QuarkAudio Technical Report

Many existing audio processing and generation models rely on task-specific architectures, resulting in fragmented development efforts and limited extensibility. It is therefore promising to design a unified framework capable of handling multiple tasks, while providing robust instruction and audio understanding and high-quality audio generation. This requires a compatible paradigm design, a powerful backbone, and a high-fidelity audio reconstruction module. To meet these requirements, this technical report introduces QuarkAudio, a decoder-only autoregressive (AR) LM-based generative framework that unifies multiple tasks. The framework includes a unified discrete audio tokenizer, H-Codec, which incorporates self-supervised learning (SSL) representations into the tokenization and reconstruction process. We further propose several improvements to H-Codec, such as a dynamic frame-rate mechanism and extending the audio sampling rate to 48 kHz. QuarkAudio unifies tasks by using task-specific conditional information as the conditioning sequence of the decoder-only LM, and predicting discrete target audio tokens in an AR manner. The framework supports a wide range of audio processing and generation tasks, including speech restoration (SR), target speaker extraction (TSE), speech separation (SS), voice conversion (VC), and language-queried audio source separation (LASS). In addition, we extend downstream tasks to universal free-form audio editing guided by natural language instructions (including speech semantic editing and audio event editing). Experimental results show that H-Codec achieves high-quality audio reconstruction with a low frame rate, improving both the efficiency and performance of downstream audio generation, and that QuarkAudio delivers competitive or comparable performance to state-of-the-art task-specific or multi-task systems across multiple tasks.

UniSE: A Unified Framework for Decoder-only Autoregressive LM-based Speech Enhancement

The development of neural audio codecs (NACs) has largely promoted applications of language models (LMs) to speech processing and understanding. However, there lacks the verification on the effectiveness of autoregressive (AR) LMbased models in unifying different sub-tasks of speech enhancement (SE). In this work, we propose UniSE, a unified decoder-only LM-based framework to handle different SE tasks including speech restoration, target speaker extraction and speech separation. It takes input speech features as conditions and generates discrete tokens of the target speech using AR modeling, which facilitates a compatibility between distinct learning patterns of multiple tasks. Experiments on several benchmarks indicate the proposed UniSE can achieve competitive performance compared to discriminative and generative baselines, showing the capacity of LMs in unifying SE tasks. The demo page is available here: https://github.com/hyyan2k/UniSE.

A Composite Predictive-Generative Approach to Monaural Universal Speech Enhancement

It is promising to design a single model that can suppress various distortions and improve speech quality, i.e., universal speech enhancement (USE). Compared to supervised learning-based predictive methods, diffusion-based generative models have shown greater potential due to the generative capacities from degraded speech with severely damaged information. However, artifacts may be introduced in highly adverse conditions, and diffusion models often suffer from a heavy computational burden due to many steps for inference. In order to jointly leverage the superiority of prediction and generation and overcome the respective defects, in this work we propose a universal speech enhancement model called PGUSE by combining predictive and generative modeling. Our model consists of two branches: the predictive branch directly predicts clean samples from degraded signals, while the generative branch optimizes the denoising objective of diffusion models. We utilize the output fusion and truncated diffusion scheme to effectively integrate predictive and generative modeling, where the former directly combines results from both branches and the latter modifies the reverse diffusion process with initial estimates from the predictive branch. Extensive experiments on several datasets verify the superiority of the proposed model over state-of-the-art baselines, demonstrating the complementarity and benefits of combining predictive and generative modeling.

The USTC-NERCSLIP Systems for The ICMC-ASR Challenge

This report describes the submitted system to the In-Car Multi-Channel Automatic Speech Recognition (ICMC-ASR) challenge, which considers the ASR task with multi-speaker overlapping and Mandarin accent dynamics in the ICMC case. We implement the front-end speaker diarization using the self-supervised learning representation based multi-speaker embedding and beamforming using the speaker position, respectively. For ASR, we employ an iterative pseudo-label generation method based on fusion model to obtain text labels of unsupervised data. To mitigate the impact of accent, an Accent-ASR framework is proposed, which captures pronunciation-related accent features at a fine-grained level and linguistic information at a coarse-grained level. On the ICMC-ASR eval set, the proposed system achieves a CER of 13.16% on track 1 and a cpCER of 21.48% on track 2, which significantly outperforms the official baseline system and obtains the first rank on both tracks.