ASR, ACD, PDD, NER and MOS: VoIP Quality Metrics Explained

02 Aug, 2026
Understand ASR, Average Call Duration, PDD, NER and MOS, with a practical method for comparing outbound voice traffic and investigating real call examples.

A route report becomes useful when it helps explain a real call. ASR, ACD, PDD, NER and MOS describe different parts of the calling experience. Read them together, keep their definitions consistent and compare traffic that is genuinely comparable.

ASR: how many attempts were answered?

Answer-Seizure Ratio expresses answered seizures as a percentage of total seizures. Check how your reporting system counts attempts and retries before comparing exports. In an illustrative sample of 100 counted attempts and 42 answers, the ratio is 42%.

Subscriber behavior affects the result: a busy or unanswered call can reduce ASR without proving a route fault. Compare destination networks, calling times and traffic profiles before deciding what changed. Keep the sample size beside the percentage.

ACD: how long did answered calls last?

ACD means Average Call Duration in this context. Divide the total duration of answered calls by the number of answered calls, using the same duration field throughout the report. If there are no answered calls, the average is undefined.

A change in ACD is a prompt to inspect calls. Short conversations may reflect the purpose of the call, the recipient's response or a technical interruption. Compare call outcomes with actual examples instead of assigning one explanation to every short call.

PDD: where does the waiting begin and end?

Post-Dial Delay measures a call-setup interval. Define its start and end events in your report. The first progress response, audible ringback and the answer are different events; a system measuring one should not be compared directly with a system measuring another.

For a slow-call example, keep a timeline: when the device sent the call, when progress arrived and when the user heard audio. SIP distinguishes progress responses such as 180 Ringing and 183 Session Progress from a successful final response. A trace and the user's experience should be examined together.

NER: separate network outcomes from subscriber behavior

Network Effectiveness Ratio is intended to assess how effectively a network handles attempts, including outcomes such as busy or no answer when classified appropriately. Ask how the report maps termination causes into this metric. An unexplained list of SIP codes is not enough to guarantee that two providers calculate NER identically.

Use the breakdown behind the ratio. A change concentrated in one destination or response category gives you a more specific investigation than an account-wide average.

MOS: understand the measurement method

Mean Opinion Score describes perceived quality. A listener-test result and an algorithmic estimate are different kinds of evidence. Record the method used, the codec and the measurement point when comparing values. Treat a single score as part of the evidence alongside two-way listening tests and media statistics.

Read media measurements alongside call outcomes

Packet loss, variation in packet arrival timing and delay help explain audio problems. Keep the two directions of a call separate where your tools allow it. Also check the endpoint and local connection when several routes show the same symptom from one workstation.

  • Use a consistent reporting window and timezone.
  • Separate destinations and comparable traffic profiles.
  • Retain raw counts, not only averages and percentages.
  • Pair unusual metrics with a small set of call examples.
  • Change one relevant variable and repeat the comparison.

Apply the review to SIPEASY outbound calling

SIPEASY's focus is outgoing traffic through CLI, nCLI and TDM route offerings. Registration and outbound route access are KYC-free. Share the destination, timestamp, call identifier and observed issue through the contact page when discussing a test or investigating a result.

For teams using PBX tools, the paid packages include recording, SPY and Whisper with device and storage allowances defined by the chosen package. Use your actual workflow to decide which tools are relevant to the review.

Try Standard for one month free. Sign in to your SIPEASY panel and activate the offer yourself using its dedicated button. Standard normally costs $100/month. Outgoing calls remain chargeable at the applicable tariff. See current packages and call rates.

Technical references

ITU-T E.425 describes network performance monitoring, including ASR and NER. RFC 3261 defines SIP signaling, and RFC 3550 describes RTP and RTCP reporting.

Tags :