Lorenzo Palestini

A novel approach to channel decorrelation for stereo Acoustic Echo Cancellation based on missing fundamental theory

Decorrelation is a well known issue in the context of Stereophonic Acoustic Echo Cancellation: it is related to the problem of uniquely identifying each pair of room acoustic paths, due to high inter-channel coherence. In this paper, a novel approach to decorrelate a stereo signal based on the missing fundamental phenomenon is proposed. An adaptive algorithm is employed to track the behavior of one of the two channels, ensuring a continuous decorrelation without affecting the stereo quality. Several results are presented comparing our approach with Masked Noise injection method in terms of Magnitude Square Coherence, Itakura Saito measure and Convergence Speed of adaptive filters in order to confirm the validity of the proposed approach.

Multipoint equalization of digital car audio systems

In this paper a novel approach for automotive sound system enhancement based on multipoint equalization is presented. Beginning from well known techniques applied to room equalization issues, a frequency domain algorithm that combines fractional octave smoothing of measured impulse responses (IRs) in multiple locations is developed. The IRs are calculated using well positioned microphones inside a car taking into account the properties of the environment. Several results are presented considering different methods to combine the IRs and confirming the validity of the approach in comparison with single point equalization.

A New Approach to Digital Audio Equalization

The design and implementation of a M-band linear phase digital audio equalizer system is presented. Beginning from analysis/synthesis filterbank, an innovative uniform and non uniform bands audio equalizer is derived using multirate properties. In literature fixed frequency response equalization has well-known problems due to algorithms implementation. The idea of this work derives from different techniques employed in filter banks to avoid aliasing in the case of adaptive filtering in each band. The effectiveness of the algorithm is shown comparing it with a simple filterbank and with an octave band equalizer based on frequency domain technique. The solution presented here has several advantages in terms of linear phase and uniform frequency response avoiding ripple between adjacent bands.

Real-time implementation of Wave Field Synthesis for sound reproduction systems

Wave field synthesis (WFS) is a reproduction technique based on a high number of loudspeakers which allows accurate reconstruction of the desired sound field. Many theoretical researches about WFS have been carried out, but its study from a practical point of view has been often ignored. In this paper we propose a real-time approach to WFS, focusing on the investigation of efficient solutions with the aim of decreasing the high computational complexity. In particular we will present two different array geometries: a rectangular one constituted by four linear arrays and a circular one and the results of their real-time implementation will be shown.

Demonstrating hArtes project approach through an Advanced Car Information System

The need for an advanced car information system has been recently emerging, able to handle issues left open by car information systems already on the market, and to overcome their limitations. In the hArtes framework an innovative approach is proposed, based on a general purpose system with support for human interfaces, an advanced remote application control (RAC) interface following a client-server paradigm, and (sub)systems implementing specific demanding functions. In this context a complex demonstrator, composed by 4 physical systems whose interfacing software is derived from Nu-Tech DSP platform, is presented able to run computational-demanding CIS applications. The proposed architecture shows several advantages, such as increased flexibility, ease of integration with personal devices, software and hardware (individually replaceable modules) upgradeability.

A Novel Approach to Active Noise Control Based on Wave Domain Adaptive Filtering

Wave Field Analysis and Synthesis are methods which permit 3D sound field recording and reproduction. They are based on the precise extrapolation and reconstruction of the desired wave field by using arrays of microphones and loudspeakers. In order to utilize these techniques in real world applications (e.g. cinema, home theatre, teleconferencing) it is necessary to apply multi-channel Digital Signal Processing algorithms, already developed for traditional systems. In this paper we approach the problem of Active Noise Control (ANC) from the mentioned point of view: we will present two cancellation schemes depending on whether the noise sources are localized in the recording room or in the reproduction environment. The two algorithms make use of the spatio-temporal transforms, on which Wave Domain Adaptive Filtering (WDAF) is based, in order to reduce the computational complexity. It will be shown through numerical simulations that it is possible to suppress noise sources localized either inside or outside the recording/reproduction area.

Advanced CIS architecture and algorithms for enhanced in-car audio listening

Car Infotainment Systems (CIS) have been gaining more and more attention by the scientific and industrial community: many different architectures have been proposed so far, and recently, some of the authors have introduced a new conceptual system, namely advanced CIS (ACIS), which seems to positively respond to the major needs of the related market. As already pointed out in previous works, several functionalities are offered within it like the audio listening experience enhancement, which suitably exploits the ACIS properties, as the employment of professional audio equipment and several PCs offering stability and flexibility with external systems. In particular the availability of several I/O audio channels and of large computing capabilities allows considering complex procedures for audio processing. In this paper, a complete set of algorithms to enhance audio reproduction with relation to the physical properties of the car environment are discussed in terms of achievable performances when made running on such ACIS (effectively implemented on a real car). Several tests have been performed on purpose and satisfactory results obtained.

In Car Audio

In the last decade automotive audio has been gaining great attention by the scientific and industrial communities. In this context, a new approach to test and develop advanced audio algorithms for an heterogeneous embedded platform has been proposed within the European hArtes project. A real audio laboratory installed in a real car (hArtes CarLab) has been developed employing professional audio equipment. The algorithms can be tested and validated on a PC exploiting each application as a plug-in of the real time NU-Tech framework. Then a set of tools (hArtes Toolchain) can be used to generate code for the embedded platform starting from the plug-in implementation. An overview of the whole system is here presented, taking into consideration a complete set of audio algorithms developed for the advanced car infotainment system (ACIS) that is composed of three main different applications regarding the In Car listening and communication experience. Starting from a high level description of the algorithms, several implementations on different levels of hardware abstraction are presented, along with empirical results on both the design process undergone and the performance results achieved.

A theoretical analysis of a buffer frame size conversion algorithm for audio applications ensuring minimum latency

In implementing digital signal processing (DSP) algorithms for audio real-time applications, one is frequently faced with problems regarding incompatibilities between the hardware buffer length (the internal buffer of a professional sound card) and the software buffer size imposed by the underlying algorithm (due to i.e. multirate or FFT constraints). This mismatch is solved by proper frame size conversion algorithms which inevitably introduce delay. In this context, this paper presents a buffering scheme together with a theoretical proof of the minimum delay property shown by it. Some examples derived from frequently encountered issues in DSP applications are reported.