Abstract
Although resynthesis may seem a simple analysis/synthesis process, it is a quite complex task, even more when it comes to recreating a singing voice. This paper presents a system whose goal is to start with an original audio stream of someone singing and recreate the same performance (melody, phonetics, dynam-ics) using an internal vocal sound library (choir or solo voice). By extracting dynamics and pitch information, and looking for phonetic similarities between the original audio frames and the frames of the sound library, a completely new audio stream is created. The obtained audio results, although not perfect (mainly due to the existence of audio artifacts), show that this technologi-cal approach may become an extremely powerful audio tool.

Abstract
The quality of low bit-rate coding algorithms standardized under ISO/MPEG has been carefully verified by subjective listening tests using audio material of the highest technical quality. When using digital audio compression in the context of television sound systems, however, the question arises, as to how the coders would perform with imperfect audio signals, such as those obtained from impaired film material. This paper describes studies into the effects of the more severe impairments present in film soundtracks upon MPEG-I Layer-2 and Layer-3 audio coding, which have been carried out within the European RACE `COUGAR' project programme R2122.

Abstract
This paper presents new results improving by a factor of 10 the accuracy of an Odd-DFT based frequency estimation algorithm. These results are shown to be robust to the influence of additive noise and compare favorably to other non-iterative frequency domain estimation algorithms. A perspective is given oil possible application areas, namely those involving real-time constraints.

Abstract
Consumer audio applications such as satellite radio broadcasts, multi-channel audio streaming and playback systems coupled with the need to meet stringent bandwidth requirements are eliciting newer challenges in parametric multichannel audio coding schemes. This paper describes the continuation of our research concerning the Immersive Soundfield Rendition (ISR) system and the different enhancements in various algorithmic components. The need to maintain a constant bit rate for many applications requires a rate control mechanism. The various strategies utilized in the rate control mechanism are presented. In addition, an innovative phase compensated down-mixing scheme has been incorporated in the ISR system so as to generate a high quality carrier signal. Enhancements have been made to the blind up-mixing scheme and to considerable gains have been made in terms of acoustic diversity. The various enhancements of the ISR system and its performance are detailed. Audio demonstrations are available at http://www.atc-labs.com/isr.

Abstract
We recently introduced a novel Automatic Noise Reduction (ANR) algorithm for the removal of wideband stationary/non-stationary noise from audio [1]. Current noise reduction techniques exhibit certain undesirable characteristics. Distortion and/or alteration of the audio characteristics is a common problem. User intervention in identifying the noise profile is sometimes necessary. ANR uses a novel framework employing dominant component subtraction and restoration and performs better than conventional techniques in subjective tests. Here we describe three enhancements to ANR. The first of these increases the level of noise removal for the special case of stationary background noise. The second is a new tool for improving the temporal envelope coherence and yields additional noise removal. The third is a multi-band processing tool for conditioning time-frequency envelope for reduced listener fatigue.

Abstract
In this paper we present a new approach to the problem of isolated vowel recognition in real-time. Language learning and speech therapy are examples of application areas that require real-time biofeedback of acoustic features. As the performance of known approaches usually drops for child speakers, we evaluated different alternatives of feature extraction and pattern recognition techniques, including PCA, LDA, ANN and Bayesian classification. In addition, we studied the explicit inclusion of pitch as a main parameter in both simulation and the real-time feature extraction process. Best results were obtained with our dataset when MFCCs are mapped, using LDA, to a 4-dimensional sub-space that is followed by Bayesian classification. An interactive game was designed that implements the selected real-time vowel recognition technique.