Llasa: Scaling Train-Time and Inference-Time Compute for Llama-based Speech Synthesis

Recent advances in text-based large language models (LLMs), particularly in the GPT series and the o1 model, have demonstrated the effectiveness of scaling both training-time and inference-time compute. However, current state-of-the-art TTS systems leveraging LLMs are often multi-stage, requiring separate models (e.g., diffusion models after LLM), complicating the decision of whether to scale a particular model during training or testing. This work makes the following contributions: First, we explore the scaling of train-time and inference-time compute for speech synthesis. Second, we propose a simple framework Llasa for speech synthesis that employs a single-layer vector quantizer (VQ) codec and a single Transformer architecture to fully align with standard LLMs such as Llama. Our experiments reveal that scaling train-time compute for Llasa consistently improves the naturalness of synthesized speech and enables the generation of more complex and accurate prosody patterns. Furthermore, from the perspective of scaling inference-time compute, we employ speech understanding models as verifiers during the search, finding that scaling inference-time compute shifts the sampling modes toward the preferences of specific verifiers, thereby improving emotional expressiveness, timbre consistency, and content accuracy. In addition, we released the checkpoint and training code for our TTS model (1B, 3B, 8B) and codec model publicly available.

CosyAudio: Improving Audio Generation with Confidence Scores and Synthetic Captions

Text-to-Audio (TTA) generation is an emerging area within AI-generated content (AIGC), where audio is created from natural language descriptions. Despite growing interest, developing robust TTA models remains challenging due to the scarcity of well-labeled datasets and the prevalence of noisy or inaccurate captions in large-scale, weakly labeled corpora. To address these challenges, we propose CosyAudio, a novel framework that utilizes confidence scores and synthetic captions to enhance the quality of audio generation. CosyAudio consists of two core components: AudioCapTeller and an audio generator. AudioCapTeller generates synthetic captions for audio and provides confidence scores to evaluate their accuracy. The audio generator uses these synthetic captions and confidence scores to enable quality-aware audio generation. Additionally, we introduce a self-evolving training strategy that iteratively optimizes CosyAudio across both well-labeled and weakly-labeled datasets. Initially trained with well-labeled data, AudioCapTeller leverages its assessment capabilities on weakly-labeled datasets for high-quality filtering and reinforcement learning, which further improves its performance. The well-trained AudioCapTeller refines corpora by generating new captions and confidence scores, serving for the audio generator training. Extensive experiments on open-source datasets demonstrate that CosyAudio outperforms existing models in automated audio captioning, generates more faithful audio, and exhibits strong generalization across diverse scenarios.

ZSVC: Zero-shot Style Voice Conversion with Disentangled Latent Diffusion Models and Adversarial Training

Style voice conversion aims to transform the speaking style of source speech into a desired style while keeping the original speaker's identity. However, previous style voice conversion approaches primarily focus on well-defined domains such as emotional aspects, limiting their practical applications. In this study, we present ZSVC, a novel Zero-shot Style Voice Conversion approach that utilizes a speech codec and a latent diffusion model with speech prompting mechanism to facilitate in-context learning for speaking style conversion. To disentangle speaking style and speaker timbre, we introduce information bottleneck to filter speaking style in the source speech and employ Uncertainty Modeling Adaptive Instance Normalization (UMAdaIN) to perturb the speaker timbre in the style prompt. Moreover, we propose a novel adversarial training strategy to enhance in-context learning and improve style similarity. Experiments conducted on 44,000 hours of speech data demonstrate the superior performance of ZSVC in generating speech with diverse speaking styles in zero-shot scenarios.

Autoregressive Speech Synthesis with Next-Distribution Prediction

We introduce KALL-E, a novel autoregressive (AR) language modeling approach with next-distribution prediction for text-to-speech (TTS) synthesis. Unlike existing methods, KALL-E directly models and predicts the continuous speech distribution conditioned on text without relying on VAE- or diffusion-based components. Specifically, we use WaveVAE to extract continuous speech distributions from waveforms instead of using discrete speech tokens. A single AR language model predicts these continuous speech distributions from text, with a Kullback-Leibler divergence loss as the constraint. Experimental results show that KALL-E outperforms open-source implementations of YourTTS, VALL-E, NaturalSpeech 2, and CosyVoice in terms of naturalness and speaker similarity in zero-shot TTS scenarios. Moreover, KALL-E demonstrates exceptional zero-shot capabilities in emotion and accent cloning. Importantly, KALL-E presents a more straightforward and effective paradigm for using continuous speech representations in TTS. Audio samples are available at: \url{https://zxf-icpc.github.io/kalle/}.

YingSound: Video-Guided Sound Effects Generation with Multi-modal Chain-of-Thought Controls

Generating sound effects for product-level videos, where only a small amount of labeled data is available for diverse scenes, requires the production of high-quality sounds in few-shot settings. To tackle the challenge of limited labeled data in real-world scenes, we introduce YingSound, a foundation model designed for video-guided sound generation that supports high-quality audio generation in few-shot settings. Specifically, YingSound consists of two major modules. The first module uses a conditional flow matching transformer to achieve effective semantic alignment in sound generation across audio and visual modalities. This module aims to build a learnable audio-visual aggregator (AVA) that integrates high-resolution visual features with corresponding audio features at multiple stages. The second module is developed with a proposed multi-modal visual-audio chain-of-thought (CoT) approach to generate finer sound effects in few-shot settings. Finally, an industry-standard video-to-audio (V2A) dataset that encompasses various real-world scenarios is presented. We show that YingSound effectively generates high-quality synchronized sounds across diverse conditional inputs through automated evaluations and human studies. Project Page: \url{https://giantailab.github.io/yingsound/}

CoDiff-VC: A Codec-Assisted Diffusion Model for Zero-shot Voice Conversion

Zero-shot voice conversion (VC) aims to convert the original speaker's timbre to any target speaker while keeping the linguistic content. Current mainstream zero-shot voice conversion approaches depend on pre-trained recognition models to disentangle linguistic content and speaker representation. This results in a timbre residue within the decoupled linguistic content and inadequacies in speaker representation modeling. In this study, we propose CoDiff-VC, an end-to-end framework for zero-shot voice conversion that integrates a speech codec and a diffusion model to produce high-fidelity waveforms. Our approach involves employing a single-codebook codec to separate linguistic content from the source speech. To enhance content disentanglement, we introduce Mix-Style layer normalization (MSLN) to perturb the original timbre. Additionally, we incorporate a multi-scale speaker timbre modeling approach to ensure timbre consistency and improve voice detail similarity. To improve speech quality and speaker similarity, we introduce dual classifier-free guidance, providing both content and timbre guidance during the generation process. Objective and subjective experiments affirm that CoDiff-VC significantly improves speaker similarity, generating natural and higher-quality speech.

The NPU-HWC System for the ISCSLP 2024 Inspirational and Convincing Audio Generation Challenge

This paper presents the NPU-HWC system submitted to the ISCSLP 2024 Inspirational and Convincing Audio Generation Challenge 2024 (ICAGC). Our system consists of two modules: a speech generator for Track 1 and a background audio generator for Track 2. In Track 1, we employ Single-Codec to tokenize the speech into discrete tokens and use a language-model-based approach to achieve zero-shot speaking style cloning. The Single-Codec effectively decouples timbre and speaking style at the token level, reducing the acoustic modeling burden on the autoregressive language model. Additionally, we use DSPGAN to upsample 16 kHz mel-spectrograms to high-fidelity 48 kHz waveforms. In Track 2, we propose a background audio generator based on large language models (LLMs). This system produces scene-appropriate accompaniment descriptions, synthesizes background audio with Tango 2, and integrates it with the speech generated by our Track 1 system. Our submission achieves the second place and the first place in Track 1 and Track 2 respectively.

Vec-Tok-VC+: Residual-enhanced Robust Zero-shot Voice Conversion with Progressive Constraints in a Dual-mode Training Strategy

Zero-shot voice conversion (VC) aims to transform source speech into arbitrary unseen target voice while keeping the linguistic content unchanged. Recent VC methods have made significant progress, but semantic losses in the decoupling process as well as training-inference mismatch still hinder conversion performance. In this paper, we propose Vec-Tok-VC+, a novel prompt-based zero-shot VC model improved from Vec-Tok Codec, achieving voice conversion given only a 3s target speaker prompt. We design a residual-enhanced K-Means decoupler to enhance the semantic content extraction with a two-layer clustering process. Besides, we employ teacher-guided refinement to simulate the conversion process to eliminate the training-inference mismatch, forming a dual-mode training strategy. Furthermore, we design a multi-codebook progressive loss function to constrain the layer-wise output of the model from coarse to fine to improve speaker similarity and content accuracy. Objective and subjective evaluations demonstrate that Vec-Tok-VC+ outperforms the strong baselines in naturalness, intelligibility, and speaker similarity.

Text-aware and Context-aware Expressive Audiobook Speech Synthesis

Recent advances in text-to-speech have significantly improved the expressiveness of synthetic speech. However, a major challenge remains in generating speech that captures the diverse styles exhibited by professional narrators in audiobooks without relying on manually labeled data or reference speech. To address this problem, we propose a text-aware and context-aware(TACA) style modeling approach for expressive audiobook speech synthesis. We first establish a text-aware style space to cover diverse styles via contrastive learning with the supervision of the speech style. Meanwhile, we adopt a context encoder to incorporate cross-sentence information and the style embedding obtained from text. Finally, we introduce the context encoder to two typical TTS models, VITS-based TTS and language model-based TTS. Experimental results demonstrate that our proposed approach can effectively capture diverse styles and coherent prosody, and consequently improves naturalness and expressiveness in audiobook speech synthesis.

Single-Codec: Single-Codebook Speech Codec towards High-Performance Speech Generation

The multi-codebook speech codec enables the application of large language models (LLM) in TTS but bottlenecks efficiency and robustness due to multi-sequence prediction. To avoid this obstacle, we propose Single-Codec, a single-codebook single-sequence codec, which employs a disentangled VQ-VAE to decouple speech into a time-invariant embedding and a phonetically-rich discrete sequence. Furthermore, the encoder is enhanced with 1) contextual modeling with a BLSTM module to exploit the temporal information, 2) a hybrid sampling module to alleviate distortion from upsampling and downsampling, and 3) a resampling module to encourage discrete units to carry more phonetic information. Compared with multi-codebook codecs, e.g., EnCodec and TiCodec, Single-Codec demonstrates higher reconstruction quality with a lower bandwidth of only 304bps. The effectiveness of Single-Code is further validated by LLM-TTS experiments, showing improved naturalness and intelligibility.