* encode/store audio as part of video file
* better match write_audio() with write()
* handle different FFmpeg versions, flush audio encoder, suppress encoder QAvg/info messages
* use audio_buffer.size() instead of keeping track of size separately
* no more for loops
* save to qcam and rlog
* assert audio support check
* microphone --> soundPressure, audioData --> rawAudioData
* deque much more efficient if buffer ever >> frame_size, ~ same performance for defaults
* cleanup and fix time scaling
* initialize audio separately and pass sample_rate in
* update comments
* ensure header is written before writing audio
* buffer audio frame but do not process before header written
* handle buffer overflow now that we are using as an actual buffer
* spelling
Dean Lee