Speech separation and speaker diarization have strong similarities. In particular with respect to end-to-end neural diarization (EEND) methods. Separation aims at extracting each speaker from overlapped speech, while diarization identifies time boundaries of speech segments produced by the same speaker. In this paper, we carry out an analysis of the use of speech separation guided diarization (SSGD) where diarization is performed simply by separating the speakers signals and applying voice activity detection. In particular we compare two speech separation (SSep) models, both in offline and online settings. In the online setting we consider both the use of continuous source separation (CSS) and causal SSep models architectures. As an additional contribution, we show a simple post-processing algorithm which reduces significantly the false alarm errors of a SSGD pipeline. We perform our experiments on Fisher Corpus Part 1 and CALLHOME datasets evaluating both separation and diarization metrics. Notably, without fine-tuning, our SSGD DPRNN-based online model achieves 12.7% DER on CALLHOME, comparable with state-of-the-art EEND models despite having considerably lower latency, i.e., 50 ms vs 1 s.
Data plays a crucial role in machine learning. However, in real-world applications, there are several problems with data, e.g., data are of low quality; a limited number of data points lead to under-fitting of the machine learning model; it is hard to access the data due to privacy, safety and regulatory concerns. \textit{Synthetic data generation} offers a promising new avenue, as it can be shared and used in ways that real-world data cannot. This paper systematically reviews the existing works that leverage machine learning models for synthetic data generation. Specifically, we discuss the synthetic data generation works from several perspectives: (i) applications, including computer vision, speech, natural language, healthcare, and business; (ii) machine learning methods, particularly neural network architectures and deep generative models; (iii) privacy and fairness issue. In addition, we identify the challenges and opportunities in this emerging field and suggest future research directions.
Unpaired data has shown to be beneficial for low-resource automatic speech recognition~(ASR), which can be involved in the design of hybrid models with multi-task training or language model dependent pre-training. In this work, we leverage unpaired data to train a general sequence-to-sequence model. Unpaired speech and text are used in the form of data pairs by generating the corresponding missing parts in prior to model training. Inspired by the complementarity of speech-PseudoLabel pair and SynthesizedAudio-text pair in both acoustic features and linguistic features, we propose a complementary joint training~(CJT) method that trains a model alternatively with two data pairs. Furthermore, label masking for pseudo-labels and gradient restriction for synthesized audio are proposed to further cope with the deviations from real data, termed as CJT++. Experimental results show that compared to speech-only training, the proposed basic CJT achieves great performance improvements on clean/other test sets, and the CJT++ re-training yields further performance enhancements. It is also apparent that the proposed method outperforms the wav2vec2.0 model with the same model size and beam size, particularly in extreme low-resource cases.
Denoising diffusion probabilistic models (DDPMs) have recently achieved leading performances in many generative tasks. However, the inherited iterative sampling process costs hindered their applications to speech synthesis. This paper proposes FastDiff, a fast conditional diffusion model for high-quality speech synthesis. FastDiff employs a stack of time-aware location-variable convolutions of diverse receptive field patterns to efficiently model long-term time dependencies with adaptive conditions. A noise schedule predictor is also adopted to reduce the sampling steps without sacrificing the generation quality. Based on FastDiff, we design an end-to-end text-to-speech synthesizer, FastDiff-TTS, which generates high-fidelity speech waveforms without any intermediate feature (e.g., Mel-spectrogram). Our evaluation of FastDiff demonstrates the state-of-the-art results with higher-quality (MOS 4.28) speech samples. Also, FastDiff enables a sampling speed of 58x faster than real-time on a V100 GPU, making diffusion models practically applicable to speech synthesis deployment for the first time. We further show that FastDiff generalized well to the mel-spectrogram inversion of unseen speakers, and FastDiff-TTS outperformed other competing methods in end-to-end text-to-speech synthesis. Audio samples are available at \url{https://FastDiff.github.io/}.
The smaller memory bandwidth in smart devices prompts development of smaller Automatic Speech Recognition (ASR) models. To obtain a smaller model, one can employ the model compression techniques. Knowledge distillation (KD) is a popular model compression approach that has shown to achieve smaller model size with relatively lesser degradation in the model performance. In this approach, knowledge is distilled from a trained large size teacher model to a smaller size student model. Also, the transducer based models have recently shown to perform well for on-device streaming ASR task, while the conformer models are efficient in handling long term dependencies. Hence in this work we employ a streaming transducer architecture with conformer as the encoder. We propose a multi-stage progressive approach to compress the conformer transducer model using KD. We progressively update our teacher model with the distilled student model in a multi-stage setup. On standard LibriSpeech dataset, our experimental results have successfully achieved compression rates greater than 60% without significant degradation in the performance compared to the larger teacher model.
Internal Language Model Estimation (ILME) based language model (LM) fusion has been shown significantly improved recognition results over conventional shallow fusion in both intra-domain and cross-domain speech recognition tasks. In this paper, we attempt to apply our ILME method to cross-domain code-switching speech recognition (CSSR) work. Specifically, our curiosity comes from several aspects. First, we are curious about how effective the ILME-based LM fusion is for both intra-domain and cross-domain CSSR tasks. We verify this with or without merging two code-switching domains. More importantly, we train an end-to-end (E2E) speech recognition model by means of merging two monolingual data sets and observe the efficacy of the proposed ILME-based LM fusion for CSSR. Experimental results on SEAME that is from Southeast Asian and another Chinese Mainland CS data set demonstrate the effectiveness of the proposed ILME-based LM fusion method.
Recently, convolution-augmented transformer (Conformer) has achieved promising performance in automatic speech recognition (ASR) and time-domain speech enhancement (SE), as it can capture both local and global dependencies in the speech signal. In this paper, we propose a conformer-based metric generative adversarial network (CMGAN) for SE in the time-frequency (TF) domain. In the generator, we utilize two-stage conformer blocks to aggregate all magnitude and complex spectrogram information by modeling both time and frequency dependencies. The estimation of magnitude and complex spectrogram is decoupled in the decoder stage and then jointly incorporated to reconstruct the enhanced speech. In addition, a metric discriminator is employed to further improve the quality of the enhanced estimated speech by optimizing the generator with respect to a corresponding evaluation score. Quantitative analysis on Voice Bank+DEMAND dataset indicates the capability of CMGAN in outperforming various previous models with a margin, i.e., PESQ of 3.41 and SSNR of 11.10 dB.
The zero-shot scenario for speech generation aims at synthesizing a novel unseen voice with only one utterance of the target speaker. Although the challenges of adapting new voices in zero-shot scenario exist in both stages -- acoustic modeling and vocoder, previous works usually consider the problem from only one stage. In this paper, we extend our previous Glow-WaveGAN to Glow-WaveGAN 2, aiming to solve the problem from both stages for high-quality zero-shot text-to-speech and any-to-any voice conversion. We first build a universal WaveGAN model for extracting latent distribution $p(z)$ of speech and reconstructing waveform from it. Then a flow-based acoustic model only needs to learn the same $p(z)$ from texts, which naturally avoids the mismatch between the acoustic model and the vocoder, resulting in high-quality generated speech without model fine-tuning. Based on a continuous speaker space and the reversible property of flows, the conditional distribution can be obtained for any speaker, and thus we can further conduct high-quality zero-shot speech generation for new speakers. We particularly investigate two methods to construct the speaker space, namely pre-trained speaker encoder and jointly-trained speaker encoder. The superiority of Glow-WaveGAN 2 has been proved through TTS and VC experiments conducted on LibriTTS corpus and VTCK corpus.
We investigate robustness properties of pre-trained neural models for automatic speech recognition. Real life data in machine learning is usually very noisy and almost never clean, which can be attributed to various factors depending on the domain, e.g. outliers, random noise and adversarial noise. Therefore, the models we develop for various tasks should be robust to such kinds of noisy data, which led to the thriving field of robust machine learning. We consider this important issue in the setting of automatic speech recognition. With the increasing popularity of pre-trained models, it's an important question to analyze and understand the robustness of such models to noise. In this work, we perform a robustness analysis of the pre-trained neural models wav2vec2, HuBERT and DistilHuBERT on the LibriSpeech and TIMIT datasets. We use different kinds of noising mechanisms and measure the model performances as quantified by the inference time and the standard Word Error Rate metric. We also do an in-depth layer-wise analysis of the wav2vec2 model when injecting noise in between layers, enabling us to predict at a high level what each layer learns. Finally for this model, we visualize the propagation of errors across the layers and compare how it behaves on clean versus noisy data. Our experiments conform the predictions of Pasad et al. [2021] and also raise interesting directions for future work.
In recent years, there has been an increased popularity in image and speech generation using diffusion models. However, directly generating music waveforms from free-form text prompts is still under-explored. In this paper, we propose the first text-to-waveform music generation model that can receive arbitrary texts using diffusion models. We incorporate the free-form textual prompt as the condition to guide the waveform generation process of diffusion models. To solve the problem of lacking such text-music parallel data, we collect a dataset of text-music pairs from the Internet with weak supervision. Besides, we compare the effect of two prompt formats of conditioning texts (music tags and free-form texts) and prove the superior performance of our method in terms of text-music relevance. We further demonstrate that our generated music in the waveform domain outperforms previous works by a large margin in terms of diversity, quality, and text-music relevance.