Andrea Primavera

A low latency implementation of a non-uniform partitioned convolution algorithm for room acoustic simulation

Finite impulse response convolution is one of the most widely used operations in digital signal processing field for filtering operations. Low computationally demanding techniques are essential for calculating convolutions with low input/output latency in real scenarios, considering that the real-time requirements are strictly related to the impulse response length. In this context, a complete overview of the state of the art relative to the algorithms for fast computation of convolution is described here. Then, a novel perceptual approach employed to reduce the computational cost of fast convolution algorithms is here presented. It is based on the pre-processing of a selected impulse response and it allows to further reduce the number of complex multiplications considering the energy decay relief and the absolute threshold of hearing, as psychoacoustic constraints. Several results are reported in terms of computational cost and perceived audio quality in order to prove the effectiveness of the proposed approach also introducing comparisons with the existing techniques of the state of the art.

An efficient DSP-based implementation of a fast convolution approach with non uniform partitioning

Finite impulse response convolution is one of the most widely used operation in digital signal processing field for filtering operations. In this context, low computationally demanding techniques become essential for calculating convolutions with low input/output latency in real scenarios, considering that the real time requirements are strictly related to the impulse response length. In this context, an efficient DSP implementation of a fast convolution approach is presented with the aim of lowering the workload required in applications like reverberation. It is based on a non uniform partitioning of the impulse response and a psychoacoustic technique derived from the human ear sensitivity. Several results are reported in order to prove the effectiveness of the proposed approach also introducing comparisons with the existing techniques of the state of the art.

Objective and Subjective Investigation on a Novel Method for Digital Reverberator Parameters Estimation

Reverberation is a well known effect that has an important role in our listening experience. A great deal of research has been devoted in the last decades aiming to artificially reproduce the reverberation effect exploiting a hybrid reverberation structure. In this context, several automatic procedures have been presented in the literature in order to derive the reverberator structure considering the mixing time evaluation and the minimization functions definition for the late reverberation device. Taking into consideration these aspects, a deep analysis of hybrid digital reverberator audio quality is here proposed, introducing a new parameter for the definition of the mixing time and two new cost functions for the definition of the late reverberation parameters. More in detail, starting from the considerations derived from a previous accurate approach based on the mel frequency cepstral coefficients, the new cost functions are based on the evaluation of the perceptual linear predictive and power normalized cepstral coefficients. Several results are reported, in terms of objective measure, performance analysis and subjective measures, taking into consideration different real impulse responses and various input stimuli and making a comparison with the state of the art. In particular, the obtained results show that a good accuracy can be achieved also considering a low number of coefficients, therefore improving the computational performance.

An efficient implementation of acoustic crosstalk cancellation for 3D audio rendering

The paper deals with the development of an efficient real time system for the reproduction of a spatialized audio field taking into account the listeners position. The system is composed of two parts: a sound rendering system based on a crosstalk canceller that is required in order to have a spatialized audio reproduction and a listener position tracking system in order to model the crosstalk canceller parameters. Then, an efficient implementation of a time domain crosstalk cancellation algorithm is presented considering an improved version of the recursive ambiophonics crosstalk elimination algorithm. A real time application is proposed introducing a Kinect control, capable to accurately track the listener position and changing the crosstalk parameters related to its position. Several results are presented comparing the proposed approach with the state of the art in order to confirm its validity.

Hybrid Reverberator Using Multiple Impulse Responses for Audio Rendering Improvement

In the recent years, hybrid reverberation algorithms have been widely explored aiming to reproduce the acoustic behavior of real environment at low computational load. On this basis, exploiting the advantages introduced from hybrid reverberation structures, a novel approach for the reproduction of moving listener position through impulse responses (IR) interpolation has been presented in this paper. In particular, the presented methodology allows to remove redundant information in large IR database also decreasing the memory usage and the computational complexity required to perform the auralization operation. The effectiveness of the proposed approach has been proved taking into account a real IR database and also providing comparison with the existing state-of-art techniques in terms of objective and subjective measures.

Advanced audio spatializer combined with a multipoint equalization system

The paper deals with the development of a real time system for the reproduction of an immersive audio field considering the crosstalk cancellation and the room response equalization issues. In particular, the real-time system is composed of two parts: a crosstalk cancellation network and a combined multipoint equalization structure. The former is required in order to have a spatialized audio and it is based on the free-field relationship in order to not introduce a timbre alteration. The latter is used to improve the objective and subjective quality of sound reproduction systems by compensating the room and loudspeakers transfer function. Both steps are based on a-priori analysis of the real environment using real impulse responses measured in different positions. In particular, an offline procedure capable of determining the tuning parameters for the crosstalk network and of deriving the final filters for the equalization structure, is adopted. Several results are presented in order to show the effectiveness of the proposed algorithms considering objective and subjective evaluations and comparing the presented approach with the state of the art.

An efficient time varying hybrid reverberator for room acoustic simulation

This paper deals with the development of an efficient real time system for the simulation of time varying sound fields produced by moving sources and receiver in virtual reverberant environments. The system is composed of two parts: the impulse responses generator based on image-source model and a time varying hybrid reverberator employed to reproduce the spatial impression through impulse responses interpolation. In detail, the proposed approach overcomes the computational limitations related to use of the image-source method removing redundant information in the generated impulse responses and thus, decreasing the computational complexity required to perform the audio rendering operation. The effectiveness of the proposed approach has been proved taking into account different virtual reverberant environments and providing comparison with the state-of-art techniques in terms of objective and subjective measures.

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.