The aim of this project is the design of low power optimized systems fordigital signal processing of acoustic signals. Such systems are required in PRO-DASP: Power Reduction for Digital Audio Signal Processing PRO-DASP: Verlustleistungsminimierung für digitale Audio-Signalverarbeitung medical help devices, mobile communication as well as in information processing systems and for multimedia applications.

For these applications a design methodology will be developed and formalizedacross all levels of abstraction, reflecting the interactions between algorithm and architecure at this point of system- and hardware-design. Additional optimization on module- and circuit-level result in a high applicability of this methodology. This is done through development and consistent implementation of highly optimized typical signal processing components and low power driven design decisions on the architecture level.

As a typical application for low power audio signal processing in thefrequency domain an psychoacoustical motivated design is used. It consists of a filterbank-analysis with following timevariant amplify-regulation of the specific channels and next a filterbank-summation (reconstruction). The analysis part of this design is already successfully in use for different applications (for example automatic speech recognition, audio- and speech-quality measurement) and can be used in combination with the new developed resynthesis-part for many applications in the field of human-computer-interaction and telecommunication as well as a prototyp of digital hearing aids.