https://sapl.gist.ac.kr Speech and Audio Processing Lab. [2026 상반기 정기발표회] 

----------------------------------

08.18 (화) 10:30-12:00 

ㅁ주제 : 

- Subband-wise Attention-based Neural MVDR Beamforming for Dynamic Sources

- MIMO Speech Separation for Moving Sources via SRP-PHAT-Based Spatial Information Preservation

ㅁ발표자 : 강경묵

----------------------------------

08.19 (수) 14:00-15:30

ㅁ주제 : Struggling Process of Music Source Separation

ㅁ발표자 : 손주혜

15:30-17:00

ㅁ주제 : Voice Conversion

ㅁ발표자 : 창길현

----------------------------------

08.21 (금) 10:30-12:00

ㅁ주제 : Improving Music Source Separation with Instrument-Aware EQ Augmentation

ㅁ발표자 : 전제우

13:30-15:00

ㅁ주제 : BFM TTS

ㅁ발표자 : 유종문

]]> Fri, 21 Aug 2026 10:10:37 +0000 [랩 세미나] 08.14 (금) 10:30 ~ ㅁ주제 : Disentangled Authenticity Representation for Partially Deepfake Audio Localization (ICASSP 2026) ㅁ발표자 : 이현서]]> Fri, 14 Aug 2026 13:46:22 +0000 [랩 세미나] 08.07 (금) 10:30 ~ ㅁ주제 : MATCHA-TTS: A Fast TTS Architecture with Conditional Flow Matching ㅁ발표자 : 김정연]]> Fri, 07 Aug 2026 10:03:29 +0000 [랩 세미나] 07.31 (금) 10:30 ~ ㅁ주제 : CodecSep: Prompt-Driven Universal Sound Separation on Neural Audio Codec Latents ㅁ발표자 : 전제우 [신입생 세미나] 07.28 (화) 14:00 ~ ㅁ주제 : Aligner-Encoders: Self-Attention Transformers Can Be Self-Transducers ㅁ발표자 : 양예진 07.29 (수) 14:00 ~ ㅁ주제 : MEANVOICEFLOW: ONE-STEP NONPARALLEL VOICE CONVERSION WITH MEAN FLOWS (ICASSP 2026) ㅁ발표자 : 창길현]]> Fri, 31 Jul 2026 10:01:28 +0000 [랩 세미나] 07.24 (금) 10:00 ~ ㅁ주제 : Is There a Better Source Distribution than Gaussian? ㅁ발표자 : 유종문 [신입생 세미나] 07.21 (화) 14:00 ~ ㅁ주제 : MATCHA-TTS: A Fast TTS Architecture with Conditional Flow Matching (ICASSP 2024) ㅁ발표자 : 김정연 07.23 (목) 14:00 ~ ㅁ주제 : ECAPA-TDNN: Emphasized Channel Attention, Propagation and Aggregation in TDNN Based Speaker Verification (Interspeech 2020) ㅁ발표자 : 이현서]]> Fri, 24 Jul 2026 16:23:48 +0000 [랩 세미나] 07.16 (금) 14:30 ~ ㅁ주제 : DLF: Disentangled-Language-Focused Multimodal Sentiment Analysis ㅁ발표자 : 우현섭]]> Sun, 19 Jul 2026 19:43:55 +0000 [랩 세미나] 07.10 (금) 10:30 ~ ㅁ주제 : GENSE: Generative Speech Enhancement via Language Models Using Hierarchical Modeling, ICLR 2025 ㅁ발표자 : 손주혜 [신입생 세미나] 07.07 (화) 14:00 ~ ㅁ주제 : Recap of 2026 Freshman Seminar - Chapter 7. Frequency Domain Representations [4/4] ㅁ발표자 : 이현서]]> Fri, 10 Jul 2026 10:05:38 +0000 [랩 세미나] 07.03 (금) 14:00 ~ ㅁ주제 : Locally Aligned Rectified Flow Model for Speech Enhancement Towards Single-Step Diffusion (Interspeech 2024) MeanFlowSE: one-step generative speech enhancement via conditional mean flow (ICASSP 2026) ㅁ발표자 : 이성규 [신입생 세미나] 06.30 (화) 14:00 ~ ㅁ주제 : Recap of 2026 Freshman Seminar – Chapter 7. Frequency Domain Representations [2/4] ㅁ발표자 : 창길현 07.02 (목) 14:00 ~ ㅁ주제 : Recap of 2026 Freshman Seminar – Chapter 7. Frequency Domain Representations [3/4] ㅁ발표자 : 양예진]]> Thu, 02 Jul 2026 15:59:35 +0000 [랩 세미나] 06.26(금) 14:00 ~ ㅁ주제 : Target Speaker Extraction ㅁ발표자 : 한상욱 [신입생 세미나] 06.26(목) 10:30 ~ ㅁ주제 : Recap of 2026 Freshman Seminar – Chapter 7. Frequency Domain Representations [1/4] ㅁ발표자 : 김정연 ]]> Fri, 26 Jun 2026 16:05:51 +0000 Fri, 19 Jun 2026 13:49:38 +0000 [랩 세미나] 06.12(금) 16:00 ~ ㅁ주제 : To Fuse or Drop? Dual-Path Learning for Resolving Modality Conflict in Multimodal Emotion Recognition ㅁ발표자 : 우현섭 [신입생 세미나] 06.11(목) 10:00 ~ ㅁ주제 : Speech Quality Assessment (DSSP Lecture Note 07) ㅁ발표자 : 양예진]]> Fri, 12 Jun 2026 10:01:58 +0000 [랩 세미나] 06.05(금) 10:00 ~ ㅁ주제 : Domain Partitioning Meets Parameter-Efficient Fine-Tuning A Novel Method for Improved Language-Queried Audio Source Separation ㅁ발표자 : 전제우 [신입생 세미나] 06.04(목) 10:00 ~ ㅁ주제 : Homomorphic Signal Processing and Cepstral Analysis ㅁ발표자 : 이현서]]> Fri, 05 Jun 2026 09:26:17 +0000 [랩 세미나] 05.29(금) 10:30 ~

ㅁ주제 : ICASSP 2026 - "Triage Knowledge Distillation for Speaker Verification",  "Int-MeanFlow: Few-step Speech Generation with Integral Velocity Distillation" 

ㅁ발표자 : 김기혁

[신입생 세미나] 05.28(목) 10:30 ~

ㅁ주제 : 7. Frequency-Domain Representations (Part 3)

(7.9) Implementation of the FBS Method Using the FFT

(7.10) OLA Revisited

(7.11) Modifications of the STFT

ㅁ발표자 : 창길현 

]]> Tue, 02 Jun 2026 10:53:54 +0000 [랩 세미나] 05.21(금) 10:00 ~

ㅁ주제 : ICASSP 2026 - "HyFlowSE: Hybrid End-To-End Flow-Matching Speech Enhancement via Generative-Discriminative Learning", "Towards Real-Time Generative Speech Restoration with Flow-Matching"

ㅁ발표자 : 채종욱

[신입생 세미나] 05.14(목) 10:30 ~

ㅁ주제 : 7. Frequency-Domain Representations (Part 2)

7.5. Overlap Addition Method of Synthesis

7.6. Filter Bank Summation Method of Synthesis

7.7 Time-Decimated Filter Banks

7.8 Two-Channel Filter Banks

ㅁ발표자 : 김정연

]]> Tue, 02 Jun 2026 10:53:13 +0000 [랩 세미나] 05.15(금) 10:30 ~ ㅁ주제 : F5-TTS: A Fairytaler that Fakes Fluent and Faithful Speech with Flow Matching ㅁ발표자 : 유종문 [신입생 세미나] 05.14(목) 10:30 ~ ㅁ주제 : 7. Frequency-Domain Representations (Part 1) ㅁ발표자 : 이현서]]> Fri, 15 May 2026 10:35:50 +0000 [신입생 세미나] 05.07(목) 10:30 ~ ㅁ주제 : 6.4 Short-Time Zero-Crossing Rate 6.5 The Short-Time Autocorrelation Function 6.6 The Modified Short-Time Autocorrelation Function 6.7 The Short-Time Average Magnitude Difference Function ㅁ발표자 : 양예진  ]]> Thu, 07 May 2026 12:40:03 +0000 [신입생 세미나] 04.30(목) 10:30 ~ ㅁ주제 : (6.1) Introduction (6.2) Short-time analysis of speech (6.3) Short-time energy and short-time magnitude ㅁ발표자 : 창길현]]> Thu, 30 Apr 2026 09:56:00 +0000 [랩 세미나] 04.24(금) 10:30 ~ ㅁ주제 : Target Matching Based Generative Model for Speech Enhancement ㅁ발표자 : 김기혁]]> Fri, 24 Apr 2026 09:54:44 +0000 [랩 세미나] 04.17(금) 10:30 ~ ㅁ주제 : Blockwise Flow Matching: Improving Flow Matching Models For Efficient High-Quality Generation (NeurIPS 2025.) ㅁ발표자 : 채종욱 [신입생 세미나] 04.16(목) 10:00 ~ ㅁ주제 : Review of Speech Processing Lecture – STFT & Linear Prediction ㅁ발표자 : 이현서, 김정연]]> Fri, 17 Apr 2026 09:55:59 +0000 [랩 세미나] 04.10(금) 10:30 ~ ㅁ주제 : Input-Adaptive Spectral feature Compression by Sequence Modeling for Source Separation ㅁ발표자 : 손주혜 [신입생 세미나] 04.09(목) 10:30 ~ ㅁ주제 : Chapter 4.5-4.6 Auditory Models / Lecture 3. Speech Perception ㅁ발표자 : 양예진]]> Fri, 10 Apr 2026 15:29:08 +0000