Noise Reduction (NR)

Speech Enhancement - Noise Reduction (NR)

Problem: Microphone Signals Contaminated by Background Noise

  • Numerous kinds of background noise possible, e.g. stationary, wind, impulsive, talkers, babble
  • No generic solution suitable for all use cases exists
  • Trade-off between noise reduction and artifacts

Approach: Single channel noise reduction

  • Combined statistical and codebook-based noise reduction
  • Online learning of unknown non-stationary noise
  • Optional fusion of multiple estimators

Results

  • Low signal delay, low complexity
  • Significantly improved speech intelligibility

 

Audio Examples

Our offer

Algorithms

  • Low complexity statistical noise reduction
  • Novel codebook-based noise reduction

Services

  • Implementation on mobile platforms

Literature

[heese16a]
Florian Heese
Speech Signal Enhancement by Information Combining
Dissertation, 2016

[nelke16a]
Christoph Matthias Nelke
Wind Noise Reduction – Signal Processing Concepts –

Dissertation, 2016

[nelke15a]
Christoph Matthias Nelke, Patrick A. Naylor, and Peter Vary
Corpus Based Reconstruction of Speech Degraded by Wind Noise
Proceedings of European Signal Processing Conference (EUSIPCO), August 2015

[nelke15]
Christoph Matthias Nelke and Peter Vary
Wind Noise Short Term Power Spectrum Estimation Using Pitch Adaptive Inverse Binary Masks
Proceedings of IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), April 2015

[heese15]
Florian Heese and Peter Vary
Noise PSD Estimation By Logarithmic Baseline Tracing
Proceedings of IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), April 2015

[heese15a]
Florian Heese, Markus Niermann, and Peter Vary
Speech-Codebook Based Soft Voice Activity Detection
Proceedings of IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), April 2015

[heese14]
Florian Heese, Christoph Matthias Nelke, Markus Niermann, and Peter Vary
Selflearning Codebook Speech Enhancement
ITG-Fachtagung Sprachkommunikation, September 2014

[nelke14]
Christoph Matthias Nelke, Navin Chatlani, Christophe Beaugeant, and Peter Vary
Single Microphone Wind Noise PSD Estimation Using Signal Centroids
Proceedings of IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), Mai 2014

[nelke13]
Christoph Matthias Nelke, Christophe Beaugeant, and Peter Vary
Dual Microphone Noise PSD Estimation for Mobile Phones in Hands-Free Position Exploiting the Coherence and Speech Presence Probability
Proceedings of IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), Mai 2013

[martin01c]
Martin, Rainer
Noise Power Spectral Density Estimation Based on Optimal Smoothing and Minimum Statistics
IEEE Transactions on Speech and Audio Processing, Juli 2001