate artificial intelligence (AI) into ATC in order to reduce the workload of ATCos. However, the development of data-driven AI systems for ATC demands large-scale annotated datasets, which are currently lacking in the field. This paper explores the lessons learned from the ATCO2 project, a project that aimed to develop a unique platform to collect and preprocess large amounts of ATC data from airspace in real time. Audio and surveillance data were collected from publicly accessible radio frequency channels with VHF receivers owned by a community of volunteers and later uploaded to Opensky Network servers, which can be considered an "unlimited source" of data. In addition, this paper reviews previous work from ATCO2 partners, including (i) robust automatic speech recognition, (ii) natural language processing, (iii) English language identification of ATC communications, and (iv) the integration of surveillance data such as ADS-B. We believe that the pipeline developed during the ATCO2 project, along with the open-sourcing of its data, will encourage research in the ATC field. A sample of the ATCO2 corpus is available on the following website: https://www.atco2.org/data, while the full corpus can be purchased through ELDA at http://catalog.elra.info/en-us/repository/browse/ELRA-S0484. We demonstrated that ATCO2 is an appropriate dataset to develop ASR engines when little or near to no ATC in-domain data is available. For instance, with the CNN-TDNNf kaldi model, we reached the performance of as low as 17.9% and 24.9% WER on public ATC datasets which is 6.6/7.6% better than "out-of-domain" but supervised CNN-TDNNf model.
Speaker anonymization aims to conceal a speaker's identity while preserving content information in speech. Current mainstream neural-network speaker anonymization systems disentangle speech into prosody-related, content, and speaker representations. The speaker representation is then anonymized by a selection-based speaker anonymizer that uses a mean vector over a set of randomly selected speaker vectors from an external pool of English speakers. However, the resulting anonymized vectors are subject to severe privacy leakage against powerful attackers, reduction in speaker diversity, and language mismatch problems for unseen language speaker anonymization. To generate diverse, language-neutral speaker vectors, this paper proposes an anonymizer based on an orthogonal Householder neural network (OHNN). Specifically, the OHNN acts like a rotation to transform the original speaker vectors into anonymized speaker vectors, which are constrained to follow the distribution over the original speaker vector space. A basic classification loss is introduced to ensure that anonymized speaker vectors from different speakers have unique speaker identities. To further protect speaker identities, an improved classification loss and similarity loss are used to push original-anonymized sample pairs away from each other. Experiments on VoicePrivacy Challenge datasets in English and the AISHELL-3 dataset in Mandarin demonstrate the proposed anonymizer's effectiveness.
Most deep learning pipelines are built on real-valued operations to deal with real-valued inputs such as images, speech or music signals. However, a lot of applications naturally make use of complex-valued signals or images, such as MRI or remote sensing. Additionally the Fourier transform of signals is complex-valued and has numerous applications. We aim to make deep learning directly applicable to these complex-valued signals without using projections into $\mathbb{R}^2$. Thus we add to the recent developments of complex-valued neural networks by presenting building blocks to transfer the transformer architecture to the complex domain. We present multiple versions of a complex-valued Scaled Dot-Product Attention mechanism as well as a complex-valued layer normalization. We test on a classification and a sequence generation task on the MusicNet dataset and show improved robustness to overfitting while maintaining on-par performance when compared to the real-valued transformer architecture.
Recently, it is quite common to integrate Chinese sequence labeling results to enhance syntactic and semantic parsing. However, little attention has been paid to the utility of hierarchy and structure information encoded in syntactic and semantic features for Chinese sequence labeling tasks. In this paper, we propose a novel framework to encode syntactic structure features and semantic information for Chinese sequence labeling tasks with graph convolutional networks (GCN). Experiments on five benchmark datasets, including Chinese word segmentation and part-of-speech tagging, demonstrate that our model can effectively improve the performance of Chinese labeling tasks.
Conformer-based end-to-end models have become ubiquitous these days and are commonly used in both streaming and non-streaming automatic speech recognition (ASR). Techniques like dual-mode and dynamic chunk training helped unify streaming and non-streaming systems. However, there remains a performance gap between streaming with a full and limited past context. To address this issue, we propose the integration of a novel dynamic contextual carry-over mechanism in a state-of-the-art (SOTA) unified ASR system. Our proposed dynamic context Conformer (DCTX-Conformer) utilizes a non-overlapping contextual carry-over mechanism that takes into account both the left context of a chunk and one or more preceding context embeddings. We outperform the SOTA by a relative 25.0% word error rate, with a negligible latency impact due to the additional context embeddings.
Cross-speaker style transfer in speech synthesis aims at transferring a style from source speaker to synthesized speech of a target speaker's timbre. In most previous methods, the synthesized fine-grained prosody features often represent the source speaker's average style, similar to the one-to-many problem(i.e., multiple prosody variations correspond to the same text). In response to this problem, a strength-controlled semi-supervised style extractor is proposed to disentangle the style from content and timbre, improving the representation and interpretability of the global style embedding, which can alleviate the one-to-many mapping and data imbalance problems in prosody prediction. A hierarchical prosody predictor is proposed to improve prosody modeling. We find that better style transfer can be achieved by using the source speaker's prosody features that are easily predicted. Additionally, a speaker-transfer-wise cycle consistency loss is proposed to assist the model in learning unseen style-timbre combinations during the training phase. Experimental results show that the method outperforms the baseline. We provide a website with audio samples.
Modern text-to-speech (TTS) systems use deep learning to synthesize speech increasingly approaching human quality, but they require a database of high quality audio-text sentence pairs for training. Malayalam, the official language of the Indian state of Kerala and spoken by 35+ million people, is a low resource language in terms of available corpora for TTS systems. In this paper, we present IMaSC, a Malayalam text and speech corpora containing approximately 50 hours of recorded speech. With 8 speakers and a total of 34,473 text-audio pairs, IMaSC is larger than every other publicly available alternative. We evaluated the database by using it to train TTS models for each speaker based on a modern deep learning architecture. Via subjective evaluation, we show that our models perform significantly better in terms of naturalness compared to previous studies and publicly available models, with an average mean opinion score of 4.50, indicating that the synthesized speech is close to human quality.
This paper presents FastFit, a novel neural vocoder architecture that replaces the U-Net encoder with multiple short-time Fourier transforms (STFTs) to achieve faster generation rates without sacrificing sample quality. We replaced each encoder block with an STFT, with parameters equal to the temporal resolution of each decoder block, leading to the skip connection. FastFit reduces the number of parameters and the generation time of the model by almost half while maintaining high fidelity. Through objective and subjective evaluations, we demonstrated that the proposed model achieves nearly twice the generation speed of baseline iteration-based vocoders while maintaining high sound quality. We further showed that FastFit produces sound qualities similar to those of other baselines in text-to-speech evaluation scenarios, including multi-speaker and zero-shot text-to-speech.
In active speaker detection (ASD), we would like to detect whether an on-screen person is speaking based on audio-visual cues. Previous studies have primarily focused on modeling audio-visual synchronization cue, which depends on the video quality of the lip region of a speaker. In real-world applications, it is possible that we can also have the reference speech of the on-screen speaker. To benefit from both facial cue and reference speech, we propose the Target Speaker TalkNet (TS-TalkNet), which leverages a pre-enrolled speaker embedding to complement the audio-visual synchronization cue in detecting whether the target speaker is speaking. Our framework outperforms the popular model, TalkNet on two datasets, achieving absolute improvements of 1.6\% in mAP on the AVA-ActiveSpeaker validation set, and 0.8\%, 0.4\%, and 0.8\% in terms of AP, AUC and EER on the ASW test set, respectively. Code is available at \href{https://github.com/Jiang-Yidi/TS-TalkNet/}{\color{red}{https://github.com/Jiang-Yidi/TS-TalkNet/}}.
The syllable is a perceptually salient unit in speech. Since both the syllable and its acoustic correlate, i.e., the speech envelope, have a preferred range of rhythmicity between 4 and 8 Hz, it is hypothesized that theta-band neural oscillations play a major role in extracting syllables based on the envelope. A literature survey, however, reveals inconsistent evidence about the relationship between speech envelope and syllables, and the current study revisits this question by analyzing large speech corpora. It is shown that the center frequency of speech envelope, characterized by the modulation spectrum, reliably correlates with the rate of syllables only when the analysis is pooled over minutes of speech recordings. In contrast, in the time domain, a component of the speech envelope is reliably phase-locked to syllable onsets. Based on a speaker-independent model, the timing of syllable onsets explains about 24% variance of the speech envelope. These results indicate that local features in the speech envelope, instead of the modulation spectrum, are a more reliable acoustic correlate of syllables.