Roderick van Domburg

I've pushed a fix for this to #786. The issue was that this specific M4A file - in ISO5 format - had set its frames metadata to 0, causing total...

Not directly related to scope of this PR, but as I was reviewing it I thought we can do better than sleeping for 5 ms in the iterator implementation. Here's...

You’re right that the burst-filling behavior mitigates some of the latency concerns, but I think your analysis assumes data is transferred to cpal exactly when the microphone thread wakes up....

> Part of me just wants to swap everything out for an mpsc::channel. But that probably would impact perf.... (Take a look at the impl of mpsc it is not...

Out of all the solutions, with a condvar you'll have lowest jitter, lowest system load, without the performance penalty of MPSC/SPSC synchronization 🤷‍♂️

I’ll add it before merge, probably this evening. To your point of readability, `ringbuf` will definitely be easier on the eyes, but I’ll remember to add some code comments what...

Here you go. This also adds the logic to always write full blocks (buffer periods) if we know the buffer size. By the way, I think it should be possible...

> This also adds the logic to always write full blocks (buffer periods) if we know the buffer size. Sleeping over this, I simplified the logic to just notify on...

Would you help me review that PR over at cpal? That community is pretty dormant. I’m trying to revive it, but for now, I don’t expect much help there. When...

I tend to agree that the power in this feature would be precisely to expose it to users. To that extent it'd be worthwhile to have an example added that...