How to check VoIP route quality: ASR, ACD, PDD, MOS and packet loss

Two routes to the same destination can have similar prices and very different results. Checking route quality is how wholesale carriers decide where to send traffic, when to raise a ticket with a vendor and when to move volume elsewhere. This guide explains the core metrics — ASR, ACD, PDD, MOS and packet loss — what each one tells you, and how to combine them into a reliable view of a route.

ASR – Answer-Seizure Ratio

What it is: the share of call attempts that are answered, usually shown as a percentage:

ASR = answered calls ÷ total call attempts × 100

What it tells you: how often calls on the route actually connect. A sudden drop often points to a vendor problem, capacity limits or a routing change.

Watch out: ASR depends heavily on the destination and on your customers’ traffic profile (for example, the share of mobile numbers, invalid numbers or busy subscribers). Compare a route with other routes to the same destination carrying similar traffic. Also remember that an unusually high ASR is not always good news — it can be a sign of False Answer Supervision.

ACD – Average Call Duration

What it is: the average length of answered calls:

ACD = total minutes of answered calls ÷ number of answered calls

What it tells you: whether people are having real conversations. Poor audio, one-way audio or dropped calls usually shorten ACD, because callers hang up early.

Watch out: like ASR, ACD varies by destination and traffic type. Call-centre traffic naturally has shorter calls than person-to-person calling. Look at the distribution of durations too: many calls ending after a similar short time can indicate a problem that the average hides.

PDD – Post-Dial Delay

What it is: the time between sending the call (the SIP INVITE) and the caller receiving an indication that the call is progressing, such as ringing (180 Ringing or 183 Session Progress).

What it tells you: how long the caller waits in silence. Long PDD often means the call is passing through many hops or that a vendor is retrying through several paths. Callers may hang up before the call even rings, which lowers ASR.

Watch out: a PDD that is suspiciously short combined with an immediate answer can be a FAS indicator rather than a sign of excellent quality.

Packet loss

What it is: the percentage of RTP voice packets that never arrive at the far end.

What it tells you: how much of the audio is missing. Even small amounts of loss can be audible as clipping or robotic speech, especially when losses come in bursts.

Watch out: check whether loss is constant or appears at particular times of day; peak-hour congestion on a vendor’s network is a common cause.

Jitter and latency

Two further network measurements help explain audio problems. Latency is how long packets take to travel; high latency makes conversations feel awkward, with people talking over each other. Jitter is the variation in packet arrival times; too much jitter overwhelms the receiving jitter buffer and behaves like packet loss.

MOS – Mean Opinion Score

What it is: a score from 1 (bad) to 5 (excellent) describing perceived voice quality. Originally it came from human listening tests; in networks it is usually estimated from measurements such as packet loss, jitter, latency and the codec in use, for example using the ITU-T E-model.

What it tells you: a single, easy-to-compare figure for how calls on a route are likely to sound.

Watch out: the maximum achievable MOS depends on the codec — a compressed codec such as G.729 scores lower than G.711 even on a perfect network. Compare MOS between routes using the same codec, and treat it as an estimate rather than a guarantee.

Putting the metrics together

No single number describes a route. A practical approach is to read them together:

Passive monitoring vs test calls

Passive monitoring measures your real traffic as it flows. It reflects exactly what your customers experience and costs nothing extra, so it should be your main source of truth for routes that carry traffic.

Test calls are useful for routes that carry no traffic yet — for example, a new vendor or a backup route. Keep them operator-controlled and limited, and use them to confirm specific things such as caller ID delivery, ringback behaviour and audio quality.

A simple route-quality routine

1. Watch live KPIs (ASR, ACD, PDD, packet loss, MOS) per route, refreshed frequently.
2. Compare routes per destination rather than against a global target.
3. Give each route a clear status such as PASS, WATCH or FAIL so the NOC knows where to focus.
4. Investigate with evidence – SIP traces and recordings for the calls behind a bad number.
5. Test new routes before they receive volume, and flag them clearly until they have a track record.
6. Review trends in hourly and daily reports, and adjust routing based on measured performance and cost.

How VoxInspector helps

VoxInspector brings these checks together for wholesale carriers on VOS3000: passive live monitoring of ASR, ACD, PDD, packet loss and MOS, automatic PASS / WATCH / FAIL live tests for routes carrying traffic, on-demand manual test calls, FAS detection, and Best Route suggestions per country. For routes and interconnects, see VoxComms wholesale voice services, or request a VoxInspector demo.