Perceptual analysis and evaluation of Quality of Experience (QoE) for speech services

Motivation & Project Description

• Perceptual analysis and evaluation of the Quality of Experience (QoE) for important speech services:
  • Speech Transmission: Super-wideband speech quality dimensions
  • Speaker Characterization: User specific perception (age, gender, personality)
  • Speech Generation Factors: Text-to-speech (TTS) quality features
  • Interaction Factors: Multi-party conferencing quality aspects
• Focus: Evaluating dimensions of QoE and their interaction

Expected Outcome:

• Super-wideband E-model for network planning and monitoring
• Improvement of DIAL (former ITU-T Rec. P.863 candidate by France Télécom R&D, Kiel University, and T-Labs)
• ITU-T contributions and recommendations (ITU-T SG 12, Q.8/12, Q.9/12, and Q.18/12)
• Automatic user characterization similar to (manual) professional assessment centers
• Recommendations for TTS synthesis and multiparty conferencing system development
• Scientific publications

Time Frame: 01/2011-12/2012

T-labs Team Members: Florian Hinterleitner, Blazej Lewcio, Sebastian Möller, Tim Polzehl,
Alexander Raake (AIPA), Janto Skowronek (AIPA), Marcel Wältermann

Students: Henrik von Coler, Erik Detzner, Claudia Pischel

Partners: Fachgebiet Assessment of IP-based Applications (AIPA)

Funding by: Deutsche Telekom Laboratories