The recent ubiquitous adoption of remote conferencing has been accompanied by omnipresent frustration with distorted or otherwise unclear voice communication. Audio enhancement can compensate for low-quality input signals from, for example, small true wireless earbuds, by applying noise suppression techniques. Such processing relies on voice activity detection (VAD) with low latency and the added capability of discriminating the wearer's voice from others - a task of significant computational complexity. The tight energy budget of devices as small as modern earphones, however, requires any system attempting to tackle this problem to do so with minimal power and processing overhead, while not relying on speaker-specific voice samples and training due to usability concerns. This paper presents the design and implementation of a custom research platform for low-power wireless earbuds based on novel, commercial, MEMS bone-conduction microphones. Such microphones can record the wearer's speech with much greater isolation, enabling personalized voice activity detection and further audio enhancement applications. Furthermore, the paper accurately evaluates a proposed low-power personalized speech detection algorithm based on bone conduction data and a recurrent neural network running on the implemented research platform. This algorithm is compared to an approach based on traditional microphone input. The performance of the bone conduction system, achieving detection of speech within 12.8ms at an accuracy of 95\% is evaluated. Different SoC choices are contrasted, with the final implementation based on the cutting-edge Ambiq Apollo 4 Blue SoC achieving 2.64mW average power consumption at 14uJ per inference, reaching 43h of battery life on a miniature 32mAh li-ion cell and without duty cycling.
This paper describes the DKU-MSXF submission to track 4 of the VoxCeleb Speaker Recognition Challenge 2023 (VoxSRC-23). Our system pipeline contains voice activity detection, clustering-based diarization, overlapped speech detection, and target-speaker voice activity detection, where each procedure has a fused output from 3 sub-models. Finally, we fuse different clustering-based and TSVAD-based diarization systems using DOVER-Lap and achieve the 4.30% diarization error rate (DER), which ranks first place on track 4 of the challenge leaderboard.
Although automatic emotion recognition (AER) has recently drawn significant research interest, most current AER studies use manually segmented utterances, which are usually unavailable for dialogue systems. This paper proposes integrating AER with automatic speech recognition (ASR) and speaker diarisation (SD) in a jointly-trained system. Distinct output layers are built for four sub-tasks including AER, ASR, voice activity detection and speaker classification based on a shared encoder. Taking the audio of a conversation as input, the integrated system finds all speech segments and transcribes the corresponding emotion classes, word sequences, and speaker identities. Two metrics are proposed to evaluate AER performance with automatic segmentation based on time-weighted emotion and speaker classification errors. Results on the IEMOCAP dataset show that the proposed system consistently outperforms two baselines with separately trained single-task systems on AER, ASR and SD.
Snoring is a common disorder that affects people's social and marital lives. The annoyance caused by snoring can be partially solved with active noise control systems. In this context, the present work aims at introducing an enhanced system based on the use of a convolutional recurrent neural network for snoring activity detection and a delayless subband approach for active snoring cancellation. Thanks to several experiments conducted using real snoring signals, this work shows that the active snoring cancellation system achieves better performance when the snoring activity detection stage is turned on, demonstrating the beneficial effect of a preliminary snoring detection stage in the perspective of snoring cancellation.
We introduce "ivrit.ai", a comprehensive Hebrew speech dataset, addressing the distinct lack of extensive, high-quality resources for advancing Automated Speech Recognition (ASR) technology in Hebrew. With over 3,300 speech hours and a over a thousand diverse speakers, ivrit.ai offers a substantial compilation of Hebrew speech across various contexts. It is delivered in three forms to cater to varying research needs: raw unprocessed audio; data post-Voice Activity Detection, and partially transcribed data. The dataset stands out for its legal accessibility, permitting use at no cost, thereby serving as a crucial resource for researchers, developers, and commercial entities. ivrit.ai opens up numerous applications, offering vast potential to enhance AI capabilities in Hebrew. Future efforts aim to expand ivrit.ai further, thereby advancing Hebrew's standing in AI research and technology.
The analysis of conversations recorded in everyday life requires privacy protection. In this contribution, we explore a privacy-preserving feature extraction method based on input feature dimension reduction, spectral smoothing and the low-cost speaker anonymization technique based on McAdams coefficient. We assess the utility of the feature extraction methods with a voice activity detection and a speaker diarization system, while privacy protection is determined with a speech recognition and a speaker verification model. We show that the combination of McAdams coefficient and spectral smoothing maintains the utility while improving privacy.
Speaker diarization is the task of answering Who spoke and when? in an audio stream. Pipeline systems rely on speech segmentation to extract speakers' segments and achieve robust speaker diarization. This paper proposes a common framework to solve three segmentation tasks in the distant speech scenario: Voice Activity Detection (VAD), Overlapped Speech Detection (OSD), and Speaker Change Detection (SCD). In the literature, a few studies investigate the multi-microphone distant speech scenario. In this work, we propose a new set of spatial features based on direction-of-arrival estimations in the circular harmonic domain (CH-DOA). These spatial features are extracted from multi-microphone audio data and combined with standard acoustic features. Experiments on the AMI meeting corpus show that CH-DOA can improve the segmentation while being robust in the case of deactivated microphones.
Parallel neurosymbolic architectures have been applied effectively in NLP by distilling knowledge from a logic theory into a deep model.However, prior art faces several limitations including supporting restricted forms of logic theories and relying on the assumption of independence between the logic and the deep network. We present Concordia, a framework overcoming the limitations of prior art. Concordia is agnostic both to the deep network and the logic theory offering support for a wide range of probabilistic theories. Our framework can support supervised training of both components and unsupervised training of the neural component. Concordia has been successfully applied to tasks beyond NLP and data classification, improving the accuracy of state-of-the-art on collective activity detection, entity linking and recommendation tasks.
Voice activity detection (VAD) improves the performance of speaker verification (SV) by preserving speech segments and attenuating the effects of non-speech. However, this scheme is not ideal: (1) it fails in noisy environments or multi-speaker conversations; (2) it is trained based on inaccurate non-SV sensitive labels. To address this, we propose a speaker verification-based voice activity detection (SVVAD) framework that can adapt the speech features according to which are most informative for SV. To achieve this, we introduce a label-free training method with triplet-like losses that completely avoids the performance degradation of SV due to incorrect labeling. Extensive experiments show that SVVAD significantly outperforms the baseline in terms of equal error rate (EER) under conditions where other speakers are mixed at different ratios. Moreover, the decision boundaries reveal the importance of the different parts of speech, which are largely consistent with human judgments.
Automatic Speech Recognition (ASR) plays a crucial role in voice-based applications. For applications requiring real-time feedback like Voice Search, streaming capability becomes vital. While LSTM/RNN and CTC based ASR systems are commonly employed for low-latency streaming applications, they often exhibit lower accuracy compared to state-of-the-art models due to a lack of future audio frames. In this work, we focus on developing accurate LSTM, attention, and CTC based streaming ASR models for large-scale Hinglish (a blend of Hindi and English) Voice Search. We investigate various modifications in vanilla LSTM training which enhance the system's accuracy while preserving its streaming capabilities. We also address the critical requirement of end-of-speech (EOS) detection in streaming applications. We present a simple training and inference strategy for end-to-end CTC models that enables joint ASR and EOS detection. The evaluation of our model on Flipkart's Voice Search, which handles substantial traffic of approximately 6 million queries per day, demonstrates significant performance gains over the vanilla LSTM-CTC model. Our model achieves a word error rate (WER) of 3.69% without EOS and 4.78% with EOS while also reducing the search latency by approximately ~1300 ms (equivalent to 46.64% reduction) when compared to an independent voice activity detection (VAD) model.
Voice Activity Detection (VAD) is becoming an essential front-end component in various speech processing systems. As those systems are commonly deployed in environments with diverse noise types and low signal-to-noise ratios (SNRs), an effective VAD method should perform robust detection of speech region out of noisy background signals. In this paper, we propose adversarial domain adaptive VAD (ADA-VAD), which is a deep neural network (DNN) based VAD method highly robust to audio samples with various noise types and low SNRs. The proposed method trains DNN models for a VAD task in a supervised manner. Simultaneously, to mitigate the performance degradation due to back-ground noises, the adversarial domain adaptation method is adopted to match the domain discrepancy between noisy and clean audio stream in an unsupervised manner. The results show that ADA-VAD achieves an average of 3.6%p and 7%p higher AUC than models trained with manually extracted features on the AVA-speech dataset and a speech database synthesized with an unseen noise database, respectively.
For speech interaction, voice activity detection (VAD) is often used as a front-end. However, traditional VAD algorithms usually need to wait for a continuous tail silence to reach a preset maximum duration before segmentation, resulting in a large latency that affects user experience. In this paper, we propose a novel semantic VAD for low-latency segmentation. Different from existing methods, a frame-level punctuation prediction task is added to the semantic VAD, and the artificial endpoint is included in the classification category in addition to the often-used speech presence and absence. To enhance the semantic information of the model, we also incorporate an automatic speech recognition (ASR) related semantic loss. Evaluations on an internal dataset show that the proposed method can reduce the average latency by 53.3% without significant deterioration of character error rate in the back-end ASR compared to the traditional VAD approach.
Most existing studies on joint activity detection and channel estimation for grant-free massive random access (RA) systems assume perfect synchronization among all active users, which is hard to achieve in practice. Therefore, this paper considers asynchronous grant-free massive RA systems and develops novel algorithms for joint user activity detection, synchronization delay detection, and channel estimation. In particular, the framework of orthogonal approximate message passing (OAMP) is first utilized to deal with the non-independent and identically distributed (i.i.d.) pilot matrix in asynchronous grant-free massive RA systems, and an OAMP-based algorithm capable of leveraging the common sparsity among the received pilot signals from multiple base station antennas is developed. To reduce the computational complexity, a memory AMP (MAMP)based algorithm is further proposed that eliminates the matrix inversions in the OAMP-based algorithm. Simulation results demonstrate the effectiveness of the two proposed algorithms over the baseline methods. Besides, the MAMP-based algorithm reduces 37% of the computations while maintaining comparable detection/estimation accuracy, compared with the OAMP-based algorithm.
This paper introduces FunASR, an open-source speech recognition toolkit designed to bridge the gap between academic research and industrial applications. FunASR offers models trained on large-scale industrial corpora and the ability to deploy them in applications. The toolkit's flagship model, Paraformer, is a non-autoregressive end-to-end speech recognition model that has been trained on a manually annotated Mandarin speech recognition dataset that contains 60,000 hours of speech. To improve the performance of Paraformer, we have added timestamp prediction and hotword customization capabilities to the standard Paraformer backbone. In addition, to facilitate model deployment, we have open-sourced a voice activity detection model based on the Feedforward Sequential Memory Network (FSMN-VAD) and a text post-processing punctuation model based on the controllable time-delay Transformer (CT-Transformer), both of which were trained on industrial corpora. These functional modules provide a solid foundation for building high-precision long audio speech recognition services. Compared to other models trained on open datasets, Paraformer demonstrates superior performance.
Grant-free non-orthogonal multiple access has been regarded as a viable approach to accommodate access for a massive number of machine-type devices with small data packets. The sporadic activation of the devices creates a multiuser setup where it is suitable to use compressed sensing in order to detect the active devices and decode their data. We consider asynchronous access of machine-type devices that send data packets of different frame sizes, leading to data length diversity. We address the composite problem of activity detection, channel estimation, and data recovery by posing it as a structured sparse recovery, having three-level sparsity caused by sporadic activity, symbol delay, and data length diversity. We approach the problem through approximate message passing with a backward propagation algorithm (AMP-BP), tailored to exploit the sparsity, and in particular the data length diversity. Moreover, we unfold the proposed AMP-BP into a model-driven network, termed learned AMP-BP (LAMP-BP), which enhances detection performance. The results show that the proposed LAMP-BP outperforms existing methods in activity detection and data recovery accuracy.
Compressed sensing multi-user detection (CS-MUD) algorithms play a key role in optimizing grant-free (GF) non-orthogonal multiple access (NOMA) for massive machine-type communications (mMTC). However, current CS-MUD algorithms cannot be efficiently parallelized, leading to computationally expensive implementations of joint activity detection and channel estimation (JADCE) as the number of deployed machine-type devices (MTDs) increases. To address this, the present work proposes novel JADCE algorithms that can be applied in parallel for different clusters of MTDs by exploiting the structure of the pilot sequences. These are the approximation error method (AEM)-alternating direction method of multipliers (ADMM), and AEM-sparse Bayesian learning (SBL). Results presented in terms of the normalized mean square error and the probability of miss detection show comparable performance to the conventional algorithms. However, both AEM-ADMM and AEM-SBL algorithms have significantly reduced computational complexity and run times, thus, facilitating network scalability.
This paper investigates the problem of activity detection and channel estimation in cooperative multi-cell massive access systems with temporally correlated activity, where all access points (APs) are connected to a central unit via fronthaul links. We propose to perform user-centric AP cooperation for computation burden alleviation and introduce a generalized sliding-window detection strategy for fully exploiting the temporal correlation in activity. By establishing the probabilistic model associated with the factor graph representation, we propose a scalable Dynamic Compressed Sensing-based Multiple Measurement Vector Generalized Approximate Message Passing (DCS-MMV-GAMP) algorithm from the perspective of Bayesian inference. Therein, the activity likelihood is refined by performing standard message passing among the activities in the spatial-temporal domain and GAMP is employed for efficient channel estimation. Furthermore, we develop two schemes of quantize-and-forward (QF) and detect-and-forward (DF) based on DCS-MMV-GAMP for the finite-fronthaul-capacity scenario, which are extensively evaluated under various system limits. Numerical results verify the significant superiority of the proposed approach over the benchmarks. Moreover, it is revealed that QF can usually realize superior performance when the antenna number is small, whereas DF shifts to be preferable with limited fronthaul capacity if the large-scale antenna arrays are equipped.
Personal voice activity detection has received increased attention due to the growing popularity of personal mobile devices and smart speakers. PVAD is often an integral element to speech enhancement and recognition for these applications in which lightweight signal processing is only enabled for the target user. However, in real-world scenarios, the detection performance may degrade because of competing speakers, background noise, and reverberation. To address this problem, we proposed to use equivalent rectangular bandwidth ERB-scaled spatial coherence as the input feature to train an array configuration-agnostic PVAD network. Whereas the network model requires only 112k parameters, it exhibits excellent detection performance and robustness in adverse acoustic conditions. Notably, the proposed ARCA-PVAD system is scalable to array configurations. Experimental results have demonstrated the superior performance achieved by the proposed ARCA-PVAD system over a baseline in terms of the area under receiver operating characteristic curve and equal error rate.
There is an increasing demand of massive machine-type communication (mMTC) to provide scalable access for a large number of devices, which has prompted extensive investigation on grant-free massive random access (RA) in 5G and beyond wireless networks. Although many efficient signal processing algorithms have been developed, the limited radio resource for pilot transmission in grant-free massive RA systems makes accurate user activity detection and channel estimation challenging, which thereby compromises the communication reliability. In this paper, we adopt retransmission as a means to improve the quality of service (QoS) for grant-free massive RA. Specifically, by jointly leveraging the user activity correlation between adjacent transmission blocks and the historical channel estimation results, we first develop an activity-correlation-aware receiver for grant-free massive RA systems with retransmission based on the correlated approximate message passing (AMP) algorithm. Then, we analyze the performance of the proposed receiver, including the user activity detection, channel estimation, and data error, by resorting to the state evolution of the correlated AMP algorithm and the random matrix theory (RMT). Our analysis admits a tight closed-form approximation for frame error rate (FER) evaluation. Simulation results corroborate our theoretical analysis and demonstrate the effectiveness of the proposed receiver for grant-free massive RA with retransmission, compared with a conventional design that disregards the critical user activity correlation.
Noise reduction is a crucial aspect of hearing aids, which researchers have been striving to address over the years. However, most existing noise reduction algorithms have primarily been evaluated using English. Considering the linguistic differences between English and Sinhala languages, including variation in syllable structures and vowel duration, it is very important to assess the performance of noise reduction tailored to the Sinhala language. This paper presents a comprehensive analysis between wavelet transformation and adaptive filters for noise reduction in Sinhala languages. We investigate the performance of ten wavelet families with soft and hard thresholding methods against adaptive filters with Normalized Least Mean Square, Least Mean Square Average Normalized Least Mean Square, Recursive Least Square, and Adaptive Filtering Averaging optimization algorithms along with cepstral and energy-based voice activity detection algorithms. The performance evaluation is done using objective metrics; Signal to Noise Ratio (SNR) and Perceptual Evaluation of Speech Quality (PESQ) and a subjective metric; Mean Opinion Score (MOS). A newly recorded Sinhala language audio dataset and the NOIZEUS database by the University of Texas, Dallas were used for the evaluation. Our code is available at https://github.com/ChathukiKet/Evaluation-of-Noise-Reduction-Methods