Perceptual Audio Evaluation

13 October, 2021 @ 16.00 - 17.30

What is good audio quality and how do we measure it? Join us for this webinar about perceptual audio and learn about the subjective evaluation techniques in use now, and those in development for the future.

We have a star panel of speakers from Sense Lab, Google and GN Hearing.

Photo: FORCE Technology

Form: Webinar

Date: 13 October 2021

Time: 16.00 – 17.30

Price: free

Language: English

Place: Online via Zoom

Participation link will be sent after you sign up.

 

Program:

  • Nick Zacharov will give a broad overview and introduction to the subject of perceptual audio evaluation, which just happens to be the name of his renowned book. He will discuss some of the techniques that they use at SenseLab, where he is the lead technologist. 

  • Jan Skoglund will present his work on low bit-rate speech compression using AI, and the evaluation methods that are required for this kind of technology. The training of networks using existing measures is inadequate for evaluation of audio quality.

  • Charlotte Sørensen will speak about objective metrics for speech quality and intelligibility  prediction – what they are, the current state-of-the-art methods, the advantages and limitations. She will also cover non-intrusive metrics – those which can predict the perceptual performance of a noisy audio signal without requiring a clean speech reference signal (making it useful for real-time processing).

Participant profile:
  • Those working in product development
  • Acoustic and DSP engineers
  • Those working with audio quality and evaluation methods

You will meet:

In case of any questions regarding sign-up, please contact Murielle De Smedt, mds@danishsound.org

Nick is an acoustic/electroacoustic and perceptual evaluation expert with 30 years of engineering and leadership experience and is the co-author of “Perceptual Audio Evaluation”, the bible of evaluation. He specializes in sound quality, sensory evaluation, multi-sensory perception, modelling, machine learning, innovative product development, and we hear that he is an excellent team leader. He is an active member of the audio community and has over 90 publication and patents.

Jan Skoglund is a software engineer, manager, researcher and musician. He has been working at Google for the past 10 years, and before that as the R&D Manager at Global IP Solutions.

Charlotte Sørensen received her B.Sc. and M.Sc. degree in Biomedical Engineering and Informatics, and the Ph.D. degree from Aalborg University, Aalborg, Denmark, in 2012, 2014 and 2020, respectively. While working toward the Ph.D. degree she was employed with GN Hearing A/S, Ballerup, Denmark and associated with the Audio Analysis Lab, Aalborg University. The title of her thesis was “Non-Intrusive Speech Intelligibility Prediction”. She is currently employed as a Research Scientist at GN Hearing A/S, where her main research interests include speech signal processing with emphasis on objective metrics for perceptual evaluation of speech and speech enhancement for hearing aid applications.

Innovationskraft
When you participate in this event, your time will be used as co-financing for the project Innovationskraft, which is funded by Danmarks Erhvervsfremmebestyrelse and Uddannelses- og Forskningsstyrelsen at a standard rate. Read more about Innovationskraft 

Danish Sound Cluster

Hold dig opdateret

Tilmeld vores nyhedsbrev: