On a 90-seat IP telephony rollout, we’re seeing occasional robot audio on Wi‑Fi 6 handsets; currently running Opus 16 kHz at 20 ms ptime, adaptive jitter buffer 60–120 ms, DSCP EF (46) with LLQ, targeting MOS 4.2+. Has anyone gotten cleaner handoffs by switching to G.711 with a fixed about 80 ms buffer or bumping ptime to 30 ms without tanking talker latency?
I wouldn’t jump to G.711 on Wi‑Fi 6 — the airtime hit usually makes roams worse. > handoffs by switching to G.711 with a fixed about 80 ms buffer or bumping ptime to Stay on Opus 16 kHz/20 ms, enable in‑band FEC, and cap the jitter buffer around 80–90 ms; that stopped the robot audio for us at similar scale. Are 802.11r/k/v and U‑APSD enabled on the voice SSID?
Quick check: are you negotiating 802.11r FT over‑the‑air on the voice SSID? Keeping Opus 20 ms and enabling FT OTA has consistently shaved enough roam time that the 60–120 ms adaptive buffer stops producing ‘robot audio’ for us. Reference if helpful: https://www.arubanetworks.com/techdocs/ArubaOS_87_Web_Help/Content/arubaos-solutions/FT/FT.htm.