Speech Coding And Communications Technologies

Mobile tools for communication have become characteristic feature of our age. In the industrial world, being always available, always on-line has become the new standard. A central part and the first wave of mobile communication were technologies that allowed digital transmission of speech or speech coding. Today, these technologies have become so self-evident that we hardly ever realize how important and central they are to our society and culture.

By speech coding we refer to digital transmission and storage of speech signals. It includes modelling the speech source, extraction of perceptually important features, compression and resynthesis. Traditionally speech coding has been a separate field from audio coding, but as multimedia applications found their way to mobile devices, it became clear that speech and audio should be jointly coded. This trend led in 2012 to the standardization of MPEG Unified Speech and Audio Coding, which is further developed and marketed as extended HE-AAC by the Audio and Multimedia division at Fraunhofer IIS.

The objective of the Speech Coding group headed by Prof. Dr. Tom Bäckström is research and further development of speech processing tools especially for mobile devices. Due to the strong traditions at Fraunhofer IIS, research on Unified Speech and Audio Coding is naturally the main focus of this group. It is also a field where internal co-operation with the Audio Signal Analysis group of Prof. Dr. Bernd Edler and the Audio Coding group of Prof. Dr. Jürgen Herre is strongest. Moreover, in co-operation with the Perception-Based Spatial Audio Signal Processing group of Prof. Dr. Emanuël Habets and the communication experts at Fraunhofer IIS production models to spatial pick-up of speech are being employed.

One recent example of the successful cooperation between AudioLabs and Fraunhofer IIS is the adoption of the Enhanced Voice Service (EVS) standard. EVS is an audio codec especially designed for Voice over LTE services. It enables phone calls with Full-HD Voice quality, bringing call fidelity up to the same level as today’s other digital media services. Integrating state-of-the-art speech and audio coding technology, EVS removes the limitations of bandlimited and voice-centric codecs that were previously used in mobile communications. AudioLabs research built the foundation of the Fraunhofer contributions to the EVS standard.