|
|
Università degli Studi di Milano |
|
|
|
DICo Dipartimento di Informatica e Comunicazione Via
Comelico 39 – 20135 Milano
|
|
DSP
Application Day 2006
Milano
- September 18, 2006
As in previous editions, DSP Application Day is intended to stimulate students’, researchers’, and industrial developers’ interest on digital-signal-processing and real-time-embedded-systems hardware and software technology. The conference and the seminar are designed to present and teach on state-of-art, emerging, and innovative silicon and software technologies.
Sponsor
Organization:
Mario Malcangi DICo – Università degli Studi
di Milano
Scientific Commettee:
Daniele Bagni Xilinx
Goffredo Haus DICo - Università degli Studi di
Milano
Mario Malcangi DICo - Università degli Studi
di Milano
Gian Paolo Rossi DICo - Università degli Studi di
Milano
Francesco Tisato DISCo - Università degli Studi
di Milano Bicocca
C O N F E R E N C E
9:15 Opening:
The digital design revolution: high throughput with a slow clock
Mario Malcangi -
DICo - Dipartimento di Informatica e Comunicazione
Università degli
Studi di Milano
9:30 Neural Network-based
speech synthesis
David Frontini, Mario Malcangi – DSP&RTS/DICo
– Università degli Studi di Milano
Neural
networks and other softcomputing methodologies (mainly fuzzy logic and genetic
algorithms) demonstrate to be optimal to solve pattern matching problems in
audio and speech processing applications. Very few research has been done to
target audio and speech synthesis. Softcomputing computational model can be an
optimal solution to audio synthesis above all to reduce memory and
computational power requirements. A neural network-based text-to-speech
processor has been proposed and compared to a rule-based system.
10:00 Evaluation of the
Perceived Quality of Hearing Aids
Giancarlo Vercellesi, Lorenzo Picinali, Antonio
Mancuso
LIM/DICo – Università degli Studi di Milano
The higher cultural and social level of people affected by hypoacusis, leads to a growth of the request of high quality hearing aid systems: to achieve this quality enhancement, the evolution of technologies in the hearing aid and audiometric fields is moving towards the digital world. Even if digital audio prosthesis are not new, the algorithms for the enhancement of speech quality and intelligibility are always being improved. We propose a methodology for the evaluation of the perceived quality of hearing aids: the simulation of a real two-dimensional soundscape is done using a quadraphonic loudspeaker array, then a speech signal (extracted from the Italian Audiometric Vocal Audio Archive) is reproduced from a single loudspeaker placed in different positions around the virtual soundscape, and everything is recorded through a Dummy Head microphone system with hearing aids placed on the two ears. The analysis of the signal recorded in these simulations is carried out through a complex comparison of various parameters extracted from the signals in output from the Dummy Head microphone and from the hearing aids.
10:30 DoMiNo network
media processor, the industry's first programmable, single-chip,
multi-stream, multi-format, MPEG
audio/video/system codec architecture
Gregory Maertens - LSI Logic
The LSI Logic DoMiNo network media processor is
the industry's first programmable, single-chip, multi-stream, multi-format,
MPEG audio/video/system codec (encoder/decoder) architecture. This
powerful and flexible architecture has been designed to trigger - very much
like the domino effect - an endless variety of digital audio/visual (A/V)
products for the consumer, prosumer and professional markets. LSI Logic DoMiNo
is a true system-on-a-chip architecture, reducing cost, design complexity,
power consumption, and time-to-market through its high level of system
integration and its C-Ware (TM) high-level programming environment. LSI Logic
DoMiNo has true multi-stream and multi-format A/V codec capabilities that
enable unique features on consumer electronics. The presentation will give an
overview of DoMiNo's architecture as well as some of its applications using
commercial products.
11.00 VoIP
overWLAN: a use case based on ST Nomadik chipset
Diego
Melpignano – STMicroelectronics, Advanced System Technology
This session presents an application of DSP
technology to an emerging service: Voice-over-IP in Wireless LAN networks. In this
context, energy consumption is a key factor and DSP’s contribute to
considerable savings compared to general purpose CPU. ST Nomadik ™ application
processor embeds an ARM core and two DSP’s, one for video and one for audio.
This System on Chip can be used in combination with low-power wireless
peripherals to implement a VoIP application using Wireless LAN connectivity.
The ARM core runs the signalling protocol and the main OS, whereas the audio
DSP is responsible for voice encoding and decoding. The software interface to
control Multimedia components complies with the OpenMAX Integration Layer API,
recently standardized by the Khronos Group. This talk presents the overall
architecture of the Nomadik System on Chip, discusses the characteristics of
the VoIP-WLAN application and shows the software partitioning that has been
prototyped.
11:30 Coffee break
12:00 Verification and Validation of a Simulink
Model
Rossana Pacchiodo – The MathWorks
The activity of modelling has to be supported
by a strategy of verification and validation. We say that a model is “verified
and validated”, when it’s compliant with behavioural requirements described
during the project phase (verification) and works properly under all the
possible operating conditions (validation).
Verification & validation need a strict
control of all requirements and, in the meantime, an intensive activity of
testing. Simulink is the perfect environment to link requirements written on paper
with the corresponding executable model. Moreover, it offers the more suitable
tools for testing process. The validation activity can be measured taking count
of severe criterions, keeping the same modelling environment and reusing every
step of the validation activity to test the code generated from the model.
Design choices can be documented straightaway through an automatic generation
process.
12.30 A DSP-based method for transient restoration in recorded music
Massimo Magrini*,
Tommaso Giunti*, Graziano Bertini*, Andreas Rauber**, Thomas Lidy**
*ISTI-CNR
Pisa, ** TU Wien
The set of audio treatment methods commonly used in
the mastering process of commercial music tends to increase the loudness
perception of the final audio product, obtaining a “fat” sound that can be
played with sufficient quality on low level audio devices, such as small radios
or computer loudspeaker. The side effects of these audio processes are the loss
in dynamic range and poor transients. We have proposed a method for time domain
transient restorations of recorded music, easily implementable with low cost
Digital Signal Processors i.e. in stand alone device or included in HW (iPod
like) or SW (Winamp like) audio player applications. The way to automatically
tune the effect by on-the-fly features extraction from the songs played is
under study. This goal is developed jointly with a research group of TU Wien
working in music classification and retrival, in the framework of EU MUSCLE-NOE
project.
13:00 Deployment of
operating systems and real-time kernels on new-generations DSPs
Umberto
Zangheri – ZP engineering
Modern product design always requires more
integration; a novel trend in modern DSP development is based on the usage of a
full operating system on the same processor, for sophisticated control
functions, or the utilization of a real-time, multi-tasking kernel.
This paper details some of the issues that
exist when working with DSPs and the uClinux operating system, and a comparison
of pros and cons in using a real-time kernel on high-performance processors.
13:30 Lunch
|
S E M I N A R Model-Based Prototyping for Video Processing
Systems PDF Stefano
Olivieri – The MathWorks The goal of
this presentation is to show the Model-Based Design methodology for the
implementation and verification of an application on FPGA. After
building the Simulink model of a video processing algorithm, we demonstrate
how the automatic HDL code generation tools allow to move transparently to
its implementation on a FPGA based prototyping system. 14.30 Start 16:00 Coffee Break 16:30 Continuation 17.30 End |