Publications

2026

1. ICASSP

   UniverSR: Unified and Versatile Audio Super-Resolution via Vocoder-Free Flow Matching

   Woongjib Choi, Sangmin Lee, Hyungseob Lim, and 1 more author

   In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2026

   Abs Code LINK Demo

   In this paper, we present a vocoder-free framework for audio super-resolution that employs a flow matching generative model to capture the conditional distribution of complex-valued spectral coefficients. Unlike conventional two-stage diffusion-based approaches that pre dict a mel-spectrogram and then rely on a pre-trained neural vocoder to synthesize waveforms, our method directly reconstructs waveforms via the inverse Short-Time Fourier Transform (iSTFT), thereby eliminating the dependence on a separate vocoder. This design not only simplifies end-to-end optimization but also overcomes a critical bottleneck of two-stage pipelines, where the final audio quality is fundamentally constrained by vocoder performance. Experiments show that our model consistently produces high-fidelity 48 kHz audio across diverse upsampling factors, achieving state-of-the-art performance on both speech and general audio datasets.

2025

1. Preprint

   SAGE-LD: Towards Scalable and Generalizable End-to-End Language Diarization via Simulated Data Augmentation

   Sangmin Lee, Woongjib Choi, Jihyun Kim, and 1 more author

   2025

   Preprint

   Abs LINK

   In this paper, we present a neural spoken language diarization model that supports an unconstrained span of languages within a single framework. Our approach integrates a learnable query-based architecture grounded in multilingual awareness, with large-scale pretraining on simulated code-switching data. By jointly leveraging these two components, our method overcomes the limitations of conventional approaches in data scarcity and architecture optimization, and generalizes effectively to real-world multilingual settings across diverse environments. Experimental results demonstrate that our approach achieves state-of-the-art performance on several language diarization benchmarks, with a relative performance improvement of 23% to 52% over previous methods. We believe that this work not only advances research in language diarization but also establishes a foundational framework for code-switching speech technologies.

2. INTERSPEECH

   Neural Spectral Band Generation for Audio Coding

   Woongjib Choi, Byeong Hyeon Kim, Hyungseob Lim, and 2 more authors

   In Proceedings of INTERSPEECH, 2025

   Abs LINK

   Spectral band replication (SBR) enables bit-efficient coding by generating high-frequency bands from the low-frequency ones. However, it only utilizes coarse spectral features upon a subband-wise signal replication, limiting adaptability to diverse acoustic signals. In this paper, we explore the efficacy of a deep neural network (DNN)-based generative approach for coding the high-frequency bands, which we call neural spectral band generation (n-SBG). Specifically, we propose a DNN-based encoder-decoder structure to extract and quantize the side information related to the high-frequency components and generate the components given both the side information and the decoded core-band signals. The whole coding pipeline is optimized with generative adversarial criteria to enable the generation of perceptually plausible sound. From experiments using AAC as the core codec, we show that the proposed method achieves a better perceptual quality than HE-AAC-v1 with much less side information.

2024

1. Domestic

   DNN-based Speech Codec Enhancement Utilizing Conditional Flow Matching

   Jaehoon Shin, Woongjib Choi, Byeong Hyeon Kim, and 2 more authors

   In Proceedings of the Korean Institute of Broadcast and Media Engineers (KIBME) Conference, 2024

   Abs LINK

   본 논문에서는 신호처리 기반 음성 코덱의 낮은 비트전송률에서의 복원 품질을 향상하기 위해 조건부 플로우 매칭 기반의 후처리 필터를 적용하고 그 효과를 확인한다. 최근 확산 기반 생성 모델을 활용하여 오디오 코덱의 손실 압축으로 인해 왜곡된 신호의 품질을 향상하는 연구가 이루어지고 있으나, 주로 신경망 기반 오디오 코덱에 편중되어 있으며 기존 상용 코덱에 대한 적용 사례는 부족하다. 본 연구에서는 플로우 매칭 기반 생성 모델을 신호처리 기반 음성 코덱의 후처리 필터로 적용하여, 낮은 비트전송률에서도 높은 비트전송률의 복원 품질에 준하는 품질 개선을 달성하고자 한다. 또한, 품질 개선 폭을 최대한 유지하면서 연산량의 효율성을 확보하기 위한 경량화 방법을 탐구하였으며, 실험을 통해 제안한 방법의 타당성을 확인하였다.

2. Domestic

   Non-Blind Speech Bandwidth Extension Using Pre-Trained Self-Supervised Features

   Woongjib Choi, Byeong Hyeon Kim, and Hong-Goo Kang

   In Proceedings of the Institute of Electronics and Information Engineers (IEIE) Conference, 2024

   Abs LINK

   Bandwidth extension (BWE) is a technique that estimates high-frequency components from provided low-frequency components. Most existing deep neural network-based BWE methods rely solely on information from the band-limited audio signal. Drawing inspiration from spectral band replication (SBR), we propose a novel BWE framework that employs self-supervised pre-trained features as conditional inputs in the BWE process. Our experiments demonstrate that the proposed approach outperforms the baseline SEANet architecture in terms of log-spectral distance (LSD) and PESQ. Additionally, we explore the methodologies for extracting and quantizing self-supervised learning (SSL) features to effectively utilize them as conditional inputs.