Several real-time voice stacks, head to head. Tap Connect on any one — it immediately reads the same short Signal aloud, then stays live so you can talk to it and ask follow-ups. Compare the latency and the feel, then pick the winner.
Tap Connect on any stack — it reads the report, then talk to it. Compare the latency and the feel. Use one at a time (mics can cross-talk). The latency shown is the gap from the moment you stop speaking to the first audio back (last turn + running average), plus time-to-first-audio on the opening read.