We present AmberNet, a compact end-to-end neural network for Spoken Language Identification. AmberNet consists of 1D depth-wise separable convolutions and Squeeze-and-Excitation layers with global context, followed by statistics pooling and linear layers. AmberNet achieves performance similar to state-of-the-art(SOTA) models on VoxLingua107 dataset, while being 10x smaller. AmberNet can be adapted to unseen languages and new acoustic conditions with simple finetuning. It attains SOTA accuracy of 75.8% on FLEURS benchmark. We show the model is easily scalable to achieve a better trade-off between accuracy and speed. We further inspect the model's sensitivity to input length and show that AmberNet performs well even on short utterances.
Automatic speech recognition models are often adapted to improve their accuracy in a new domain. A potential drawback of model adaptation to new domains is catastrophic forgetting, where the Word Error Rate on the original domain is significantly degraded. This paper addresses the situation when we want to simultaneously adapt automatic speech recognition models to a new domain and limit the degradation of accuracy on the original domain without access to the original training dataset. We propose several techniques such as a limited training strategy and regularized adapter modules for the Transducer encoder, prediction, and joiner network. We apply these methods to the Google Speech Commands and to the UK and Ireland English Dialect speech data set and obtain strong results on the new target domain while limiting the degradation on the original domain.
Inverse text normalization (ITN) is an essential post-processing step in automatic speech recognition (ASR). It converts numbers, dates, abbreviations, and other semiotic classes from the spoken form generated by ASR to their written forms. One can consider ITN as a Machine Translation task and use neural sequence-to-sequence models to solve it. Unfortunately, such neural models are prone to hallucinations that could lead to unacceptable errors. To mitigate this issue, we propose a single-pass token classifier model that regards ITN as a tagging task. The model assigns a replacement fragment to every input token or marks it for deletion or copying without changes. We present a dataset preparation method based on the granular alignment of ITN examples. The proposed model is less prone to hallucination errors. The model is trained on the Google Text Normalization dataset and achieves state-of-the-art sentence accuracy on both English and Russian test sets. One-to-one correspondence between tags and input words improves the interpretability of the model's predictions, simplifies debugging, and allows for post-processing corrections. The model is simpler than sequence-to-sequence models and easier to optimize in production settings. The model and the code to prepare the dataset is published as part of NeMo project.
Despite recent progress in generative adversarial network(GAN)-based vocoders, where the model generates raw waveform conditioned on mel spectrogram, it is still challenging to synthesize high-fidelity audio for numerous speakers across varied recording environments. In this work, we present BigVGAN, a universal vocoder that generalizes well under various unseen conditions in zero-shot setting. We introduce periodic nonlinearities and anti-aliased representation into the generator, which brings the desired inductive bias for waveform synthesis and significantly improves audio quality. Based on our improved generator and the state-of-the-art discriminators, we train our GAN vocoder at the largest scale up to 112M parameters, which is unprecedented in the literature. In particular, we identify and address the training instabilities specific to such scale, while maintaining high-fidelity output without over-regularization. Our BigVGAN achieves the state-of-the-art zero-shot performance for various out-of-distribution scenarios, including new speakers, novel languages, singing voices, music and instrumental audio in unseen (even noisy) recording environments. We will release our code and model at: https://github.com/NVIDIA/BigVGAN
Speaker diarization systems are challenged by a trade-off between the temporal resolution and the fidelity of the speaker representation. By obtaining a superior temporal resolution with an enhanced accuracy, a multi-scale approach is a way to cope with such a trade-off. In this paper, we propose a more advanced multi-scale diarization system based on a multi-scale diarization decoder. There are two main contributions in this study that significantly improve the diarization performance. First, we use multi-scale clustering as an initialization to estimate the number of speakers and obtain the average speaker representation vector for each speaker and each scale. Next, we propose the use of 1-D convolutional neural networks that dynamically determine the importance of each scale at each time step. To handle a variable number of speakers and overlapping speech, the proposed system can estimate the number of existing speakers. Our proposed system achieves a state-of-art performance on the CALLHOME and AMI MixHeadset datasets, with 3.92% and 1.05% diarization error rates, respectively.
Text normalization (TN) systems in production are largely rule-based using weighted finite-state transducers (WFST). However, WFST-based systems struggle with ambiguous input when the normalized form is context-dependent. On the other hand, neural text normalization systems can take context into account but they suffer from unrecoverable errors and require labeled normalization datasets, which are hard to collect. We propose a new hybrid approach that combines the benefits of rule-based and neural systems. First, a non-deterministic WFST outputs all normalization candidates, and then a neural language model picks the best one -- similar to shallow fusion for automatic speech recognition. While the WFST prevents unrecoverable errors, the language model resolves contextual ambiguity. The approach is easy to extend and we show it is effective. It achieves comparable or better results than existing state-of-the-art TN models.
This paper describes Mixer-TTS, a non-autoregressive model for mel-spectrogram generation. The model is based on the MLP-Mixer architecture adapted for speech synthesis. The basic Mixer-TTS contains pitch and duration predictors, with the latter being trained with an unsupervised TTS alignment framework. Alongside the basic model, we propose the extended version which additionally uses token embeddings from a pre-trained language model. Basic Mixer-TTS and its extended version achieve a mean opinion score (MOS) of 4.05 and 4.11, respectively, compared to a MOS of 4.27 of original LJSpeech samples. Both versions have a small number of parameters and enable much faster speech synthesis compared to the models with similar quality.
Training neural text-to-speech (TTS) models for a new speaker typically requires several hours of high quality speech data. Prior works on voice cloning attempt to address this challenge by adapting pre-trained multi-speaker TTS models for a new voice, using a few minutes of speech data of the new speaker. However, publicly available large multi-speaker datasets are often noisy, thereby resulting in TTS models that are not suitable for use in products. We address this challenge by proposing transfer-learning guidelines for adapting high quality single-speaker TTS models for a new speaker, using only a few minutes of speech data. We conduct an extensive study using different amounts of data for a new speaker and evaluate the synthesized speech in terms of naturalness and voice/style similarity to the target speaker. We find that fine-tuning a single-speaker TTS model on just 30 minutes of data, can yield comparable performance to a model trained from scratch on more than 27 hours of data for both male and female target speakers.
In this paper, we propose TitaNet, a novel neural network architecture for extracting speaker representations. We employ 1D depth-wise separable convolutions with Squeeze-and-Excitation (SE) layers with global context followed by channel attention based statistics pooling layer to map variable-length utterances to a fixed-length embedding (t-vector). TitaNet is a scalable architecture and achieves state-of-the-art performance on speaker verification task with an equal error rate (EER) of 0.68% on the VoxCeleb1 trial file and also on speaker diarization tasks with diarization error rate (DER) of 1.73% on AMI-MixHeadset, 1.99% on AMI-Lapel and 1.11% on CH109. Furthermore, we investigate various sizes of TitaNet and present a light TitaNet-S model with only 6M parameters that achieve near state-of-the-art results in diarization tasks.