Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

How much to Dereverberate? Low-Latency Single-Channel Speech Enhancement in Distant Microphone Scenarios

Satvik Venkatesh, Philip Coleman, Arthur Benilov, Simon Brown, Selim Sheta, Frederic Roskam

TL;DR

This work tackles real-time low-latency single-channel dereverberation for distant-microphone scenarios in large rooms. It introduces volume-based sampling of $T_{60}$ and a windowing strategy for room impulse responses to create realistic training targets, and it analyzes how preserving early reflections via an offset interacts with $T^{max}_{60}$. The approach demonstrates feasibility and achieves state-of-the-art performance for distant-mic SE, highlighting the critical roles of room volume, distance, and reflection handling. The findings offer practical guidance for data generation and parameter choices, with implications for real-world applications like lectures and stage acoustics and avenues for future domain adaptation and listening evaluations.

Abstract

Dereverberation is an important sub-task of Speech Enhancement (SE) to improve the signal's intelligibility and quality. However, it remains challenging because the reverberation is highly correlated with the signal. Furthermore, the single-channel SE literature has predominantly focused on rooms with short reverb times (typically under 1 second), smaller rooms (under volumes of 1000 cubic meters) and relatively short distances (up to 2 meters). In this paper, we explore real-time low-latency single-channel SE under distant microphone scenarios, such as 5 to 10 meters, and focus on conference rooms and theatres, with larger room dimensions and reverberation times. Such a setup is useful for applications such as lecture demonstrations, drama, and to enhance stage acoustics. First, we show that single-channel SE in such challenging scenarios is feasible. Second, we investigate the relationship between room volume and reverberation time, and demonstrate its importance when randomly simulating room impulse responses. Lastly, we show that for dereverberation with short decay times, preserving early reflections before decaying the transfer function of the room improves overall signal quality.

How much to Dereverberate? Low-Latency Single-Channel Speech Enhancement in Distant Microphone Scenarios

TL;DR

This work tackles real-time low-latency single-channel dereverberation for distant-microphone scenarios in large rooms. It introduces volume-based sampling of and a windowing strategy for room impulse responses to create realistic training targets, and it analyzes how preserving early reflections via an offset interacts with . The approach demonstrates feasibility and achieves state-of-the-art performance for distant-mic SE, highlighting the critical roles of room volume, distance, and reflection handling. The findings offer practical guidance for data generation and parameter choices, with implications for real-world applications like lectures and stage acoustics and avenues for future domain adaptation and listening evaluations.

Abstract

Dereverberation is an important sub-task of Speech Enhancement (SE) to improve the signal's intelligibility and quality. However, it remains challenging because the reverberation is highly correlated with the signal. Furthermore, the single-channel SE literature has predominantly focused on rooms with short reverb times (typically under 1 second), smaller rooms (under volumes of 1000 cubic meters) and relatively short distances (up to 2 meters). In this paper, we explore real-time low-latency single-channel SE under distant microphone scenarios, such as 5 to 10 meters, and focus on conference rooms and theatres, with larger room dimensions and reverberation times. Such a setup is useful for applications such as lecture demonstrations, drama, and to enhance stage acoustics. First, we show that single-channel SE in such challenging scenarios is feasible. Second, we investigate the relationship between room volume and reverberation time, and demonstrate its importance when randomly simulating room impulse responses. Lastly, we show that for dereverberation with short decay times, preserving early reflections before decaying the transfer function of the room improves overall signal quality.
Paper Structure (15 sections, 2 equations, 3 figures, 3 tables)

This paper contains 15 sections, 2 equations, 3 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Proposed Method
  3. Volume-based $T_{60}$ Sampling
  4. Windowing Room Impulse Responses
  5. Experimental Setup
  6. Dataset
  7. Nework Architectures
  8. Parameters for Dereverberation
  9. Results
  10. Volume-based $T_{60}$ Sampling
  11. Window Design for Dereverberation
  12. Close Microphone
  13. Distant Microphone
  14. Comparison with State-of-the-art
  15. Conclusion

Figures (3)

  • Figure 1: The proposed relationship between room volume and $T_{60}$ for conference rooms, curve-fitted with values from harris1957handbookcirillo2007acousticsothman2012influence.
  • Figure 2: Gain curves applied for reverb suppression.
  • Figure 3: DNSMOS OVRL, SIG, and BAK scores for the model trained with different dereverberation parameters tested under close and distant microphone scenarios in small and large rooms. The asterisks indicate the level of significance (*: $p < 0.05$, **: $p < 0.001$).