The Joint Workshop on Hands-free Speech Communication and Microphone Arrays has been held on May 6-8, 2008, at Castello del Buonconsiglio in Trento, Italy. It follows the HSCMA held at Rutgers University in 2005, and continues the tradition of previous workshops on Hands-free Speech Communication (HSC) and Microphone Arrays (MA). The HSCMA 2008 workshop is mainly devoted to presenting recent advances in speech and signal processing techniques based upon multi-microphone systems, and to distant-talking speech communication and human/machine interaction. It brings together researchers and practitioners from universities and industry working in robust speech recognition, speech enhancement, high-quality sound capture, or multiple-input/multiple-output (MIMO) acoustic signal processing.

The technical program has been scheduled to assign one day to multi-microphone processing methods, one day to hands-free speech communication issues, and one day to topics which are relevant either to both areas, or to new research directions.

UNIVPM (Università Politecnica delle Marche) was at HSCMA 2008, to present 3 papers related to hArtes.

Simone Cifani, Emanuele Principi, Cesare Rocchi, Stefano Squartini and Francesco Piazza
"A Multichannel Noise Reduction Front-end based on Psychoacoustics for Robust Speech Recognition in Highly Noisy Environments"

Paolo Peretti, Lorenzo Palestini, Stefania Cecchi and Francesco Piazza
"A Subband Approach to Wave Domain Adaptative Filtering"

Alessandro Bastari, Stefano Squartini and Francesco Piazza
"Joint Acoustic Feedback Cancellation and Noise Reduction within the Prediction Error Method Framework"

The past April 17th of April, Fabrizio Ferrandi with the Politecnico di Milano presented in ESA, in the Noordwijk laboriatories (Netherland), the hArtes approach to task partitioning  in a presentation withthe title"PandA: a framework for task partitioning of MPSoCs architectures".

This lecture presented the embedded system design toolchain for automatic generation of parallel code runnable on symmetric multiprocessor  systems from an initial sequential specification written using the C  language, i.e. the task partitioning toolchain currently under developement in the  Artes European project.

The lecture presented in particular the technique estimating the execution time of software through the use of source code instrumentation. This technique is based on the use of virtual Instructions (VI) associated with each C statement. Through experimental validation, latencies are associated to each VI and are used to predict the performance of unknown application programs. It was shown how this allows to take into consideration also inter-instruction interactions, thus enhancing prediction accuracy.