MOS Calculator: VoIP Call Quality Estimator
Enter your network latency, jitter and packet loss to estimate the Mean Opinion Score (MOS) and R-factor of your VoIP calls, and see what's dragging quality down.
How to improve it
MOS scale reference
| MOS | R-factor | Quality |
|---|---|---|
| 4.3 – 4.5 | 90 – 100 | Excellent |
| 4.0 – 4.3 | 80 – 90 | Good |
| 3.6 – 4.0 | 70 – 80 | Fair |
| 3.1 – 3.6 | 60 – 70 | Poor |
| 1.0 – 3.1 | 0 – 60 | Bad |
How the estimate is calculated
This tool uses a simplified version of the ITU-T G.107 E-model. It begins with a base R-factor of 93.2 and subtracts:
- Delay impairment (Id), from mouth-to-ear delay: one-way network latency + a jitter buffer of about twice the jitter + packetisation and codec delay. Id = 0.024·d, plus 0.11·(d − 177.3) when d exceeds 177.3 ms.
- Effective equipment impairment (Ie-eff) = Ie + (95 − Ie) · Ppl ÷ (Ppl + Bpl), using codec values from ITU-T G.113 Appendix I and assuming random packet loss.
R is converted to MOS with MOS = 1 + 0.035R + 7×10⁻⁶·R(R − 60)(100 − R). Real-world scores depend on loss burstiness, jitter buffer behaviour and handsets, so treat this as a planning estimate. Wideband codecs (G.722, Opus) are not modelled here, because they use an extended R scale.
Frequently asked questions
What is a good MOS score for VoIP?
4.0 or higher is good for business calls. Narrowband VoIP tops out around 4.4, so you won't see 5.0 even on a perfect network.
Is ping time the same as latency here?
Ping measures round-trip time. Voice quality depends on one-way delay, so leave 'This is round-trip time' ticked when you enter a ping result and the tool will halve it.
What matters most for call quality?
Packet loss usually does the most damage, followed by high delay. Jitter matters mostly because a large jitter buffer adds delay.
Why is my MOS lower with G.729?
G.729 compresses audio to 8 kbps, which costs some quality (Ie = 11) before any network problems. It's a trade-off for using about a third of G.711's bandwidth.