Work out jitter budget instantly with clear inputs, formula shown and shareable results.
A jitter buffer converts variable arrival timing into a steady playout stream by delaying every packet, so it trades latency for smoothness. Sizing it at roughly 2.5 times the measured jitter covers the great majority of arrivals; anything later is discarded. That is why ITU-T G.114 caps one-way voice delay at 150 ms — the buffer consumes part of that budget.
Jitter buffer
required depth ~ 2.576 x measured jitter for 99% coverage; buffer packets = depth / packet interval; added latency = buffer depth
Under 30 ms is generally fine, and under 10 ms is good. Above about 50 ms, no practical buffer keeps latency within the one-way delay budget.
It adds directly to mouth-to-ear delay. Beyond about 150 ms one way, conversation becomes awkward and participants start talking over each other.