arXiv subjects

Jung-Woo Choi

Publications and source records attributed to Jung-Woo Choi.

At least 19 recordsLinked to original sources

arXiv2026

SEMamba++: A General Speech Restoration Framework Leveraging Global, Local, and Periodic Spectral Patterns

General speech restoration demands techniques that can interpret complex speech structures under various distortions. While State-Space Models like SEMamba have advanced the state-of-the-art in speech denoising, they are not inherently optimized for critical speech characteristics, such as spectral periodicity or multi-resolution frequency analysis. In this work, we introduce an architecture tailored to incorporate speech-specific features as inductive biases. In particular, we propose the Global, Local, and Periodic (GLP) module, a frequency feature extraction block that effectively and efficiently leverages the properties of frequency bins. Then, we design a multi-resolution parallel time-frequency dual-processing block to capture diverse spectral patterns, and a learnable mapping to further enhance model performance. With all our ideas combined, the proposed SEMamba++ achieves the best performance among multiple baseline models while remaining computationally efficient.

eess.AS↗

arXiv2026

MeCo: One-Step MeanFlow-based Corrector for Multi-Channel Speech Separation

While discriminative models for multi-channel speech separation excel in reference-based metrics, they often exhibit suboptimal human listening quality. To address this, we propose a novel MeanFlow-based one-step generative corrector (MeCo). MeCo learns a conditional average velocity field to map discriminative estimates directly onto the clean speech manifold in a single step. To maximize one-step generation performance, we introduce Data-Space Optimization (DSO). DSO integrates an $\mathbf{x}_r$-loss, which penalizes prediction errors on longer displacement intervals to serve as a generative objective for human listening quality, with an Endpoint SI-SDR loss that directly optimizes terminal signal fidelity. Experiments demonstrate that MeCo achieves state-of-the-art (SOTA) performance with minimal computational overhead, simultaneously achieving superior signal fidelity and human listening quality in both in-domain and out-of-domain scenarios.

eess.AS↗

arXiv2026

Focus Then Listen: An Empirical Study of Plug-and-Play Audio Enhancer for Noise-Robust Large Audio Language Models

Large audio language models (LALMs) are a class of foundation models for audio understanding. Existing LALMs tend to degrade significantly in real-world noisy acoustic conditions where speech and non-speech sounds interfere. While noise-aware fine-tuning can improve robustness, it requires task-specific noisy data and expensive retraining, limiting scalability. To address this issue, we propose Focus-Then-Listen (FTL), a plug-and-play audio enhancer that improves LALMs' noise robustness. Specifically, FTL first separates the input waveform into speech and non-speech, and a modality router is applied to predict the target audio modality (e.g., speech) based on the user's instruction. Finally, a modality-aware fusion block generates a task-adaptive enhanced signal for improved downstream perception and reasoning. Experiments across multiple LALMs and tasks show that FTL improves performance across different noise levels without fine-tuning on LALMs.

cs.SD↗

arXiv2026

A Multi-Stage Separation-and-Classification Framework Guided by Complementary Acoustic-to-Semantic Clues

This report describes the system proposed for the DCASE 2026 Challenge Task 4: Spatial Semantic Segmentation of Sound Scenes (S5). Specifically, we develop a multi-stage framework in which each stage couples a separation model with a classification model. The first stage performs source separation and classification directly on the multi-channel mixture. Its outputs are then propagated to the following stage as two complementary clues that progressively refine each target estimate: (i) an enrollment clue, the separated waveform itself, serving as a low-level acoustic reference; and (ii) a class clue, the predicted label encoded as a one-hot vector. The third stage reuses the second-stage outputs under the same scheme, forming an iterative self-guided refinement process. In addition, we use a fine-grained frame-level audio embedding from an audio encoder pretrained on a large audio corpus as an additional clue to further improve the audio separation performance. On the test set, the proposed system achieves a CAPI-SDRi of 15.51 dB, a mixture accuracy of 71.09\%, and a source accuracy of 78.62\%; with an improvement of 7.02 dB, 10.38\%p and 8.22\%p compared with the challenge baseline, respectively.

eess.AS↗

arXiv2026

RAF: Relativistic Adversarial Feedback For Universal Speech Synthesis

We propose Relativistic Adversarial Feedback (RAF), a novel training objective for GAN vocoders that improves in-domain fidelity and generalization to unseen scenarios. Although modern GAN vocoders employ advanced architectures, their training objectives often fail to promote generalizable representations. RAF addresses this problem by leveraging speech self-supervised learning models to assist discriminators in evaluating sample quality, encouraging the generator to learn richer representations. Furthermore, we utilize relativistic pairing for real and fake waveforms to improve the modeling of the training data distribution. Experiments across multiple datasets show consistent gains in both objective and subjective metrics on GAN-based vocoders. Importantly, the RAF-trained BigVGAN-base outperforms the LSGAN-trained BigVGAN in perceptual quality using only 12\% of the parameters. Comparative studies further confirm the effectiveness of RAF as a training framework for GAN vocoders.

eess.AS↗

arXiv2026

Can Large Audio Language Models Understand Audio Well? Speech, Scene and Events Understanding Benchmark for LALMs

Recently, Large Audio Language Models (LALMs) have progressed rapidly, demonstrating their strong efficacy in universal audio understanding through cross-modal integration. To evaluate LALMs' audio understanding performance, researchers have proposed different benchmarks. However, key aspects for real-world interactions are underexplored in existing benchmarks, i.e., audio signals typically contain both speech and non-speech components, and energy levels of these components can vary significantly across different scenarios. Moreover, most benchmarks do not consider the joint understanding of speech, scene, and events within the same audio clip. In this work, we introduce SSEU-Bench, the first versatile audio understanding benchmark that explicitly accounts for energy differences between speech and non-speech audio, with both independent and joint understanding settings for speech, scene, and events. Furthermore, we demonstrate that some LALMs tend to underperform on certain tasks in a joint understanding setting. To address this issue, we introduce Chain-of-Thought, which effectively improves LALMs' joint audio understanding performance by decomposing complex tasks into simpler reasoning steps.

cs.SD↗

arXiv2026

Neural acoustic multipole splatting for room impulse response synthesis

Room Impulse Response (RIR) prediction at arbitrary receiver positions is essential for practical applications such as spatial audio rendering. We propose Neural Acoustic Multipole Splatting (NAMS), which synthesizes RIRs at unseen receiver positions by learning the positions of neural acoustic multipoles and predicting their emitted signals and directivities using a neural network. Representing sound fields through a combination of multipoles offers sufficient flexibility to express complex acoustic scenes while adhering to physical constraints such as the Helmholtz equation. We also introduce a pruning strategy that starts from a dense splatting of neural acoustic multipoles and progressively eliminates redundant ones during training. Experiments conducted on both real and synthetic datasets indicate that the proposed method surpasses previous approaches on most metrics while maintaining rapid inference. Ablation studies reveal that multipole splatting with pruning achieves better performance than the monopole model with just 20% of the poles.

eess.AS↗

arXiv2026

SoundCompass: Navigating Target Sound Extraction With Effective Directional Clue Integration In Complex Acoustic Scenes

Recent advances in target sound extraction (TSE) utilize directional clues derived from direction of arrival (DoA), which represent an inherent spatial property of sound available in any acoustic scene. However, previous DoA-based methods rely on hand-crafted features or discrete encodings, which lose fine-grained spatial information and limit adaptability. We propose SoundCompass, an effective directional clue integration framework centered on a Spectral Pairwise INteraction (SPIN) module that captures cross-channel spatial correlations in the complex spectrogram domain to preserve full spatial information in multichannel signals. The input feature expressed in terms of spatial correlations is fused with a DoA clue represented as spherical harmonics (SH) encoding. The fusion is carried out across overlapping frequency subbands, inheriting the benefits reported in the previous band-split architectures. We also incorporate the iterative refinement strategy, chain-of-inference (CoI), in the TSE framework, which recursively fuses DoA with sound event activation estimated from the previous inference stage. Experiments demonstrate that SoundCompass, combining SPIN, SH embedding, and CoI, robustly extracts target sources across diverse signal classes and spatial configurations.

eess.AS↗

arXiv2026

DISPATCH: Distilling Selective Patches for Speech Enhancement

In speech enhancement, knowledge distillation (KD) compresses models by transferring a high-capacity teacher's knowledge to a compact student. However, conventional KD methods train the student to mimic the teacher's output entirely, which forces the student to imitate the regions where the teacher performs poorly and to apply distillation to the regions where the student already performs well, which yields only marginal gains. We propose Distilling Selective Patches (DISPatch), a KD framework for speech enhancement that applies the distillation loss to spectrogram patches where the teacher outperforms the student, as determined by a Knowledge Gap Score. This approach guides optimization toward areas with the most significant potential for student improvement while minimizing the influence of regions where the teacher may provide unreliable instruction. Furthermore, we introduce Multi-Scale Selective Patches (MSSP), a frequency-dependent method that uses different patch sizes across low- and high-frequency bands to account for spectral heterogeneity. We incorporate DISPatch into conventional KD methods and observe consistent gains in compact students. Moreover, integrating DISPatch and MSSP into a state-of-the-art frequency-dependent KD method considerably improves performance across all metrics.

cs.SD↗

arXiv2025

DeepASA: An Object-Oriented Multi-Purpose Network for Auditory Scene Analysis

We propose DeepASA, a multi-purpose model for auditory scene analysis that performs multi-input multi-output (MIMO) source separation, dereverberation, sound event detection (SED), audio classification, and direction-of-arrival estimation (DoAE) within a unified framework. DeepASA is designed for complex auditory scenes where multiple, often similar, sound sources overlap in time and move dynamically in space. To achieve robust and consistent inference across tasks, we introduce an object-oriented processing (OOP) strategy. This approach encapsulates diverse auditory features into object-centric representations and refines them through a chain-of-inference (CoI) mechanism. The pipeline comprises a dynamic temporal kernel-based feature extractor, a transformer-based aggregator, and an object separator that yields per-object features. These features feed into multiple task-specific decoders. Our object-centric representations naturally resolve the parameter association ambiguity inherent in traditional track-wise processing. However, early-stage object separation can lead to failure in downstream ASA tasks. To address this, we implement temporal coherence matching (TCM) within the chain-of-inference, enabling multi-task fusion and iterative refinement of object features using estimated auditory parameters. We evaluate DeepASA on representative spatial audio benchmark datasets, including ASA2, MC-FUSS, and STARSS23. Experimental results show that our model achieves state-of-the-art performance across all evaluated tasks, demonstrating its effectiveness in both source separation and auditory parameter estimation under diverse spatial auditory scenes.

eess.AS↗

arXiv2025

Sound Separation and Classification with Object and Semantic Guidance

The spatial semantic segmentation task focuses on separating and classifying sound objects from multichannel signals. To achieve two different goals, conventional methods fine-tune a large classification model cascaded with the separation model and inject classified labels as separation clues for the next iteration step. However, such integration is not ideal, in that fine-tuning over a smaller dataset loses the diversity of large classification models, features from the source separation model are different from the inputs of the pretrained classifier, and injected one-hot class labels lack semantic depth, often leading to error propagation. To resolve these issues, we propose a Dual-Path Classifier (DPC) architecture that combines object features from a source separation model with semantic representations acquired from a pretrained classification model without fine-tuning. We also introduce a Semantic Clue Encoder (SCE) that enriches the semantic depth of injected clues. Our system achieves a state-of-the-art 11.19 dB CA-SDRi and enhanced semantic fidelity on the DCASE 2025 task4 evaluation set, surpassing the top-rank performance of 11.00 dB. These results highlight the effectiveness of integrating separator-derived features and rich semantic clues.

eess.AS↗

arXiv2025

Self-Guided Target Sound Extraction and Classification Through Universal Sound Separation Model and Multiple Clues

This paper introduces a multi-stage self-directed framework designed to address the spatial semantic segmentation of sound scene (S5) task in the DCASE 2025 Task 4 challenge. This framework integrates models focused on three distinct tasks: Universal Sound Separation (USS), Single-label Classification (SC), and Target Sound Extraction (TSE). Initially, USS breaks down a complex audio mixture into separate source waveforms. Each of these separated waveforms is then processed by a SC block, generating two critical pieces of information: the waveform itself and its corresponding class label. These serve as inputs for the TSE stage, which isolates the source that matches this information. Since these inputs are produced within the system, the extraction target is identified autonomously, removing the necessity for external guidance. The extracted waveform can be looped back into the classification task, creating a cycle of iterative refinement that progressively enhances both separability and labeling accuracy. We thus call our framework a multi-stage self-guided system due to these self-contained characteristics. On the official evaluation dataset, the proposed system achieves an 11.00 dB increase in class-aware signal-to-distortion ratio improvement (CA-SDRi) and a 55.8\% accuracy in label prediction, outperforming the ResUNetK baseline by 4.4 dB and 4.3\%, respectively, and achieving first place among all submissions.

eess.AS↗

arXiv2025

CST-former: Multidimensional Attention-based Transformer for Sound Event Localization and Detection in Real Scenes

Sound event localization and detection (SELD) is a task for the classification of sound events and the identification of direction of arrival (DoA) utilizing multichannel acoustic signals. For effective classification and localization, a channel-spectro-temporal transformer (CST-former) was suggested. CST-former employs multidimensional attention mechanisms across the spatial, spectral, and temporal domains to enlarge the model's capacity to learn the domain information essential for event detection and DoA estimation over time. In this work, we present an enhanced version of CST-former with multiscale unfolded local embedding (MSULE) developed to capture and aggregate domain information over multiple time-frequency scales. Also, we propose finetuning and post-processing techniques beneficial for conducting the SELD task over limited training datasets. In-depth ablation studies of the proposed architecture and detailed analysis on the proposed modules are carried out to validate the efficacy of multidimensional attentions on the SELD task. Empirical validation through experimentation on STARSS22 and STARSS23 datasets demonstrates the remarkable performance of CST-former and post-processing techniques without using external data.

eess.AS↗

arXiv2024

EchoScan: Scanning Complex Room Geometries via Acoustic Echoes

Accurate estimation of indoor space geometries is vital for constructing precise digital twins, whose broad industrial applications include navigation in unfamiliar environments and efficient evacuation planning, particularly in low-light conditions. This study introduces EchoScan, a deep neural network model that utilizes acoustic echoes to perform room geometry inference. Conventional sound-based techniques rely on estimating geometry-related room parameters such as wall position and room size, thereby limiting the diversity of inferable room geometries. Contrarily, EchoScan overcomes this limitation by directly inferring room floorplan maps and height maps, thereby enabling it to handle rooms with complex shapes, including curved walls. The segmentation task for predicting floorplan and height maps enables the model to leverage both low- and high-order reflections. The use of high-order reflections further allows EchoScan to infer complex room shapes when some walls of the room are unobservable from the position of an audio device. Herein, EchoScan was trained and evaluated using RIRs synthesized from complex environments, including the Manhattan and Atlanta layouts, employing a practical audio device configuration compatible with commercial, off-the-shelf devices.

cs.SD↗

arXiv2024

Multichannel-to-Multichannel Target Sound Extraction Using Direction and Timestamp Clues

We propose a multichannel-to-multichannel target sound extraction (M2M-TSE) framework for separating multichannel target signals from a multichannel mixture of sound sources. Target sound extraction (TSE) isolates a specific target signal using user-provided clues, typically focusing on single-channel extraction with class labels or temporal activation maps. However, to preserve and utilize spatial information in multichannel audio signals, it is essential to extract multichannel signals of a target sound source. Moreover, the clue for extraction can also include spatial or temporal cues like direction-of-arrival (DoA) or timestamps of source activation. To address these challenges, we present an M2M framework that extracts a multichannel sound signal based on spatio-temporal clues. We demonstrate that our transformer-based architecture can successively accomplish the M2M-TSE task for multichannel signals synthesized from audio signals of diverse classes in different room environments. Furthermore, we show that the multichannel extraction task introduces sufficient inductive bias in the DNN, allowing it to directly handle DoA clues without utilizing hand-crafted spatial features.

eess.AS↗

arXiv2024

Speech-Declipping Transformer with Complex Spectrogram and Learnerble Temporal Features

We present a transformer-based speech-declipping model that effectively recovers clipped signals across a wide range of input signal-to-distortion ratios (SDRs). While recent time-domain deep neural network (DNN)-based declippers have outperformed traditional handcrafted and spectrogram-based DNN approaches, they still struggle with low-SDR inputs. To address this, we incorporate a transformer-based architecture that operates in the time-frequency (TF) domain. The TF-transformer architecture has demonstrated remarkable performance in the speech enhancement task for low-SDR signals but cannot be optimal for the time-domain artifact like clipping. To overcome the limitations of spectrogram-based DNNs, we design an extra convolutional block that directly extracts temporal features from time-domain waveforms. The joint analysis of complex spectrogram and learned temporal features allows the model to improve performance on both high- and low-SDR inputs. Our approach also preserves the unclipped portions of the speech signal during processing, preventing degradation typically seen when only spectral information is used. In evaluations on the VoiceBank-DEMAND and DNS challenge datasets, the proposed model consistently outperformed state-of-the-art (SOTA) declipping models across various metrics, demonstrating its robustness and generalizability.

eess.AS↗

arXiv2024

DeFT-Mamba: Universal Multichannel Sound Separation and Polyphonic Audio Classification

This paper presents a framework for universal sound separation and polyphonic audio classification, addressing the challenges of separating and classifying individual sound sources in a multichannel mixture. The proposed framework, DeFT-Mamba, utilizes the dense frequency-time attentive network (DeFTAN) combined with Mamba to extract sound objects, capturing the local time-frequency relations through gated convolution block and the global time-frequency relations through position-wise Hybrid Mamba. DeFT-Mamba surpasses existing separation and classification networks by a large margin, particularly in complex scenarios involving in-class polyphony. Additionally, a classification-based source counting method is introduced to identify the presence of multiple sources, outperforming conventional threshold-based approaches. Separation refinement tuning is also proposed to improve performance further. The proposed framework is trained and tested on a multichannel universal sound separation dataset developed in this work, designed to mimic realistic environments with moving sources and varying onsets and offsets of polyphonic events.

eess.AS↗

arXiv2024

RGI-Net: 3D Room Geometry Inference from Room Impulse Responses With Hidden First-Order Reflections

Room geometry is important prior information for implementing realistic 3D audio rendering. For this reason, various room geometry inference (RGI) methods have been developed by utilizing the time-of-arrival (TOA) or time-difference-of-arrival (TDOA) information in room impulse responses (RIRs). However, the conventional RGI technique poses several assumptions, such as convex room shapes, the number of walls known in priori, and the visibility of first-order reflections. In this work, we introduce the RGI-Net which can estimate room geometries without the aforementioned assumptions. RGI-Net learns and exploits complex relationships between low-order and high-order reflections in RIRs and, thus, can estimate room shapes even when the shape is non-convex or first-order reflections are missing in the RIRs. RGI-Net includes the evaluation network that separately evaluates the presence probability of walls, so the geometry inference is possible without prior knowledge of the number of walls.

eess.AS↗