Featuring contributions from Dave Moffat and Brecht De Man
As you might know, or can guess, we’re heavily involved in the Audio Engineering Society, which is the foremost professional organisation in this field. We had a big impact at two of their recent conferences.
The 60th AES conference on Dereverberation and Reverberation of Audio Music and Speech took place in Leuven, Belgium, 3-5 February 2016.. The conference was based around a European funded project of the same name (DREAMS – http://www.dreams-itn.eu/) and aimed to bring together all expertiese in reverb and reverb removal.
The conference started out with a fantastic overview of reverberation technology, and how it has progressed over the past 50 years, by Vesa Välimäki. The day then went on to present work on object based coding and reverberation, computation dereverberation techniques.
Day two started with Thomas Brand discussing sound spatialisation and how participants are much more tolerant of reverberation in binaural listening conditions. Further work then presented on physical modelling approaches to reverberation simulation, user perception, and spatialisation of audio in the binaural context.
Day three began with Emanuël Habets, presenting on the past 50 years of reverberation removal, discussing that since we started modelling reverberation, we have also been trying to remove it from audio signals. Work was then presented on multichannel dereverberation and computational sound field modelling techniques.
The Audio Engineering group from C4DM were there in strength, presenting two papers and a demo session. David Moffat presented work on the impact dereverberation can make when combined with state of the art source separation technologies. Emmanouil Theofanis Chourdakis presented a hybrid model which, based on machine learning technologies, can intelligently apply reverberation to an audio track. Brecht De Man presented his latest research, as part of the demo session and again in a plenary lecture, on analysis of studio mixing practices, focused on analysing the perception of reverberation in multitrack mixes.
The following week was the AES Audio for Games conference in London. This is the fifth game audio conference they’ve had, and we’ve been involved in this conference series since its inception in 2009. C4DM researchers Dave Moffat, Will Wilkinson and Christian Heinrichs all presented work related to sound synthesis and procedural audio, which is becoming a big focus of our efforts (more to come!).
Brecht De Man put together an excellent report of the conference, where you can find out a lot more.