Author: Jan Abildgaard Pedersen
Authors:Chigwamba, Nyasha; Foss, Richard; Gurdan, Robby; Klinkradt, Bradley
Affiliation:Rhodes University, Grahamstown, South Africa; Universal Media Access Networks GmbH, Düsseldorf, Germany
When related audio control parameters are distributed in many devices in a high-speed audio network, a means for centralizing remote control and monitoring is required. Changes in one parameter may need to affect other parameters. The XFN protocol enables management of distributed parameters. This paper focuses on how to group related parameters that are physically separated and provides a corresponding implementation using the protocol. A comprehensive grouping theory is presented.
Download: PDF (HIGH Res) (3.4MB)
Download: PDF (LOW Res) (473KB)
Authors:Raake, Alexander; Wältermann, Marcel; Wüstenhagen, Ulf; Feiten, Bernhard
Affiliation:Telekom Innovation Laboratories (T-Labs), TU Berlin, Germany; Telekom Innovation Laboratories (T-Labs), Deutsche Telekom, Germany
Typically, audio quality assessment uses MUSHRA (Multi Stimulus with Hidden Reference and Anchors), while speech quality assessment uses ACR (Absolute Category Rating). Since many applications are transporting both speech and music, such as mobile devices, a conversion technique between the two types of ratings would be useful. Two speech and two audio quality listening tests are compared with different content types. The results illustrate when and how the two types of measurements are consistent, complementary, and inconsistent.
Download: PDF (HIGH Res) (805KB)
Download: PDF (LOW Res) (308KB)
Authors:Côté, Nicolas; Koehl, Vincent; Möller, Sebastian; Raake, Alexander; Wältermann, Marcel; Gautier-Turbin, Valérie
Affiliation:Institute of Electronics, Microelectronics and Nanotechnology, ISEN department, UMR CNRS, Lille, France; European Centre for Virtual Reality (LISyC EA 3883), University of Brest (UEB), France; Telekom Innovation Laboratories, TU Berlin, Germany; France Télécom R & D, Lannion, France
A new speech quality model, DIAL (Diagnostic Instrumental Assessment of Listening), provides diagnostic information in both narrow-band and super wide-band contexts. It is “intrusive,” assuming that the original source audio is available. Because many quality-measuring techniques collapse all degradations into a single score, they do not help developers to diagnose the basis of that score. In contrast, the proposed DIAL model uses four quality dimensions: coloration, continuity, noisiness, and loudness.
Download: PDF (HIGH Res) (769KB)
Download: PDF (LOW Res) (244KB)
Authors:Möller, Sebastian; Kettler, Frank; Gierlich, Hans-Wilhelm; Poschen, Silvia; Côté, Nicolas; Raake, Alexander; Wältermann, Marcel
Affiliation:Quality and Usability Lab, Telekom Innovation Laboratories, Technische Universität, Berlin, Germany; HEAD acoustics GmbH, Herzogenrath, Germany; Institut Supérieur de l’Électronique et du Numérique (ISEN), Lille, France; Assessment of IP-based Applications Lab, Telekom Innovation Laboratories, Technische Universität, Berlin, Germany
The E-model for predicting speech quality mouth-to-ear can be extended with additional parameters to describe the effect of imperfect noise reduction and echo cancellation. As shown by subjective tests, quality-prediction accuracy of noise reduction and echo cancelling improves. Future work is planned to better refine the proposed approach.
Download: PDF (HIGH Res) (898KB)
Download: PDF (LOW Res) (270KB)
Increasingly sophisticated analysis, modeling, and simulation tools enable sound fields to be taken apart and put back together again in a modified form. They enable detailed control over the local acoustical environment, including aspects of its spatial characteristics and perceived qualities. This has a range of possible applications, such as the enhancement of natural sounds, spatial manipulation, personalized sound zones, and systems that can adapt to their environments.
Download: PDF (423KB)