MULTICHANNEL SOUND RENDERING VIA VIRTUALIZATION IN A STEREO LOUDSPEAKER SYSTEM

Patent №

US 8,335,331

Granted

2012-12-18

Filed 2008

Owner

MICROSOFT CORPORATION

AI components

1

speech

Assignment

Recorded

Dataset

AIPD

2023_r1 edition

Application

12016944

A speaker virtualization system provides virtual surround sound using a pair of physical loudspeakers. A multiple surround audio channels input is processed using a combination of head related transfer functions and shaped reverberation to provide widening and front/back auditory clues without requiring any kind of interaural path cancellation. The system uses a 360 degree power-response head related transfer function to provide perceptual separation of the reverberant and direct paths, along with discrete, different reverberation for left and right rendering channels to provide envelopment. By eliminating interaural path cancellation, the speaker virtualization system also produces a wider virtual surround sound effect, without dependency on head position and facing.

SpeechH04S 3/002H04S 2400/01H04S 2420/01

AI classification

Speech0.61
AI hardware0.00
Machine learning0.00
Natural language0.00
Vision0.00
Planning0.00
Evolutionary computation0.00
Knowledge representation0.00

Ownership

MICROSOFT CORPORATION

assignment · 203950044

Assignors

JOHNSTON, JAMES D., LI, QUNLI, SMIRNOV, SERGE

On an employer assignment, the assignors are typically the inventors.

© 2026 NYSGPT2525 LLC