> Why would they decode it only to re-encode it? Because it's more generic and easier to implement. I would imagine to support streaming AAC from the file to

> Why would they decode it only to re-encode it? Because it's more generic and easier to implement. I would imagine to support streaming AAC from the file to the AirPods, they'd need to create a special channel for it in the OS's sound system. Even with that, I don't think it'd work in all cases. Sometimes you have to manipulate the audio. What happens when you need to mix in audio from an app, like an alert sound? The AirPods can't keep using the file's unmodified AAC stream, so you'd have to temporarily stop streaming that to them, decode it, mix in the app audio, re-encode that, and ship the modified stream to the AirPods. All the while you have to keep the timing prefect throughout all the changes or it will be even worse-sounding due to the skips. It's a lot easier to just decode-mix-reencode all the time, since that's a pretty straightforward pipeline, instead of having to switch back and forth between different audio streaming methods, and deal with loads of extra complexity.