Correct audio-metadata sync for td binaural renderer in OMASA

In ISM td binaural rendering, the metadata is updated in the third subframe to account for the audio delay, in order to have the correct audio-metadata sync. For OMASA, it is currently updated already in the first subframe.