• Publications
  • Influence
Improved Signal-to-Noise Ratio Estimation for Speech Enhancement
TLDR
A method called two-step noise reduction (TSNR) technique is proposed which solves this problem while maintaining the benefits of the decision-directed approach and a significant improvement is brought by HRNR compared to TSNR thanks to the preservation of harmonics.
Reactance domain MUSIC algorithm for electronically steerable parasitic array radiator
TLDR
Analytic and empirical results show that high-resolution DoAs estimation can be achieved by using the reactance domain MUSIC algorithm for ESPAR antennas.
A two-step noise reduction technique
TLDR
A new method, called the two-step noise reduction (TSNR) technique, is proposed, which solves the problem of single microphone speech enhancement in noisy environments while maintaining the benefits of the decision-directed approach.
MEAN TEACHER WITH DATA AUGMENTATION FOR DCASE 2019 TASK 4 Technical Report
TLDR
A mean-teacher model with convolutional neural network (CNN) and recurrent neuralnetwork (RNN) together with data augmentation and a median window tuned for each class based on prior knowledge is proposed.
Reactance-domain MUSIC for ESPAR antennas (experiment)
TLDR
In this paper, two experimental methods of estimating the arrival signal angles are proposed and a method of solving the practical problem of calibrating the ESPAR antenna output signal model is shown.
Reactance Domain MUSIC Algorithm for ESPAR Antennas
TLDR
Simulation results show that DoAs can be estimated by the reactance domain MUSIC algorithm for ESPAR antennas.
Speech enhancement using harmonic regeneration
TLDR
A new method, called harmonic regeneration noise reduction technique, which solves the problem of single microphone speech enhancement in noisy environments by calculating a fully harmonic signal based on the distorted signal using a non-linearity to regenerate harmonics in an efficient way.
A low-complexity audio fingerprinting technique for embedded applications
TLDR
The proposed audio-based fingerprinting technology can be used for automatically identifying the program being watched by capturing the sound of the TV set and to precisely estimate the timestamp of the currently broadcast moment with respect to the beginning of the program.
Reliable a posteriori signal-to-noise ratio features selection
TLDR
A new method is proposed, called reliable features selection noise reduction (RFSNR) technique, that is able to classify the a posteriori SNR estimates into two categories: the reliable features leading to speech components and the unreliable ones corresponding to musical noise only.
Speech and audio loudness depending on telephone audio bandwidth and codec — A subjective testing approach
TLDR
A new approach for the subjective assessment of the loudness of complex audio signals such as speech or music is proposed, which shows that loudness increases with the bandwidth extension up to super-wideband.
...
...