VoIP Call Quality Troubleshooting
Choppy, robotic, or dropping calls? We find the real cause — not a guess.
Bad VoIP audio is almost never the phone's fault. It is jitter, packet loss, an oversubscribed internet circuit, a switch without QoS, or a firewall rewriting SIP packets. We measure your network under real call load, identify which of those is happening, and fix it. Most NYC offices we visit have two or three of these problems stacked on top of each other.
Symptoms we diagnose every week
Each of these points to a different layer of the network, so the symptom itself is the first diagnostic clue.
- Choppy or robotic audio — jitter above 30 ms or packet loss above 1%
- One-way audio — NAT traversal or SIP ALG rewriting media paths
- Calls dropping at 30 seconds or 15 minutes — SIP timers and firewall session timeouts
- Echo on outbound calls — impedance mismatch on an analog gateway or handset headset gain
- Delay or talk-over — latency over 150 ms one-way, often a routing or VPN hairpin issue
- Audio fine internally, bad externally — upstream bandwidth saturation or carrier-side trunk issues
How we test
We do not replace hardware on a hunch. We capture evidence first, then fix the layer the data points to.
- Continuous MOS, jitter, latency, and packet loss monitoring against your SIP provider over 24–72 hours
- Packet captures at the handset, switch, and firewall so we can see exactly where frames are lost
- Switch port error and duplex audits, plus PoE budget verification
- QoS policy review: DSCP marking, trust boundaries, and whether marks survive to the WAN edge
- Firewall inspection for SIP ALG, aggressive UDP timeouts, and asymmetric routing
- Circuit validation — actual sustained upstream throughput versus what your ISP sold you
What the fix usually looks like
Once we know the cause, remediation is specific and testable. You get a written before/after with measured numbers.
- Voice VLAN separation with DSCP EF marking for RTP and CS3 for SIP signalling
- Disabling SIP ALG and setting correct UDP timeout values on the firewall
- Replacing unmanaged switches with managed PoE switches that honor QoS
- Bandwidth shaping so file sync and video conferencing cannot starve voice traffic
- Codec changes (G.722 to G.711) where bandwidth is the constraint
- Adding a second circuit or SD-WAN failover for offices where one ISP is not reliable enough
Why NYC offices see this more often
Manhattan and Brooklyn buildings frequently have shared risers, aging in-building copper, and landlord-supplied internet with contention ratios nobody discloses. We work in these buildings daily and know which conditions to test for first — including cabling that was certified years ago and has since been patched into by other tenants.
Frequently asked
How fast can you diagnose a call quality problem?+
We can usually identify the cause within one onsite visit. Intermittent issues need 24–72 hours of monitoring before we can point to a specific cause with confidence.
What jitter and packet loss numbers are acceptable for VoIP?+
Target under 30 ms jitter, under 150 ms one-way latency, and under 1% packet loss. Above those thresholds, callers start hearing audible degradation.
Do we need to change VoIP providers to fix this?+
Rarely. In most cases the problem is inside your own network or firewall. We only recommend a provider change when monitoring proves the fault is upstream of your circuit.
Will you work with our existing IT company?+
Yes. We routinely hand our packet captures and QoS recommendations to in-house or outsourced IT teams and can implement the changes ourselves if they prefer.