[Non-BE] Resolve "ISM to BINAURAL_ROOM_IR not BE between 20ms and 5ms rendering"
Reason why this change is needed
In case of head rotation with BINAURAL_ROOM_IR for discrete ISM the rendering is done as ISM->rotation of azi/ele according to the head rotation->rendering to virtual LS->CREND with BRIR. This is at the moment different for 5ms and 20ms rendering, since the rendering to virtual LS has a different rotation/rendering interpolation dependend on the rendering frame size asked. Furthermore it is not inline with the TD object renderer where the meta data update from the previous to the current azi/ele is dependent on a certain subframe index, while here always the current azi/ele is used.
Description of the change
Rewrite the render to LS function to really work on 5ms subframes with correct interpolators and rotation for 5ms output with head rotation. Add meta data update as in the TD object renderer.
Affected operating points
- ISM to BINAURAL_ROOM_IR with head/combinded orientation
- Produced bitstream is BE compared to previous state (no encoder change).
Related to #968 (closed)