In order to improve conference quality, one could use interaural level & time differences (basics of psychoacoustic sound localization commonly written as ILD and ITD, https://en.wikipedia.org/wiki/Sound_localization#ITD_and_IID ).
Basically, a virtual meeting room could be setup with all attendees located throughout a circle. As Rainbow can identify the speaker, the sound could be virtually located where the speaker is by using very simple audio filters. More advance processing could use HRTF (head related transfer function) and give a better result.
Several limitations to that:
- if several people are using one sending device (e.g. one "simple" phone in a conference room), the corresponding virtual attendee would represent all participants who are in this room.
- sound localization would not be possible without a stereo headset but it will not decrease conversational quality anyway as the delay between our ears is lower than a millisecond.
Please sign in to leave a comment.