Inertially stable virtual auditory space for spatial audio applications
During an initialization of a head pose tracker for a spatial audio system, a spatial audio ambience bed is rotated about a boresight vector to align the boresight vector with a center channel of the ambience bed. The boresight is computed using source device motion data and headset motion data. The ambience bed includes the center channel and one or more other channels. An ambience bed reference frame is aligned with a horizontal plane of a headset reference frame, such that the ambience bed is horizontally level with a user's ears. A first estimated gravity direction is fixed (made constant) in the ambience bed reference frame. During head pose tracking, the ambience bed reference frame is rolled about the boresight vector to align a second estimated gravity direction in the headset reference frame with the first estimated gravity direction fixed in the ambience bed reference frame.
1. A method comprising:
during an initialization of a head pose tracker for a spatial audio system that includes a source device and a headset coupled to the source device:
rotating, using one or more processors, a spatial audio ambience bed in a three-dimensional virtual auditory space about a boresight vector to align the boresight vector with a center channel of the ambience bed, and to align an ambience bed reference frame with a horizontal plane of a headset reference frame, such that the ambience bed is horizontally level with a user's ears, and fixing a first estimated gravity direction in the ambience bed reference frame; and
during head pose tracking:
rolling, using the one or more processors, the ambience bed reference frame about the boresight vector to align a second estimated gravity direction in the headset reference frame with the first estimated gravity direction fixed in the ambience bed reference frame, wherein the boresight vector is estimated using source device motion data and headset motion data, and the ambience bed includes the center channel and one or more other channels.
2. The method of claim 1 , wherein the center channel and the one or more other channels are located in a common plane of the ambience bed.
3. The method of claim 1 , wherein the ambience bed is configured for a 5.1 audio format.
4. The method of claim 1 , wherein the source device is a mobile device that presents visual content synchronized with spatial audio played through the center channel and the one or more other channels of the ambience bed in the three-dimensional virtual auditory space.
5. The method of claim 1 , wherein the estimated boresight vector determines an orientation of the ambience bed, and the ambience bed determines locations of audio channels around the user, such that when the center channel of the ambience bed is aligned with the boresight vector the center channel is rendered at an estimated location of source device.
6. The method of claim 1 , wherein the first or second estimated gravity directions are computed by:
determining a gravity direction using acceleration measurements output by an accelerometer of the source device, wherein the gravity direction is determined during a stationary or quiescence time interval when the source device is not moving;
computing a specific force vector based on an average of the acceleration measurements;
determining a reference gravity direction based on the specific force vector;
predicting an attitude of the source device based on a rotation rate of the source device and the reference gravity direction, wherein the rotation rate is output by an angular rate sensor of the source device; and
estimating the first or second gravity directions by rotating the determined gravity direction into an inertial reference frame using the predicted attitude of the source device.
7. A system comprising:
one or more processors;
memory storing instructions that when executed by the one or more processors, cause the one or more processors to perform operations comprising:
during an initialization of a head pose tracker for a spatial audio system that includes a source device and a headset coupled to the source device:
rotating a spatial audio ambience bed in a three-dimensional virtual auditory space about a boresight vector to align the boresight vector with a center channel of the ambience bed, and to align an ambience bed reference frame with a horizontal plane of a headset reference frame, such that the ambience bed is horizontally level with a user's ears, and fixing a first estimated gravity direction in the ambience bed reference frame; and
during head pose tracking:
rolling the ambience bed reference frame about the boresight vector to align a second estimated gravity direction in the headset reference frame with the first estimated gravity direction fixed in the ambience bed reference frame, wherein the boresight vector is estimated using source device motion data and headset motion data, and the ambience bed includes the center channel and one or more other channels.
8. The system of claim 7 , wherein the center channel and the one or more other channels are located in a common plane of the ambience bed.
9. The system of claim 7 , wherein the ambience bed is configured for a 5.1 audio format.
10. The system of claim 7 , wherein the source device is a mobile device that presents visual content synchronized with spatial audio played through the center channel and the one or more other channels of the ambience bed in the three-dimensional virtual auditory space.
11. The system of claim 7 , wherein the estimated boresight vector determines an orientation of the ambience bed, and the ambience bed determines locations of audio channels around the user, such that when the center channel of the ambience bed is aligned with the boresight vector the center channel is rendered at an estimated location of source device.
12. The system of claim 7 , wherein the first or second estimated gravity directions are computed by:
determining a gravity direction using acceleration measurements output by an accelerometer of the source device, wherein the gravity direction is determined during a stationary or quiescence time interval when the source device is not moving;
computing a specific force vector based on an average of the acceleration measurements;
determining a reference gravity direction based on the specific force vector;
predicting an attitude of the source device based on a rotation rate of the source device and the reference gravity direction, wherein the rotation rate is output by an angular rate sensor of the source device; and
estimating the first or second gravity directions by rotating the determined gravity direction into an inertial reference frame using the predicted attitude of the source device.