Last Monday June 29, 2020 —exactly 1200 our time in Argentina, 1600 in UK- was an historical day for us, because we had the privilege to have a videoconference with Francis Rumsey: we had a memorable experience listening to advice and concepts directly from the man himself. Here is a general topic-resumed meeting content:
Concerning the 3D audio perception in an interactive audiovisual program: trained users hear spatial location; untrained, general users listen to other things, like timbral quality. So what we ask and why? Why we are obsessed with things like localization when users in general don't focus on that in the first place? In user interaction in virtual or augmented reality, the need to make complicated tasks doesn't allow to focus on this kind of approach of sound quality (a space localization approach). The question that matters most probably is: does the sound help to successfully complete the task the user made?
Concerning the 3D audio perception in music only, a classic passive user program: the jump from mono to stereo is pronounced. The jump from stereo to surround is smaller for music, adding more channels adds more smaller differences, talking specifically with speakers. In binaural, things are apparently easer in the first place (only two channels), but without head tracking most binaural reproduction tends to be in the back only, it's difficult to perceive frontal sources without it. Ahead of that, the type of binaural render used for deliver a good, universal 3D experience is the critical part, and there is no commercial evidence of a real interest of music consumers (mainstream, entertainment music) ahead of two channel stereo, since all the efforts since quadrophonic systems till today. Still, the binaural possibilities are great in these times, but perhaps the key is to deliver a switchable, intermediate spatial format in the end of the signal chain (and Ambisonics is a good example for that), to switch between headphones and speakers in a competent way.
Concerning the auditory perception of 3D audio of trained and untrained listeners: parallelism with the food industry=trained subjects do better the detail description of the technical qualities that allow the creation of a flavor profile: perhaps sound experts as subjects can do the same with sound. But if we ask to this people what they like, it's a very different question: that question it's made to ordinary people (the consumers). If you ask to common audio consumers what it's the localization of a source, of a spectral clue, etc. the responses will not be as good as with experts.
Thank you very much Francis for sharing your valuable time with us, talking with wisdom and profound knowledge, with great predisposition and kindness.
