For many years, research and teaching at the IKS have been focused on speech and audio communication as well as specific aspects of mobile communication. The main field is the digital processing of speech and audio signals for mobile telephony.
The block diagram below describes the signal processing components of a typical mobile phone. The components are classified into three categories: Signal Enhancement, Coding, and Modulation. Further information can be obtained by following the links below.
- The beamformer makes use of a microphone array in order to amplify speech from the direction of the speaker and attenuate interfering sounds from other directions.
- The echo canceller reduces acoustical echoes from the loudspeaker.
- Noise reduction attenuates the recorded background noise.
- Dereverberation removes unwanted reverberation from the recorded speech.
- Far-end noise reduction attenuates background noise from the other side.
- Artificial bandwidth extension converts a received narrow-band signal into a wide-band signal by estimating missing parts of the spectrum.
- Intelligibility enhancement adapts the received speech based on the local noise characteristics such that it becomes more intelligible.
Coding and Modulation
- The source encoder compresses the speech signal and therefore reduces the data rate necessary for the transmission.
- The channel encoder adds redundancy in order to make the signal robust with respect to possible transmission errors.
- The signal is modulated onto the carrier frequency.
- The channel decoder exploits the redundancy added by the channel encoder in order to detect and correct transmission errors.
- The source decoder retrieves the waveform of the speech signal.
- Error concealment: In case of uncorrectable transmission errors, the affected signal frames are replaced by a repetition of previous signal frames.