SLSungho Lee

Sungho Lee

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Sungho Lee has a total of 153 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Few-step Adversarial Schrödinger Bridge for Generative Speech Enhancement

    by Seungu Han, Sungho Lee, Juheon Lee, Kyogu Lee

    Deep generative models have recently been employed for speech enhancement to generate perceptually valid clean speech on large-scale datasets. Several diffusion models have been proposed, and more recently, a tractable Schrödinger Bridge has been introduced to transport between the clean and noisy speech distributions. However, these models often suffer from an iterative reverse process and require a large number of sampling steps -- more than 50. Our investigation reveals that the performance of baseline models significantly degrades when the number of sampling steps is reduced, particularly under low-SNR conditions. We propose integrating Schrödinger Bridge with GANs to effectively mitigate this issue, achieving high-quality outputs on full-band datasets while substantially reducing the required sampling steps. Experimental results demonstrate that our proposed model outperforms existing baselines, even with a single inference step, in both denoising and dereverberation tasks.

  • Open Access

    Emosical: An Emotion-Annotated Musical Theatre Dataset

    by Hayoon Kim, Ahyeon Choi, Sungho Lee, Hyun Jin Jung, Kyogu Lee

    Hayoon Kim, Ahyeon Choi, Sungho Lee, Hyun Jin Jung, Kyogu Lee. Findings of the Association for Computational Linguistics: EMNLP 2024. 2024.

  • Open Access

    Beat-Aligned Spectrogram-to-Sequence Generation of Rhythm-Game Charts

    by Jayeon Yi, Sungho Lee, Kyogu Lee

    In the heart of "rhythm games" - games where players must perform actions in sync with a piece of music - are "charts", the directives to be given to players. We newly formulate chart generation as a sequence generation task and train a Transformer using a large dataset. We also introduce tempo-informed preprocessing and training procedures, some of which are suggested to be integral for a successful training. Our model is found to outperform the baselines on a large dataset, and is also found to benefit from pretraining and finetuning.

  • Open Access

    Yet Another Generative Model For Room Impulse Response Estimation

    by Sungho Lee, Hyeong-Seok Choi, Kyogu Lee

    Recent neural room impulse response (RIR) estimators typically comprise an encoder for reference audio analysis and a generator for RIR synthesis. Especially, it is the performance of the generator that directly influences the overall estimation quality. In this context, we explore an alternate generator architecture for improved performance. We first train an autoencoder with residual quantization to learn a discrete latent token space, where each token represents a small time-frequency patch of the RIR. Then, we cast the RIR estimation problem as a reference-conditioned autoregressive token generation task, employing transformer variants that operate across frequency, time, and quantization depth axes. This way, we address the standard blind estimation task and additional acoustic matching problem, which aims to find an RIR that matches the source signal to the target signal's reverberation characteristics. Experimental results show that our system is preferable to other baselines across various evaluation metrics.

  • Open Access

    Blind Estimation of Audio Processing Graph

    by Sungho Lee, Jaehyun Park, Seungryeol Paik, Kyogu Lee

    Musicians and audio engineers sculpt and transform their sounds by connecting multiple processors, forming an audio processing graph. However, most deep-learning methods overlook this real-world practice and assume fixed graph settings. To bridge this gap, we develop a system that reconstructs the entire graph from a given reference audio. We first generate a realistic graph-reference pair dataset and train a simple blind estimation system composed of a convolutional reference encoder and a transformer-based graph decoder. We apply our model to singing voice effects and drum mixing estimation tasks. Evaluation results show that our method can reconstruct complex signal routings, including multi-band processing and sidechaining.

  • Open Access

    Global HRTF Interpolation via Learned Affine Transformation of Hyper-conditioned Features

    by Jin Woo Lee, Sungho Lee, Kyogu Lee

    Estimating Head-Related Transfer Functions (HRTFs) of arbitrary source points is essential in immersive binaural audio rendering. Computing each individual's HRTFs is challenging, as traditional approaches require expensive time and computational resources, while modern data-driven approaches are data-hungry. Especially for the data-driven approaches, existing HRTF datasets differ in spatial sampling distributions of source positions, posing a major problem when generalizing the method across multiple datasets. To alleviate this, we propose a deep learning method based on a novel conditioning architecture. The proposed method can predict an HRTF of any position by interpolating the HRTFs of known distributions. Experimental results show that the proposed architecture improves the model's generalizability across datasets with various coordinate systems. Additional demonstrations show that the model robustly reconstructs the target HRTFs from the spatially downsampled HRTFs in both quantitative and perceptual measures.

  • Open Access

    Differentiable Artificial Reverberation

    by Sungho Lee, Hyeong-Seok Choi, Kyogu Lee

    Artificial reverberation (AR) models play a central role in various audio applications. Therefore, estimating the AR model parameters (ARPs) of a reference reverberation is a crucial task. Although a few recent deep-learning-based approaches have shown promising performance, their non-end-to-end training scheme prevents them from fully exploiting the potential of deep neural networks. This motivates the introduction of differentiable artificial reverberation (DAR) models, allowing loss gradients to be back-propagated end-to-end. However, implementing the AR models with their difference equations "as is" in the deep learning framework severely bottlenecks the training speed when executed with a parallel processor like GPU due to their infinite impulse response (IIR) components. We tackle this problem by replacing the IIR filters with finite impulse response (FIR) approximations with the frequency-sampling method. Using this technique, we implement three DAR models -- differentiable Filtered Velvet Noise (FVN), Advanced Filtered Velvet Noise (AFVN), and Delay Network (DN). For each AR model, we train its ARP estimation networks for analysis-synthesis (RIR-to-ARP) and blind estimation (reverberant-speech-to-ARP) task in an end-to-end manner with its DAR model counterpart. Experiment results show that the proposed method achieves consistent performance improvement over the non-end-to-end approaches in both objective metrics and subjective listening test results.

  • Pedicled, vascularized occipital bone graft to supplement atlantoaxial arthrodesis for the treatment of pseudoarthrosis

    by Edward M. Reece, Aditya Vedantam, Sungho Lee, Mohin Bhadkamkar, Matthew Kaufman, Michael A. Bohl, Steve W. Chang, Randall W. Porter, Nicholas Theodore, U. Kumar Kakarla, et al.


  • Open Access

    Deep Brain Stimulation for Obsessive Compulsive Disorder: Evolution of Surgical Stimulation Target Parallels Changing Model of Dysfunctional Brain Circuits

    by Patrick J. Karas, Sungho Lee, Joohi Jimenez-Shahed, Wayne K. Goodman, Ashwin Viswanathan, Sameer A. Sheth

    Obsessive compulsive disorder (OCD) is a common, disabling psychiatric disease characterized by persistent, intrusive thoughts and ritualistic, repetitive behaviors. Deep brain stimulation (DBS) is thought to alleviate OCD symptoms by modulating underlying disturbances in normal cortico-striato-thalamo-cortical (CSTC) circuitry. Stimulation of the ventral portion of the anterior limb of the internal capsule (ALIC) and underlying ventral striatum (“ventral capsule/ventral striatum” or “VC/VS” target) received U.S. FDA approval in 2009 for patients with severe, treatment-refractory OCD. Over the decades, DBS surgical outcome studies have led to an evolution in the electrical stimulation target. In parallel, advancements in neuroimaging techniques have allowed investigators to better visualize and define CSTC circuits underlying the pathophysiology of OCD. A critical analysis of these new data suggests that the therapeutic mechanism of DBS for OCD likely involves neuromodulation of a widespread cortical/subcortical network, accessible by targeting fiber bundles in the ventral ALIC that connect broad network regions. Future studies will include advances in structural and functional imaging, analysis of physiological recordings, and utilization of next-generation DBS devices. These tools will enable patient-specific optimization of DBS therapy, which will hopefully further improve outcomes.


  • Open Access

    Thoracic Duct Injury Following Cervical Spine Surgery: A Multicenter Retrospective Review

    by Adeeb Derakhshan, Daniel Lubelski, Michael P. Steinmetz, Mark Corriveau, Sungho Lee, Jonathan R. Pace, Gabriel A. Smith, Ziya Gokaslan, Mohamad Bydon, Paul M. Arnold, et al.

    Study Design: Multicenter retrospective case series. Objective: To determine the rate of thoracic duct injury during cervical spine operations. Methods: A retrospective case series study was conducted among 21 high-volume surgical centers to identify instances of thoracic duct injury during anterior cervical spine surgery. Staff at each center abstracted data for each identified case into case report forms. All case report forms were collected by the AOSpine North America Clinical Research Network Methodological Core for data processing, cleaning, and analysis. Results: Of a total of 9591 patients reviewed that underwent cervical spine surgery, 2 (0.02%) incurred iatrogenic injury to the thoracic duct. Both patients underwent a left-sided anterior cervical discectomy and fusion. The interruption of the thoracic duct was addressed intraoperatively in one patient with no residual postoperative effects. The second individual developed a chylous fluid collection approximately 2 months after the operation that required drainage via needle aspiration. Conclusions: Damage to the thoracic duct during cervical spine surgery is a relatively rare occurrence. Rapid identification of the disruption of this lymphatic vessel is critical to minimize deleterious effects of this complication.