Yanqing Liu

Chengyi Wang, Sanyuan Chen, Yu Wu et al. · microsoft-research

We introduce a language modeling approach for text to speech synthesis (TTS). Specifically, we train a neural codec language model (called Vall-E) using discrete codes derived from an off-the-shelf neural audio codec model, and regard TTS as a conditional language modeling task rather than continuous signal regression as in previous work. During the pre-training stage, we scale up the TTS training data to 60K hours of English speech which is hundreds of times larger than existing systems. Vall-E emerges in-context learning capabilities and can be used to synthesize high-quality personalized speech with only a 3-second enrolled recording of an unseen speaker as an acoustic prompt. Experiment results show that Vall-E significantly outperforms the state-of-the-art zero-shot TTS system in terms of speech naturalness and speaker similarity. In addition, we find Vall-E could preserve the speaker's emotion and acoustic environment of the acoustic prompt in synthesis. See https://aka.ms/valle for demos of our work.

Ziqiang Zhang, Long Zhou, Chengyi Wang et al. · microsoft-research

We propose a cross-lingual neural codec language model, VALL-E X, for cross-lingual speech synthesis. Specifically, we extend VALL-E and train a multi-lingual conditional codec language model to predict the acoustic token sequences of the target language speech by using both the source language speech and the target language text as prompts. VALL-E X inherits strong in-context learning capabilities and can be applied for zero-shot cross-lingual text-to-speech synthesis and zero-shot speech-to-speech translation tasks. Experimental results show that it can generate high-quality speech in the target language via just one speech utterance in the source language as a prompt while preserving the unseen speaker's voice, emotion, and acoustic environment. Moreover, VALL-E X effectively alleviates the foreign accent problems, which can be controlled by a language ID. Audio samples are available at \url{https://aka.ms/vallex}.

Xu Tan, Jiawei Chen, Haohe Liu et al. · microsoft-research

Text to speech (TTS) has made rapid progress in both academia and industry in recent years. Some questions naturally arise that whether a TTS system can achieve human-level quality, how to define/judge that quality and how to achieve it. In this paper, we answer these questions by first defining the human-level quality based on the statistical significance of subjective measure and introducing appropriate guidelines to judge it, and then developing a TTS system called NaturalSpeech that achieves human-level quality on a benchmark dataset. Specifically, we leverage a variational autoencoder (VAE) for end-to-end text to waveform generation, with several key modules to enhance the capacity of the prior from text and reduce the complexity of the posterior from speech, including phoneme pre-training, differentiable duration modeling, bidirectional prior/posterior modeling, and a memory mechanism in VAE. Experiment evaluations on popular LJSpeech dataset show that our proposed NaturalSpeech achieves -0.01 CMOS (comparative mean opinion score) to human recordings at the sentence level, with Wilcoxon signed rank test at p-level p >> 0.05, which demonstrates no statistically significant difference from human recordings for the first time on this dataset.

Kai Shen, Zeqian Ju, Xu Tan et al. · microsoft-research

Scaling text-to-speech (TTS) to large-scale, multi-speaker, and in-the-wild datasets is important to capture the diversity in human speech such as speaker identities, prosodies, and styles (e.g., singing). Current large TTS systems usually quantize speech into discrete tokens and use language models to generate these tokens one by one, which suffer from unstable prosody, word skipping/repeating issue, and poor voice quality. In this paper, we develop NaturalSpeech 2, a TTS system that leverages a neural audio codec with residual vector quantizers to get the quantized latent vectors and uses a diffusion model to generate these latent vectors conditioned on text input. To enhance the zero-shot capability that is important to achieve diverse speech synthesis, we design a speech prompting mechanism to facilitate in-context learning in the diffusion model and the duration/pitch predictor. We scale NaturalSpeech 2 to large-scale datasets with 44K hours of speech and singing data and evaluate its voice quality on unseen speakers. NaturalSpeech 2 outperforms previous TTS systems by a large margin in terms of prosody/timbre similarity, robustness, and voice quality in a zero-shot setting, and performs novel zero-shot singing synthesis with only a speech prompt. Audio samples are available at https://speechresearch.github.io/naturalspeech2.

Yichong Leng, Zhifang Guo, Kai Shen et al. · microsoft-research

Speech conveys more information than text, as the same word can be uttered in various voices to convey diverse information. Compared to traditional text-to-speech (TTS) methods relying on speech prompts (reference speech) for voice variability, using text prompts (descriptions) is more user-friendly since speech prompts can be hard to find or may not exist at all. TTS approaches based on the text prompt face two main challenges: 1) the one-to-many problem, where not all details about voice variability can be described in the text prompt, and 2) the limited availability of text prompt datasets, where vendors and large cost of data labeling are required to write text prompts for speech. In this work, we introduce PromptTTS 2 to address these challenges with a variation network to provide variability information of voice not captured by text prompts, and a prompt generation pipeline to utilize the large language models (LLM) to compose high quality text prompts. Specifically, the variation network predicts the representation extracted from the reference speech (which contains full information about voice variability) based on the text prompt representation. For the prompt generation pipeline, it generates text prompts for speech with a speech language understanding model to recognize voice attributes (e.g., gender, speed) from speech and a large language model to formulate text prompts based on the recognition results. Experiments on a large-scale (44K hours) speech dataset demonstrate that compared to the previous works, PromptTTS 2 generates voices more consistent with text prompts and supports the sampling of diverse voice variability, thereby offering users more choices on voice generation. Additionally, the prompt generation pipeline produces high-quality text prompts, eliminating the large labeling cost. The demo page of PromptTTS 2 is available online.

Yanqing Liu, Jianyang Gu, Kai Wang et al.

Dataset distillation aims to synthesize small datasets with little information loss from original large-scale ones for reducing storage and training costs. Recent state-of-the-art methods mainly constrain the sample synthesis process by matching synthetic images and the original ones regarding gradients, embedding distributions, or training trajectories. Although there are various matching objectives, currently the strategy for selecting original images is limited to naive random sampling. We argue that random sampling overlooks the evenness of the selected sample distribution, which may result in noisy or biased matching targets. Besides, the sample diversity is also not constrained by random sampling. These factors together lead to optimization instability in the distilling process and degrade the training efficiency. Accordingly, we propose a novel matching strategy named as \textbf{D}ataset distillation by \textbf{RE}present\textbf{A}tive \textbf{M}atching (DREAM), where only representative original images are selected for matching. DREAM is able to be easily plugged into popular dataset distillation frameworks and reduce the distilling iterations by more than 8 times without performance drop. Given sufficient training time, DREAM further provides significant improvements and achieves state-of-the-art performances.

Yanqing Liu, Jianyang Gu, Kai Wang et al.

Dataset distillation plays a crucial role in creating compact datasets with similar training performance compared with original large-scale ones. This is essential for addressing the challenges of data storage and training costs. Prevalent methods facilitate knowledge transfer by matching the gradients, embedding distributions, or training trajectories of synthetic images with those of the sampled original images. Although there are various matching objectives, currently the strategy for selecting original images is limited to naive random sampling. We argue that random sampling overlooks the evenness of the selected sample distribution, which may result in noisy or biased matching targets. Besides, the sample diversity is also not constrained by random sampling. Additionally, current methods predominantly focus on single-dimensional matching, where information is not fully utilized. To address these challenges, we propose a novel matching strategy called Dataset Distillation by Bidirectional REpresentAtive Matching (DREAM+), which selects representative original images for bidirectional matching. DREAM+ is applicable to a variety of mainstream dataset distillation frameworks and significantly reduces the number of distillation iterations by more than 15 times without affecting performance. Given sufficient training time, DREAM+ can further improve the performance and achieve state-of-the-art results. We have released the code at github.com/NUS-HPC-AI-Lab/DREAM+.

Lingwei Meng, Long Zhou, Shujie Liu et al.

We present MELLE, a novel continuous-valued token based language modeling approach for text-to-speech synthesis (TTS). MELLE autoregressively generates continuous mel-spectrogram frames directly from text condition, bypassing the need for vector quantization, which is typically designed for audio compression and sacrifices fidelity compared to continuous representations. Specifically, (i) instead of cross-entropy loss, we apply regression loss with a proposed spectrogram flux loss function to model the probability distribution of the continuous-valued tokens; (ii) we have incorporated variational inference into MELLE to facilitate sampling mechanisms, thereby enhancing the output diversity and model robustness. Experiments demonstrate that, compared to the two-stage codec language model VALL-E and its variants, the single-stage MELLE mitigates robustness issues by avoiding the inherent flaws of sampling vector-quantized codes, achieves superior performance across multiple metrics, and, most importantly, offers a more streamlined paradigm. The demos of our work are provided at https://aka.ms/melle.

Xiaoqiang Wang, Yanqing Liu, Jinyu Li et al.

Contextual biasing is an important and challenging task for end-to-end automatic speech recognition (ASR) systems, which aims to achieve better recognition performance by biasing the ASR system to particular context phrases such as person names, music list, proper nouns, etc. Existing methods mainly include contextual LM biasing and adding bias encoder into end-to-end ASR models. In this work, we introduce a novel approach to do contextual biasing by adding a contextual spelling correction model on top of the end-to-end ASR system. We incorporate contextual information into a sequence-to-sequence spelling correction model with a shared context encoder. Our proposed model includes two different mechanisms: autoregressive (AR) and non-autoregressive (NAR). We propose filtering algorithms to handle large-size context lists, and performance balancing mechanisms to control the biasing degree of the model. We demonstrate the proposed model is a general biasing solution which is domain-insensitive and can be adopted in different scenarios. Experiments show that the proposed method achieves as much as 51% relative word error rate (WER) reduction over ASR system and outperforms traditional biasing methods. Compared to the AR solution, the proposed NAR model reduces model size by 43.2% and speeds up inference by 2.1 times.

Xiaoqiang Wang, Yanqing Liu, Jinyu Li et al.

We previously proposed contextual spelling correction (CSC) to correct the output of end-to-end (E2E) automatic speech recognition (ASR) models with contextual information such as name, place, etc. Although CSC has achieved reasonable improvement in the biasing problem, there are still two drawbacks for further accuracy improvement. First, due to information limitation in text only hypothesis or weak performance of ASR model on rare domains, the CSC model may fail to correct phrases with similar pronunciation or anti-context cases where all biasing phrases are not present in the utterance. Second, there is a discrepancy between the training and inference of CSC. The bias list in training is randomly selected but in inference there may be more similarity between ground truth phrase and other phrases. To solve above limitations, in this paper we propose an improved non-autoregressive (NAR) spelling correction model for contextual biasing in E2E neural transducer-based ASR systems to improve the previous CSC model from two perspectives: Firstly, we incorporate acoustics information with an external attention as well as text hypotheses into CSC to better distinguish target phrase from dissimilar or irrelevant phrases. Secondly, we design a semantic aware data augmentation schema in training phrase to reduce the mismatch between training and inference to further boost the biasing accuracy. Experiments show that the improved method outperforms the baseline ASR+Biasing system by as much as 20.3% relative name recall gain and achieves stable improvement compared to the previous CSC method over different bias list name coverage ratio.

Yanqing Liu, Kai Wang, Wenqi Shao et al.

Visual-language pre-training has achieved remarkable success in many multi-modal tasks, largely attributed to the availability of large-scale image-text datasets. In this work, we demonstrate that Multi-modal Large Language Models (MLLMs) can enhance visual-language representation learning by establishing richer image-text associations for image-text datasets. Our approach is simple, utilizing MLLMs to extend multiple diverse captions for each image. To prevent the bias introduced by MLLMs' hallucinations and monotonous language styles, we propose "text shearing" to maintain the quality and availability of extended captions. In image-text retrieval, without introducing additional training cost, our method consistently obtains 5.6 ~ 35.0 and 16.8 ~ 46.1 improvement on Recall@1 under the fine-tuning and zero-shot settings, respectively. Notably, we obtain zero-shot results that are comparable to fine-tuning on target datasets, which encourages more exploration of the versatile use of MLLMs.

Zijun Wang, Haoqin Tu, Yuhan Wang et al.

This paper introduces STAR-1, a high-quality, just-1k-scale safety dataset specifically designed for large reasoning models (LRMs) like DeepSeek-R1. Built on three core principles -- diversity, deliberative reasoning, and rigorous filtering -- STAR-1 aims to address the critical needs for safety alignment in LRMs. Specifically, we begin by integrating existing open-source safety datasets from diverse sources. Then, we curate safety policies to generate policy-grounded deliberative reasoning samples. Lastly, we apply a GPT-4o-based safety scoring system to select training examples aligned with best practices. Experimental results show that fine-tuning LRMs with STAR-1 leads to an average 40% improvement in safety performance across four benchmarks, while only incurring a marginal decrease (e.g., an average of 1.1%) in reasoning ability measured across five reasoning tasks. Extensive ablation studies further validate the importance of our design principles in constructing STAR-1 and analyze its efficacy across both LRMs and traditional LLMs. Our project page is https://ucsc-vlaa.github.io/STAR-1.

Jiaqi Liu, Zipeng Ling, Shi Qiu et al.

AI agents increasingly operate over extended time horizons, yet their ability to retain, organize, and recall multimodal experiences remains a critical bottleneck. Building effective lifelong memory requires navigating a vast design space spanning architecture, retrieval strategies, prompt engineering, and data pipelines; this space is too large and interconnected for manual exploration or traditional AutoML to explore effectively. We deploy an autonomous research pipeline to discover OmniMem, a unified multimodal memory framework for lifelong AI agents. Starting from a naïve baseline (F1=0.117 on LoCoMo), the pipeline autonomously executes ${\sim}50$ experiments across two benchmarks, diagnosing failure modes, proposing architectural modifications, and repairing data pipeline bugs, all without human intervention in the inner loop. The resulting system achieves state-of-the-art on both benchmarks, improving F1 by +411% on LoCoMo (0.117$\to$0.598) and +214% on Mem-Gallery (0.254$\to$0.797) relative to the initial configurations. Critically, the most impactful discoveries are not hyperparameter adjustments: bug fixes (+175%), architectural changes (+44%), and prompt engineering (+188\% on specific categories) each individually exceed the cumulative contribution of all hyperparameter tuning, demonstrating capabilities fundamentally beyond the reach of traditional AutoML. We provide a taxonomy of six discovery types and identify four properties that make multimodal memory particularly suited for autoresearch, offering guidance for applying autonomous research pipelines to other AI system domains. Code is available at this https://github.com/aiming-lab/OmniMem.

Zeqian Ju, Yuancheng Wang, Kai Shen et al.

While recent large-scale text-to-speech (TTS) models have achieved significant progress, they still fall short in speech quality, similarity, and prosody. Considering speech intricately encompasses various attributes (e.g., content, prosody, timbre, and acoustic details) that pose significant challenges for generation, a natural idea is to factorize speech into individual subspaces representing different attributes and generate them individually. Motivated by it, we propose NaturalSpeech 3, a TTS system with novel factorized diffusion models to generate natural speech in a zero-shot way. Specifically, 1) we design a neural codec with factorized vector quantization (FVQ) to disentangle speech waveform into subspaces of content, prosody, timbre, and acoustic details; 2) we propose a factorized diffusion model to generate attributes in each subspace following its corresponding prompt. With this factorization design, NaturalSpeech 3 can effectively and efficiently model intricate speech with disentangled subspaces in a divide-and-conquer way. Experiments show that NaturalSpeech 3 outperforms the state-of-the-art TTS systems on quality, similarity, prosody, and intelligibility, and achieves on-par quality with human recordings. Furthermore, we achieve better performance by scaling to 1B parameters and 200K hours of training data.

Yanqing Liu, Xianhang Li, Zeyu Wang et al.

Previous works show that noisy, web-crawled image-text pairs may limit vision-language pretraining like CLIP and propose learning with synthetic captions as a promising alternative. Our work continues this effort, introducing two simple yet effective designs to better leverage richly described synthetic captions. Firstly, by observing a strong inverse effect in learning with synthetic captions -- the short synthetic captions can generally lead to MUCH higher performance than full-length ones -- we therefore fed only partial synthetic captions to the text encoder. Secondly, we incorporate an autoregressive captioner to mimic the recaptioning process -- by conditioning on the paired image input and web-crawled text description, the captioner learns to predict the full-length synthetic caption generated by advanced MLLMs. Experiments show that our framework significantly improves zero-shot performance in cross-modal retrieval tasks, setting new SOTA results on MSCOCO and Flickr30K. Moreover, such trained vision encoders can enhance the visual capability of LLaVA, showing strong improvements on a range of MLLM benchmarks. Our project page is https://ucsc-vlaa.github.io/CLIPS/.

Naoyuki Kanda, Xiaofei Wang, Sefik Emre Eskimez et al.

Laughter is one of the most expressive and natural aspects of human speech, conveying emotions, social cues, and humor. However, most text-to-speech (TTS) systems lack the ability to produce realistic and appropriate laughter sounds, limiting their applications and user experience. While there have been prior works to generate natural laughter, they fell short in terms of controlling the timing and variety of the laughter to be generated. In this work, we propose ELaTE, a zero-shot TTS that can generate natural laughing speech of any speaker based on a short audio prompt with precise control of laughter timing and expression. Specifically, ELaTE works on the audio prompt to mimic the voice characteristic, the text prompt to indicate the contents of the generated speech, and the input to control the laughter expression, which can be either the start and end times of laughter, or the additional audio prompt that contains laughter to be mimicked. We develop our model based on the foundation of conditional flow-matching-based zero-shot TTS, and fine-tune it with frame-level representation from a laughter detector as additional conditioning. With a simple scheme to mix small-scale laughter-conditioned data with large-scale pre-training data, we demonstrate that a pre-trained zero-shot TTS model can be readily fine-tuned to generate natural laughter with precise controllability, without losing any quality of the pre-trained zero-shot TTS model. Through objective and subjective evaluations, we show that ELaTE can generate laughing speech with significantly higher quality and controllability compared to conventional models. See https://aka.ms/elate/ for demo samples.

Hui Wang, Shujie Liu, Lingwei Meng et al.

To advance continuous-valued token modeling and temporal-coherence enforcement, we propose FELLE, an autoregressive model that integrates language modeling with token-wise flow matching. By leveraging the autoregressive nature of language models and the generative efficacy of flow matching, FELLE effectively predicts continuous-valued tokens (mel-spectrograms). For each continuous-valued token, FELLE modifies the general prior distribution in flow matching by incorporating information from the previous step, improving coherence and stability. Furthermore, to enhance synthesis quality, FELLE introduces a coarse-to-fine flow-matching mechanism, generating continuous-valued tokens hierarchically, conditioned on the language model's output. Experimental results demonstrate the potential of incorporating flow-matching techniques in autoregressive mel-spectrogram modeling, leading to significant improvements in TTS generation quality, as shown in https://aka.ms/felle.

Yifan Yang, Shujie Liu, Jinyu Li et al.

Recent zero-shot text-to-speech (TTS) systems face a common dilemma: autoregressive (AR) models suffer from slow generation and lack duration controllability, while non-autoregressive (NAR) models lack temporal modeling and typically require complex designs. In this paper, we introduce a novel pseudo-autoregressive (PAR) codec language modeling approach that unifies AR and NAR modeling. Combining explicit temporal modeling from AR with parallel generation from NAR, PAR generates dynamic-length spans at fixed time steps. Building on PAR, we propose PALLE, a two-stage TTS system that leverages PAR for initial generation followed by NAR refinement. In the first stage, PAR progressively generates speech tokens along the time dimension, with each step predicting all positions in parallel but only retaining the left-most span. In the second stage, low-confidence tokens are iteratively refined in parallel, leveraging the global contextual information. Experiments demonstrate that PALLE, trained on LibriTTS, outperforms state-of-the-art systems trained on large-scale data, including F5-TTS, E2-TTS, and MaskGCT, on the LibriSpeech test-clean set in terms of speech quality, speaker similarity, and intelligibility, while achieving up to ten times faster inference speed. Audio samples are available at https://microsoft.com/research/project/vall-e-x/palle.

Xianhang Li, Yanqing Liu, Haoqin Tu et al.

OpenAI's CLIP, released in early 2021, have long been the go-to choice of vision encoder for building multimodal foundation models. Although recent alternatives such as SigLIP have begun to challenge this status quo, to our knowledge none are fully open: their training data remains proprietary and/or their training recipes are not released. This paper fills this gap with OpenVision, a fully-open, cost-effective family of vision encoders that match or surpass the performance of OpenAI's CLIP when integrated into multimodal frameworks like LLaVA. OpenVision builds on existing works -- e.g., CLIPS for training framework and Recap-DataComp-1B for training data -- while revealing multiple key insights in enhancing encoder quality and showcasing practical benefits in advancing multimodal models. By releasing vision encoders spanning from 5.9M to 632.1M parameters, OpenVision offers practitioners a flexible trade-off between capacity and efficiency in building multimodal models: larger models deliver enhanced multimodal performance, while smaller versions enable lightweight, edge-ready multimodal deployments.

Hui Wang, Yifan Yang, Shujie Liu et al.

Recent advances in zero-shot text-to-speech (TTS) synthesis have achieved high-quality speech generation for unseen speakers, but most systems remain unsuitable for real-time applications because of their offline design. Current streaming TTS paradigms often rely on multi-stage pipelines and discrete representations, leading to increased computational cost and suboptimal system performance. In this work, we propose StreamMel, a pioneering single-stage streaming TTS framework that models continuous mel-spectrograms. By interleaving text tokens with acoustic frames, StreamMel enables low-latency, autoregressive synthesis while preserving high speaker similarity and naturalness. Experiments on LibriSpeech demonstrate that StreamMel outperforms existing streaming TTS baselines in both quality and latency. It even achieves performance comparable to offline systems while supporting efficient real-time generation, showcasing broad prospects for integration with real-time speech large language models. Audio samples are available at: https://aka.ms/StreamMel.

Yanqing Liu, Xianhang Li, Letian Zhang et al.

This paper provides a simplification on OpenVision's architecture and loss design for enhancing its training efficiency. Following the prior vision-language pretraining works CapPa and AIMv2, as well as modern multimodal designs like LLaVA, our changes are straightforward: we remove the text encoder (and therefore the contrastive loss), retaining only the captioning loss as a purely generative training signal. We name this new version OpenVision 2. The initial results are promising: despite this simplification, OpenVision 2 competitively matches the original model's performance on a broad set of multimodal benchmarks while substantially cutting both training time and memory consumption. For example, with ViT-L/14, it reduces training time by about 1.5x (from 83h to 57h), and memory usage by about 1.8x (from 24.5GB to 13.8GB, equivalently allowing the maximum batch size to grow from 2k to 8k). This superior training efficiency also allows us to scale far beyond the largest vision encoder used in OpenVision, reaching more than 1 billion parameters. We hold a strong belief that this lightweight, generative-only paradigm is compelling for future vision encoder development in multimodal foundation models.

Yanqing Liu, Ruiqing Xue, Chong Zhang et al.

While diffusion and autoregressive (AR) models have significantly advanced generative modeling, they each present distinct limitations. AR models, which rely on causal attention, cannot exploit future context and suffer from slow generation speeds. Conversely, diffusion models struggle with key-value (KV) caching. To overcome these challenges, we introduce Dragon-FM, a novel text-to-speech (TTS) design that unifies AR and flow-matching. This model processes 48 kHz audio codec tokens in chunks at a compact rate of 12.5 tokens per second. This design enables AR modeling across chunks, ensuring global coherence, while parallel flow-matching within chunks facilitates fast iterative denoising. Thus, the model leverages KV-cache across chunks and utilizes bidirectional context within each chunk. Furthermore, it bridges continuous and discrete feature modeling, demonstrating that continuous AR flow-matching can predict discrete tokens with finite scalar quantizers. This efficient codec and fast chunk-autoregressive architecture also make the model highly effective for generating long-form content, such as podcasts. Experiments on podcast datasets demonstrate its capability to efficiently generate high-quality zero-shot podcasts.

Letian Zhang, Sucheng Ren, Yanqing Liu et al.

This paper presents a family of advanced vision encoder, named OpenVision 3, that learns a single, unified visual representation that can serve both image understanding and image generation. Our core architecture is simple: we feed VAE-compressed image latents to a ViT encoder and train its output to support two complementary roles. First, the encoder output is passed to the ViT-VAE decoder to reconstruct the original image, encouraging the representation to capture generative structure. Second, the same representation is optimized with contrastive learning and image-captioning objectives, strengthening semantic features. By jointly optimizing reconstruction- and semantics-driven signals in a shared latent space, the encoder learns representations that synergize and generalize well across both regimes. We validate this unified design through extensive downstream evaluations with the encoder frozen. For multimodal understanding, we plug the encoder into the LLaVA-1.5 framework: it performs comparably with a standard CLIP vision encoder (e.g., 62.4 vs 62.2 on SeedBench, and 83.7 vs 82.9 on POPE). For generation, we test it under the RAE framework: ours substantially surpasses the standard CLIP-based encoder (e.g., gFID: 1.89 vs 2.54 on ImageNet). We hope this work can spur future research on unified modeling.

Midia Yousefi, Yao Qian, Junkun Chen et al.

End-to-end speech translation (ST), which translates source language speech directly into target language text, has garnered significant attention in recent years. Many ST applications require strict length control to ensure that the translation duration matches the length of the source audio, including both speech and pause segments. Previous methods often controlled the number of words or characters generated by the Machine Translation model to approximate the source sentence's length without considering the isochrony of pauses and speech segments, as duration can vary between languages. To address this, we present improvements to the duration alignment component of our sequence-to-sequence ST model. Our method controls translation length by predicting the duration of speech and pauses in conjunction with the translation process. This is achieved by providing timing information to the decoder, ensuring it tracks the remaining duration for speech and pauses while generating the translation. The evaluation on the Zh-En test set of CoVoST 2, demonstrates that the proposed Isochrony-Controlled ST achieves 0.92 speech overlap and 8.9 BLEU, which has only a 1.4 BLEU drop compared to the ST baseline.

Bing Han, Long Zhou, Shujie Liu et al.

With the help of discrete neural audio codecs, large language models (LLM) have increasingly been recognized as a promising methodology for zero-shot Text-to-Speech (TTS) synthesis. However, sampling based decoding strategies bring astonishing diversity to generation, but also pose robustness issues such as typos, omissions and repetition. In addition, the high sampling rate of audio also brings huge computational overhead to the inference process of autoregression. To address these issues, we propose VALL-E R, a robust and efficient zero-shot TTS system, building upon the foundation of VALL-E. Specifically, we introduce a phoneme monotonic alignment strategy to strengthen the connection between phonemes and acoustic sequence, ensuring a more precise alignment by constraining the acoustic tokens to match their associated phonemes. Furthermore, we employ a codec-merging approach to downsample the discrete codes in shallow quantization layer, thereby accelerating the decoding speed while preserving the high quality of speech output. Benefiting from these strategies, VALL-E R obtains controllablity over phonemes and demonstrates its strong robustness by approaching the WER of ground truth. In addition, it requires fewer autoregressive steps, with over 60% time reduction during inference. This research has the potential to be applied to meaningful projects, including the creation of speech for those affected by aphasia. Audio samples will be available at: https://aka.ms/valler.

Sanyuan Chen, Shujie Liu, Long Zhou et al.

This paper introduces VALL-E 2, the latest advancement in neural codec language models that marks a milestone in zero-shot text-to-speech synthesis (TTS), achieving human parity for the first time. Based on its predecessor, VALL-E, the new iteration introduces two significant enhancements: Repetition Aware Sampling refines the original nucleus sampling process by accounting for token repetition in the decoding history. It not only stabilizes the decoding but also circumvents the infinite loop issue. Grouped Code Modeling organizes codec codes into groups to effectively shorten the sequence length, which not only boosts inference speed but also addresses the challenges of long sequence modeling. Our experiments on the LibriSpeech and VCTK datasets show that VALL-E 2 surpasses previous systems in speech robustness, naturalness, and speaker similarity. It is the first of its kind to reach human parity on these benchmarks. Moreover, VALL-E 2 consistently synthesizes high-quality speech, even for sentences that are traditionally challenging due to their complexity or repetitive phrases. The advantages of this work could contribute to valuable endeavors, such as generating speech for individuals with aphasia or people with amyotrophic lateral sclerosis. See https://aka.ms/valle2 for demos of VALL-E 2.

Guangyan Zhang, Kaitao Song, Xu Tan et al.

Recently, leveraging BERT pre-training to improve the phoneme encoder in text to speech (TTS) has drawn increasing attention. However, the works apply pre-training with character-based units to enhance the TTS phoneme encoder, which is inconsistent with the TTS fine-tuning that takes phonemes as input. Pre-training only with phonemes as input can alleviate the input mismatch but lack the ability to model rich representations and semantic information due to limited phoneme vocabulary. In this paper, we propose MixedPhoneme BERT, a novel variant of the BERT model that uses mixed phoneme and sup-phoneme representations to enhance the learning capability. Specifically, we merge the adjacent phonemes into sup-phonemes and combine the phoneme sequence and the merged sup-phoneme sequence as the model input, which can enhance the model capacity to learn rich contextual representations. Experiment results demonstrate that our proposed Mixed-Phoneme BERT significantly improves the TTS performance with 0.30 CMOS gain compared with the FastSpeech 2 baseline. The Mixed-Phoneme BERT achieves 3x inference speedup and similar voice quality to the previous TTS pre-trained model PnG BERT

Yanqing Liu, Zhihang Xu, Gang Wang et al.

This paper describes the Microsoft end-to-end neural text to speech (TTS) system: DelightfulTTS for Blizzard Challenge 2021. The goal of this challenge is to synthesize natural and high-quality speech from text, and we approach this goal in two perspectives: The first is to directly model and generate waveform in 48 kHz sampling rate, which brings higher perception quality than previous systems with 16 kHz or 24 kHz sampling rate; The second is to model the variation information in speech through a systematic design, which improves the prosody and naturalness. Specifically, for 48 kHz modeling, we predict 16 kHz mel-spectrogram in acoustic model, and propose a vocoder called HiFiNet to directly generate 48 kHz waveform from predicted 16 kHz mel-spectrogram, which can better trade off training efficiency, modelling stability and voice quality. We model variation information systematically from both explicit (speaker ID, language ID, pitch and duration) and implicit (utterance-level and phoneme-level prosody) perspectives: 1) For speaker and language ID, we use lookup embedding in training and inference; 2) For pitch and duration, we extract the values from paired text-speech data in training and use two predictors to predict the values in inference; 3) For utterance-level and phoneme-level prosody, we use two reference encoders to extract the values in training, and use two separate predictors to predict the values in inference. Additionally, we introduce an improved Conformer block to better model the local and global dependency in acoustic model. For task SH1, DelightfulTTS achieves 4.17 mean score in MOS test and 4.35 in SMOS test, which indicates the effectiveness of our proposed system

Xiaoqiang Wang, Yanqing Liu, Sheng Zhao et al.

It's challenging to customize transducer-based automatic speech recognition (ASR) system with context information which is dynamic and unavailable during model training. In this work, we introduce a light-weight contextual spelling correction model to correct context-related recognition errors in transducer-based ASR systems. We incorporate the context information into the spelling correction model with a shared context encoder and use a filtering algorithm to handle large-size context lists. Experiments show that the model improves baseline ASR model performance with about 50% relative word error rate reduction, which also significantly outperforms the baseline method such as contextual LM biasing. The model also shows excellent performance for out-of-vocabulary terms not seen during training.

Mingjian Chen, Xu Tan, Bohan Li et al.

Custom voice, a specific text to speech (TTS) service in commercial speech platforms, aims to adapt a source TTS model to synthesize personal voice for a target speaker using few speech data. Custom voice presents two unique challenges for TTS adaptation: 1) to support diverse customers, the adaptation model needs to handle diverse acoustic conditions that could be very different from source speech data, and 2) to support a large number of customers, the adaptation parameters need to be small enough for each target speaker to reduce memory usage while maintaining high voice quality. In this work, we propose AdaSpeech, an adaptive TTS system for high-quality and efficient customization of new voices. We design several techniques in AdaSpeech to address the two challenges in custom voice: 1) To handle different acoustic conditions, we use two acoustic encoders to extract an utterance-level vector and a sequence of phoneme-level vectors from the target speech during training; in inference, we extract the utterance-level vector from a reference speech and use an acoustic predictor to predict the phoneme-level vectors. 2) To better trade off the adaptation parameters and voice quality, we introduce conditional layer normalization in the mel-spectrogram decoder of AdaSpeech, and fine-tune this part in addition to speaker embedding for adaptation. We pre-train the source TTS model on LibriTTS datasets and fine-tune it on VCTK and LJSpeech datasets (with different acoustic conditions from LibriTTS) with few adaptation data, e.g., 20 sentences, about 1 minute speech. Experiment results show that AdaSpeech achieves much better adaptation quality than baseline methods, with only about 5K specific parameters for each speaker, which demonstrates its effectiveness for custom voice. Audio samples are available at https://speechresearch.github.io/adaspeech/.

Jinyu Li, Rui Zhao, Zhong Meng et al.

Because of its streaming nature, recurrent neural network transducer (RNN-T) is a very promising end-to-end (E2E) model that may replace the popular hybrid model for automatic speech recognition. In this paper, we describe our recent development of RNN-T models with reduced GPU memory consumption during training, better initialization strategy, and advanced encoder modeling with future lookahead. When trained with Microsoft's 65 thousand hours of anonymized training data, the developed RNN-T model surpasses a very well trained hybrid model with both better recognition accuracy and lower latency. We further study how to customize RNN-T models to a new domain, which is important for deploying E2E models to practical scenarios. By comparing several methods leveraging text-only data in the new domain, we found that updating RNN-T's prediction and joint networks using text-to-speech generated from domain-specific text is the most effective.

Naihan Li, Shujie Liu, Yanqing Liu et al.

To speed up the inference of neural speech synthesis, non-autoregressive models receive increasing attention recently. In non-autoregressive models, additional durations of text tokens are required to make a hard alignment between the encoder and the decoder. The duration-based alignment plays a crucial role since it controls the correspondence between text tokens and spectrum frames and determines the rhythm and speed of synthesized audio. To get better duration-based alignment and improve the quality of non-autoregressive speech synthesis, in this paper, we propose a novel neural alignment model named MoboAligner. Given the pairs of the text and mel spectrum, MoboAligner tries to identify the boundaries of text tokens in the given mel spectrum frames based on the token-frame similarity in the neural semantic space with an end-to-end framework. With these boundaries, durations can be extracted and used in the training of non-autoregressive TTS models. Compared with the duration extracted by TransformerTTS, MoboAligner brings improvement for the non-autoregressive TTS model on MOS (3.74 comparing to FastSpeech's 3.44). Besides, MoboAligner is task-specified and lightweight, which reduces the parameter number by 45% and the training time consuming by 30%.

Naihan Li, Shujie Liu, Yanqing Liu et al.

Although end-to-end neural text-to-speech (TTS) methods (such as Tacotron2) are proposed and achieve state-of-the-art performance, they still suffer from two problems: 1) low efficiency during training and inference; 2) hard to model long dependency using current recurrent neural networks (RNNs). Inspired by the success of Transformer network in neural machine translation (NMT), in this paper, we introduce and adapt the multi-head attention mechanism to replace the RNN structures and also the original attention mechanism in Tacotron2. With the help of multi-head self-attention, the hidden states in the encoder and decoder are constructed in parallel, which improves the training efficiency. Meanwhile, any two inputs at different times are connected directly by self-attention mechanism, which solves the long range dependency problem effectively. Using phoneme sequences as input, our Transformer TTS network generates mel spectrograms, followed by a WaveNet vocoder to output the final audio results. Experiments are conducted to test the efficiency and performance of our new network. For the efficiency, our Transformer TTS network can speed up the training about 4.25 times faster compared with Tacotron2. For the performance, rigorous human tests show that our proposed model achieves state-of-the-art performance (outperforms Tacotron2 with a gap of 0.048) and is very close to human quality (4.39 vs 4.44 in MOS).