AES New York 2011
Paper Session P20
P20 - Audio Processing—Part 2
Saturday, October 22, 2:30 pm — 4:30 pm (Room: 1E09)
Chair:
James (JJ) Johnston
P20-1 Digital Low-Pass Filter Design with Analog-Matched Magnitude Response—Michael Massberg, Brainworx Music & Media GmbH - Leichlingen, Germany
Using the bilinear transform to derive digital low-pass filters from their analog prototypes introduces severe warping of the response once the cutoff frequency approaches the Nyquist limit. We show how to design pre-warped first and second order low-pass prototypes that, after application of the bilinear transform, provide a better match with the magnitude response of the analog original than applying the bilinear transform directly. Result plots are given and compared for different cutoff frequencies and Q factors.
Convention Paper 8551 (Purchase now)
P20-2 Performance Evaluation of Algorithms for Arbitrary Sample Rate Conversion—Andreas Franck, Fraunhofer Institute for Digital Media Technology IDMT - Ilmenau, Germany
Arbitrary sample rate conversion (ASRC) enables changes of the sampling frequency by flexible, time-varying ratios. It can be utilized advantageously in many applications of audio signal processing. Consequently, numerous algorithms for ASRC have been proposed. However, it is often difficult to choose a minimal-cost algorithm that meets the requirements of a specific application. In this paper several approaches to ASRC are reviewed. Special emphasis is placed on algorithms that enable optimal designs, which minimize the resampling error with respect to a selectable norm. Evaluations are performed to assess the computational efficiency of different algorithms as a function of the achievable quality. These analyses demonstrate that structures based on oversampling and resampling filters specifically adapted to this structure yield significant performance improvements over established algorithms.
Convention Paper 8552 (Purchase now)
P20-3 Acoustic Echo Cancellation for Surround Sound Using Spatial Decorrelation—Namgook Cho, Jaeyoun Cho, Jaewon Lee, Yongje Kim, Samsung Electronics Co., Ltd. - Korea
One of the main challenges for a stereophonic acoustic echo canceller is that it suffers from poor convergence, which is caused by strong correlation between input signals. We have proposed a new decorrelation technique that adopts spatial decorrelation to address the problem without altering the input signals. Here, we extend the results to a more generic setting, i.e., a 5.1-channel surround system. In the scheme, the input signals are decomposed and projected into the signal subspace and the noise subspace. When the decorrelated signals are fed to the adaptive filters, the interchannel coherence between the input signals decreases significantly, which provides performance improvement in echo reduction. Experiments in a real-world environment and performance comparison with state-of-the-art techniques are conducted to demonstrate the effectiveness of the proposed technique.
Convention Paper 8553 (Purchase now)
P20-4 A Fixed Beamforming Based Approach for Stereophonic Audio-Conference Systems—Matteo Pirro, Stefano Squartini, Francesco Piazza, Università Politecnia delle Marche - Ancona, Italy
Hands-free communications systems require to primarily reduce the impact of the inevitably occurring acoustic echo. Moreover, in the recent past, a certain attention has been devoted to algorithmic frameworks able to provide stereophonic acoustic rendering and so augmenting the pleasantness of the audio-conference experience. In this paper the authors propose an optimally designed fixed Beamformer (BF) based solution for Stereophonic Acoustic Echo Cancellation (SAEC), with a twofold objective in mind: reducing the echo power and maximizing the stereophonic spatial feeling. Up to the author’s knowledge this is new to the literature, and the achieved experimental results seem to confirm the effectiveness of the approach. It must be underlined, on purpose, that preliminary subjective listening tests have been carried out to evaluate the attainable audio stereo-recording quality. Moreover, the proposed solution allows to significantly reduce the overall computational cost required by the SAEC framework: indeed, BF implementation asks for few extra filtering operations with respect to the baseline approach from one hand but make the usage of the decorrelation module unessential.
Convention Paper 8554 (Purchase now)
Information Last Updated: 20111005, mei
