Compare segment encoding vs encoding the whole recording
In the current implementation, encodes PCM to MP3 20ms segment at a time then smashes it all together. I was dubious too! But this worked. The reason for doing this is that PCM is much much bigger than MP3 so storing this while I record could lead to massive queue files.
I may be leaving some performance and quality on the table, so this issue is to capture that I must compare both methods, store all PCM bits then encode vs encode PCM bits as they come.