arXiv subjects

Chang Zeng

Publications and source records attributed to Chang Zeng.

31 records · Page 2Linked to original sources

arXiv2024

InstructSing: High-Fidelity Singing Voice Generation via Instructing Yourself

It is challenging to accelerate the training process while ensuring both high-quality generated voices and acceptable inference speed. In this paper, we propose a novel neural vocoder called InstructSing, which can converge much faster compared with other neural vocoders while maintaining good performance by integrating differentiable digital signal processing and adversarial training. It includes one generator and two discriminators. Specifically, the generator incorporates a harmonic-plus-noise (HN) module to produce 8kHz audio as an instructive signal. Subsequently, the HN module is connected with an extended WaveNet by an UNet-based module, which transforms the output of the HN module to a latent variable sequence containing essential periodic and aperiodic information. In addition to the latent sequence, the extended WaveNet also takes the mel-spectrogram as input to generate 48kHz high-fidelity singing voices. In terms of discriminators, we combine a multi-period discriminator, as originally proposed in HiFiGAN, with a multi-resolution multi-band STFT discriminator. Notably, InstructSing achieves comparable voice quality to other neural vocoders but with only one-tenth of the training steps on a 4 NVIDIA V100 GPU machine\footnote{{Demo page: \href{https://wavelandspeech.github.io/instructsing/}{\texttt{https://wavelandspeech.github.io/inst\\ructsing/}}}}. We plan to open-source our code and pretrained model once the paper get accepted.

eess.AS↗

arXiv2024

HAM-TTS: Hierarchical Acoustic Modeling for Token-Based Zero-Shot Text-to-Speech with Model and Data Scaling

Token-based text-to-speech (TTS) models have emerged as a promising avenue for generating natural and realistic speech, yet they grapple with low pronunciation accuracy, speaking style and timbre inconsistency, and a substantial need for diverse training data. In response, we introduce a novel hierarchical acoustic modeling approach complemented by a tailored data augmentation strategy and train it on the combination of real and synthetic data, scaling the data size up to 650k hours, leading to the zero-shot TTS model with 0.8B parameters. Specifically, our method incorporates a latent variable sequence containing supplementary acoustic information based on refined self-supervised learning (SSL) discrete units into the TTS model by a predictor. This significantly mitigates pronunciation errors and style mutations in synthesized speech. During training, we strategically replace and duplicate segments of the data to enhance timbre uniformity. Moreover, a pretrained few-shot voice conversion model is utilized to generate a plethora of voices with identical content yet varied timbres. This facilitates the explicit learning of utterance-level one-to-many mappings, enriching speech diversity and also ensuring consistency in timbre. Comparative experiments (Demo page: https://anonymous.4open.science/w/ham-tts/)demonstrate our model's superiority over VALL-E in pronunciation precision and maintaining speaking style, as well as timbre continuity.

cs.SD↗

arXiv2023

CrossSinger: A Cross-Lingual Multi-Singer High-Fidelity Singing Voice Synthesizer Trained on Monolingual Singers

It is challenging to build a multi-singer high-fidelity singing voice synthesis system with cross-lingual ability by only using monolingual singers in the training stage. In this paper, we propose CrossSinger, which is a cross-lingual singing voice synthesizer based on Xiaoicesing2. Specifically, we utilize International Phonetic Alphabet to unify the representation for all languages of the training data. Moreover, we leverage conditional layer normalization to incorporate the language information into the model for better pronunciation when singers meet unseen languages. Additionally, gradient reversal layer (GRL) is utilized to remove singer biases included in lyrics since all singers are monolingual, which indicates singer's identity is implicitly associated with the text. The experiment is conducted on a combination of three singing voice datasets containing Japanese Kiritan dataset, English NUS-48E dataset, and one internal Chinese dataset. The result shows CrossSinger can synthesize high-fidelity songs for various singers with cross-lingual ability, including code-switch cases.

cs.SD↗

arXiv2023

HiFi-WaveGAN: Generative Adversarial Network with Auxiliary Spectrogram-Phase Loss for High-Fidelity Singing Voice Generation

Entertainment-oriented singing voice synthesis (SVS) requires a vocoder to generate high-fidelity (e.g. 48kHz) audio. However, most text-to-speech (TTS) vocoders cannot reconstruct the waveform well in this scenario. In this paper, we propose HiFi-WaveGAN to synthesize the 48kHz high-quality singing voices in real-time. Specifically, it consists of an Extended WaveNet served as a generator, a multi-period discriminator proposed in HiFiGAN, and a multi-resolution spectrogram discriminator borrowed from UnivNet. To better reconstruct the high-frequency part from the full-band mel-spectrogram, we incorporate a pulse extractor to generate the constraint for the synthesized waveform. Additionally, an auxiliary spectrogram-phase loss is utilized to approximate the real distribution further. The experimental results show that our proposed HiFi-WaveGAN obtains 4.23 in the mean opinion score (MOS) metric for the 48kHz SVS task, significantly outperforming other neural vocoders.

eess.AS↗

arXiv2023

Improving Generalization Ability of Countermeasures for New Mismatch Scenario by Combining Multiple Advanced Regularization Terms

The ability of countermeasure models to generalize from seen speech synthesis methods to unseen ones has been investigated in the ASVspoof challenge. However, a new mismatch scenario in which fake audio may be generated from real audio with unseen genres has not been studied thoroughly. To this end, we first use five different vocoders to create a new dataset called CN-Spoof based on the CN-Celeb1\&2 datasets. Then, we design two auxiliary objectives for regularization via meta-optimization and a genre alignment module, respectively, and combine them with the main anti-spoofing objective using learnable weights for multiple loss terms. The results on our cross-genre evaluation dataset for anti-spoofing show that the proposed method significantly improved the generalization ability of the countermeasures compared with the baseline system in the genre mismatch scenario.

eess.AS↗

arXiv2023

Beyond Universal Transformer: block reusing with adaptor in Transformer for automatic speech recognition

Transformer-based models have recently made significant achievements in the application of end-to-end (E2E) automatic speech recognition (ASR). It is possible to deploy the E2E ASR system on smart devices with the help of Transformer-based models. While these models still have the disadvantage of requiring a large number of model parameters. To overcome the drawback of universal Transformer models for the application of ASR on edge devices, we propose a solution that can reuse the block in Transformer models for the occasion of the small footprint ASR system, which meets the objective of accommodating resource limitations without compromising recognition accuracy. Specifically, we design a novel block-reusing strategy for speech Transformer (BRST) to enhance the effectiveness of parameters and propose an adapter module (ADM) that can produce a compact and adaptable model with only a few additional trainable parameters accompanying each reusing block. We conducted an experiment with the proposed method on the public AISHELL-1 corpus, and the results show that the proposed approach achieves the character error rate (CER) of 9.3%/6.63% with only 7.6M/8.3M parameters without and with the ADM, respectively. In addition, we also make a deeper analysis to show the effect of ADM in the general block-reusing method.

cs.SD↗

arXiv2023

Cross-modal Audio-visual Co-learning for Text-independent Speaker Verification

Visual speech (i.e., lip motion) is highly related to auditory speech due to the co-occurrence and synchronization in speech production. This paper investigates this correlation and proposes a cross-modal speech co-learning paradigm. The primary motivation of our cross-modal co-learning method is modeling one modality aided by exploiting knowledge from another modality. Specifically, two cross-modal boosters are introduced based on an audio-visual pseudo-siamese structure to learn the modality-transformed correlation. Inside each booster, a max-feature-map embedded Transformer variant is proposed for modality alignment and enhanced feature generation. The network is co-learned both from scratch and with pretrained models. Experimental results on the LRSLip3, GridLip, LomGridLip, and VoxLip datasets demonstrate that our proposed method achieves 60% and 20% average relative performance improvement over independently trained audio-only/visual-only and baseline fusion systems, respectively.

cs.SD↗

arXiv2022

Xiaoicesing 2: A High-Fidelity Singing Voice Synthesizer Based on Generative Adversarial Network

XiaoiceSing is a singing voice synthesis (SVS) system that aims at generating 48kHz singing voices. However, the mel-spectrogram generated by it is over-smoothing in middle- and high-frequency areas due to no special design for modeling the details of these parts. In this paper, we propose XiaoiceSing2, which can generate the details of middle- and high-frequency parts to better construct the full-band mel-spectrogram. Specifically, in order to alleviate this problem, XiaoiceSing2 adopts a generative adversarial network (GAN), which consists of a FastSpeech-based generator and a multi-band discriminator. We improve the feed-forward Transformer (FFT) block by adding multiple residual convolutional blocks in parallel with the self-attention block to balance the local and global features. The multi-band discriminator contains three sub-discriminators responsible for low-, middle-, and high-frequency parts of the mel-spectrogram, respectively. Each sub-discriminator is composed of several segment discriminators (SD) and detail discriminators (DD) to distinguish the audio from different aspects. The experiment on our internal 48kHz singing voice dataset shows XiaoiceSing2 significantly improves the quality of the singing voice over XiaoiceSing.

eess.AS↗

arXiv2022

Deep Spectro-temporal Artifacts for Detecting Synthesized Speech

The Audio Deep Synthesis Detection (ADD) Challenge has been held to detect generated human-like speech. With our submitted system, this paper provides an overall assessment of track 1 (Low-quality Fake Audio Detection) and track 2 (Partially Fake Audio Detection). In this paper, spectro-temporal artifacts were detected using raw temporal signals, spectral features, as well as deep embedding features. To address track 1, low-quality data augmentation, domain adaptation via finetuning, and various complementary feature information fusion were aggregated in our system. Furthermore, we analyzed the clustering characteristics of subsystems with different features by visualization method and explained the effectiveness of our proposed greedy fusion strategy. As for track 2, frame transition and smoothing were detected using self-supervised learning structure to capture the manipulation of PF attacks in the time domain. We ranked 4th and 5th in track 1 and track 2, respectively.

cs.SD↗

arXiv2022

Joint Speaker Encoder and Neural Back-end Model for Fully End-to-End Automatic Speaker Verification with Multiple Enrollment Utterances

Conventional automatic speaker verification systems can usually be decomposed into a front-end model such as time delay neural network (TDNN) for extracting speaker embeddings and a back-end model such as statistics-based probabilistic linear discriminant analysis (PLDA) or neural network-based neural PLDA (NPLDA) for similarity scoring. However, the sequential optimization of the front-end and back-end models may lead to a local minimum, which theoretically prevents the whole system from achieving the best optimization. Although some methods have been proposed for jointly optimizing the two models, such as the generalized end-to-end (GE2E) model and NPLDA E2E model, all of these methods are designed for use with a single enrollment utterance. In this paper, we propose a new E2E joint method for speaker verification especially designed for the practical case of multiple enrollment utterances. In order to leverage the intra-relationship among multiple enrollment utterances, our model comes equipped with frame-level and utterance-level attention mechanisms. We also utilize several data augmentation techniques, including conventional noise augmentation using MUSAN and RIRs datasets and a unique speaker embedding-level mixup strategy for better optimization.

eess.AS↗

arXiv2022

Spoofing-Aware Attention based ASV Back-end with Multiple Enrollment Utterances and a Sampling Strategy for the SASV Challenge 2022

Current state-of-the-art automatic speaker verification (ASV) systems are vulnerable to presentation attacks, and several countermeasures (CMs), which distinguish bona fide trials from spoofing ones, have been explored to protect ASV. However, ASV systems and CMs are generally developed and optimized independently without considering their inter-relationship. In this paper, we propose a new spoofing-aware ASV back-end module that efficiently computes a combined ASV score based on speaker similarity and CM score. In addition to the learnable fusion function of the two scores, the proposed back-end module has two types of attention components, scaled-dot and feed-forward self-attention, so that intra-relationship information of multiple enrollment utterances can also be learned at the same time. Moreover, a new effective trials-sampling strategy is designed for simulating new spoofing-aware verification scenarios introduced in the Spoof-Aware Speaker Verification (SASV) challenge 2022.

eess.AS↗

arXiv2021

Attention Back-end for Automatic Speaker Verification with Multiple Enrollment Utterances

Probabilistic linear discriminant analysis (PLDA) or cosine similarity have been widely used in traditional speaker verification systems as back-end techniques to measure pairwise similarities. To make better use of multiple enrollment utterances, we propose a novel attention back-end model, which can be used for both text-independent (TI) and text-dependent (TD) speaker verification, and employ scaled-dot self-attention and feed-forward self-attention networks as architectures that learn the intra-relationships of the enrollment utterances. In order to verify the proposed attention back-end, we conduct a series of experiments on CNCeleb and VoxCeleb datasets by combining it with several sate-of-the-art speaker encoders including TDNN and ResNet. Experimental results using multiple enrollment utterances on CNCeleb show that the proposed attention back-end model leads to lower EER and minDCF score than the PLDA and cosine similarity counterparts for each speaker encoder and an experiment on VoxCeleb indicate that our model can be used even for single enrollment case.

eess.AS↗

arXiv2021

Exploring Deep Learning for Joint Audio-Visual Lip Biometrics

Audio-visual (AV) lip biometrics is a promising authentication technique that leverages the benefits of both the audio and visual modalities in speech communication. Previous works have demonstrated the usefulness of AV lip biometrics. However, the lack of a sizeable AV database hinders the exploration of deep-learning-based audio-visual lip biometrics. To address this problem, we compile a moderate-size database using existing public databases. Meanwhile, we establish the DeepLip AV lip biometrics system realized with a convolutional neural network (CNN) based video module, a time-delay neural network (TDNN) based audio module, and a multimodal fusion module. Our experiments show that DeepLip outperforms traditional speaker recognition models in context modeling and achieves over 50% relative improvements compared with our best single modality baseline, with an equal error rate of 0.75% and 1.11% on the test datasets, respectively.

cs.MM↗