Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

VorTEX: Various overlap ratio for Target speech EXtraction

Ro-hoon Oh, Jihwan Seol, Bugeun Kim

Abstract

Target speech extraction (TSE) aims to recover a target speaker's voice from a mixture. While recent text-prompted approaches have shown promise, most approaches assume fully overlapped mixtures, limiting insight into behavior across realistic overlap ratios. We introduce VorTEX (Various overlap ratio for Target speech EXtraction), a text-prompted TSE architecture with a Decoupled Adaptive Multi-branch (DAM) Fusion block that separates primary extraction from auxiliary regularization pathways. To enable controlled analysis, we construct PORTE, a two-speaker dataset spanning overlap ratios from 0% to 100%. We further propose Suppression Ratio on Energy (SuRE), a diagnostic metric that detects suppression behavior not captured by conventional measures. Experiments show that existing models exhibit suppression or residual interference under overlap, whereas VorTEX achieves the highest separation fidelity across 20-100% overlap (e.g., 5.50 dB at 20% and 2.04 dB at 100%) while maintaining zero SuRE, indicating robust extraction without suppression-driven artifacts.

VorTEX: Various overlap ratio for Target speech EXtraction

Abstract

Target speech extraction (TSE) aims to recover a target speaker's voice from a mixture. While recent text-prompted approaches have shown promise, most approaches assume fully overlapped mixtures, limiting insight into behavior across realistic overlap ratios. We introduce VorTEX (Various overlap ratio for Target speech EXtraction), a text-prompted TSE architecture with a Decoupled Adaptive Multi-branch (DAM) Fusion block that separates primary extraction from auxiliary regularization pathways. To enable controlled analysis, we construct PORTE, a two-speaker dataset spanning overlap ratios from 0% to 100%. We further propose Suppression Ratio on Energy (SuRE), a diagnostic metric that detects suppression behavior not captured by conventional measures. Experiments show that existing models exhibit suppression or residual interference under overlap, whereas VorTEX achieves the highest separation fidelity across 20-100% overlap (e.g., 5.50 dB at 20% and 2.04 dB at 100%) while maintaining zero SuRE, indicating robust extraction without suppression-driven artifacts.
Paper Structure (19 sections, 3 equations, 3 figures, 3 tables)

This paper contains 19 sections, 3 equations, 3 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Works
  3. PORTE Dataset
  4. VorTEX Model
  5. Encoder
  6. Decoder
  7. Loss functions
  8. Experimental Setup
  9. Evaluation Metric
  10. Comparison Experiment
  11. Ablation Experiment
  12. Training Configuration
  13. Result and Discussion
  14. Perceptual Quality versus Separation Fidelity
  15. Result of Model Comparison
  16. ...and 4 more sections

Figures (3)

  • Figure 1: Architecture of VorTEX
  • Figure 2: DAM Fusion and its fusion blocks
  • Figure 3: Example mel-spectrograms