Multitask Noisy Speech Enhancement System

- Speech band equalizer
- Dynamics processing
- Noise gate
- Signal level limiter
- Clipping restoration
- Noise reduction
- Noise whitening
- Blind deconvolution
- Spectrum analyser
- Time stretching
- Spectral expander
- Fourier corrector
- Neural network corrector
- Decorrelation
- Joint approximation
- Homomorphic approximation
- Reverberation
- Synchronisation
- Normalisation
Contact info


The Multitask Noisy Speech Enhancement System is a specialised software suite designed for speech recording and for improving the quality and intelligibility of recorded speech signal. The system consists of three applications.

Restorer - the main application of the system, allowing user to improve the quality and intelligibility of the recorded speech signal, thanks to a number of advanced digital signal processing procedures implemented in the application.

Recorder - a computer hard disk recorder, optimised for speech recording. The phonation detector enables recording only those parts of the signal where speech is present, which saves disk space and simplifies the restoring process.

Browser - an application which allows user to prepare sound files containing speech recordings to the restoration process. The application makes use of the multichannel (surround) system capabilities in order to obtain better listening conditions.

The system runs on personal computer with Microsoft Windows operating system (Windows 2000 or Windows XP recommended). Minimum hardware requirements:

  • Intel Pentium 2GHz / AMD AthlonXP 2000+ CPU or better
  • 512 MB RAM
  • PC soundcard - in order to achieve best results, multichannel soundcard and home theatre system are strongly recommended.

The system may be used in cases when speech intelligibility is important and it is not possible to obtain high quality speech recordings. An example applications of the system include:

  • improving the intelligibility of recordings of radio communication between pilots and ground control - extremely important if the plane crash accident is investigated (see example);
  • forensic audio - improving the intelligibility of speech recordings used as evidence during the investigation and in the court;
  • enhancement of the speech recordings made in noisy environment during public events such as lectures, speeches, etc;
  • many others.

For more detailed description of the various parts of the system, use links in the menu.