Skip to content

VoIP One-Way Audio and Dropped Calls: Causes and Fixes

VoIP one-way audio and dropped calls: NAT, SDP, SIP ALG, firewalls, codecs, the 32 second drop and session timers, with a fix for each cause.

Updated 20 October 2026

One-way audio on VoIP calls is almost always a media path problem: NAT, a wrong IP address in the SDP, a firewall blocking RTP ports or a codec mismatch. Calls that drop at a fixed time, such as around 32 seconds or after 15 or 30 minutes, are usually signalling problems: a missing ACK or failed session refresh.

How a call carries audio

SIP sets up the call; the audio itself travels separately as RTP packets. Each side announces in the SDP body of its INVITE or 200 OK the IP address, port and codecs where it wants to receive audio. If either side announces an address the other cannot reach, or a firewall drops the packets, signalling works but audio does not. The basics of a carrier SIP connection are covered in how a SIP interconnect works.

One-way or no audio: causes and checks

CauseWhat you seeFix
Private IP in SDPSDP shows a 10.x, 172.16-31.x or 192.168.x addressConfigure the external IP on your switch or SBC, or enable NAT handling
Firewall blocks RTPPackets leave but never arrive, or the reverseOpen the RTP port range to the carrier's media IPs, not only signalling IPs
SIP ALG on a routerSDP rewritten unpredictablyDisable SIP ALG on routers and firewalls in the path
Media IPs not whitelistedCarrier sends audio from an IP you did not expectAsk the carrier for its full list of media IPs
Codec mismatchCall connects, silence or noise both waysAlign codec lists; see G.711 vs G.729
Re-INVITE handlingAudio stops after hold, transfer or a mid call changeCheck how both sides handle re-INVITE and media changes
DTMF or transcoding errorsAudio breaks after a key press or on certain callsCheck DTMF mode (RFC 2833/4733 or SIP INFO) and transcoding settings

The quickest test: capture a call with a packet capture tool and check whether RTP flows in both directions and to the IP and port announced in the SDP. If RTP arrives at your edge but not at the phone, the problem is inside your network; if it never arrives, it is on the carrier side or a firewall between you.

Calls that drop at a fixed time

  • Around 30 to 32 seconds: the side that sent the 200 OK did not receive the ACK and gives up after its retransmission timer (in standard SIP, 64 times the 500 ms base timer, which is 32 seconds). Usually a NAT or routing issue on the ACK, or a wrong Contact header.
  • At 15, 30 or 60 minutes, or another round number: session timers. One side expects a session refresh (re-INVITE or UPDATE) that never comes or is rejected. Align session timer settings, or check for a maximum call duration set by the carrier.
  • After a fixed silence: RTP timeout. A device ends the call when it receives no audio for a set time, which often happens when one-way audio is already present.

Calls that drop at random

Random drops usually come from network instability (packet loss on the link, a failing SBC), from the destination network, or from a carrier chain where one hop has capacity problems. Check whether the drops are concentrated on one route or destination and who sent the BYE in your CDRs. A rising share of short calls is often the first statistical sign, see short duration calls and MOS, jitter and packet loss.

What to send the carrier

A ticket gets fixed faster with evidence: call time in UTC, called and calling number, Call-ID, a SIP trace (pcap) showing the problem, and which direction lost audio. If only one route shows the issue with the same setup, the problem is likely on that route; add it to your notes from route testing.

VoIPMatcher manages each route it sources with the supplier, so technical issues are handled in one place. To get VoIP termination for a destination, request a VoIP route.

Frequently asked questions

Should I disable SIP ALG?

In most cases, yes. SIP ALG on routers and firewalls tries to rewrite SIP and SDP for NAT but often does it incorrectly, causing one-way audio, failed registrations and dropped calls. Disable it on every device in the path, handle NAT on your switch or session border controller instead, then retest with a packet capture.

Why does VoIP audio work one way only some of the time?

Intermittent one-way audio usually comes from NAT bindings that expire, a carrier sending media from several IPs that are not all allowed in your firewall, or re-INVITEs after hold or transfer that change the media address. Compare packet captures of a good call and a bad call to see which address or port changed.

Which RTP ports need to be open for VoIP?

It depends on the equipment. Many platforms use a UDP range somewhere between 10000 and 20000, but check your own switch or SBC configuration and the carrier's documentation. Open that range only to the carrier's media IPs, not to the whole internet, and remember that media IPs often differ from signalling IPs.

Can a firewall cause one-way audio on VoIP calls?

Yes, any firewall can, when SIP inspection or SIP ALG is enabled or when the RTP port range is closed to the carrier's media IPs. Check the vendor's documentation for turning off SIP ALG or the SIP helper, then allow the RTP range from the carrier's media addresses and retest with a packet capture on both sides.

Is one-way audio the carrier's fault?

Not usually. Most one-way audio comes from NAT, firewall or SDP settings at one end of the call. A packet capture tells you where to look: if RTP never arrives at your edge, the cause is on the carrier side or in a firewall in between; if it arrives but does not reach the phone, it is inside your network.

Need a route?

Send the destination, route type and target rate. We reply with what we can deliver.

More guides

All guides
Quality and testing SIP 487 Request Terminated: What It Means What SIP 487 Request Terminated means, the CANCEL call flow behind it, why dialers and slow routes produce many 487s and how to reduce them. Quality and testing What Is a CDR in VoIP? Monitoring Routes With CDRs What a CDR is in VoIP and how to monitor routes with call detail records: the fields you need, metrics to track per route and alerts that catch problems. Quality and testing OTP SMS Not Received: Causes and Fixes for Apps Why users do not receive your OTP SMS or get it late: filtering, sender ID, registration, slow routes, content issues and how to fix them. Quality and testing Caller ID Test: How to Check CLI on a VoIP Route How to run a caller ID test on a VoIP route: test setup, reading the received CLI, the delivery rate to agree and the warning signs of grey routes. Quality and testing Q.850 Cause Codes: List and Meaning for VoIP The common ITU-T Q.850 cause codes (1, 16, 17, 34, 41 and more) with their meaning, their SIP equivalents and how to use them to troubleshoot routes. Quality and testing SIP 408, 480 and 486: Timeouts and Busy Codes What SIP 480 Temporarily Unavailable, 408 Request Timeout and 486 Busy Here mean, their Q.850 causes, and when they point to a route problem. Quality and testing SIP Error Codes: Full List of SIP Response Codes Every SIP response code from RFC 3261 in one table, with its meaning and typical cause on VoIP routes, plus what 503, 487, 408, 480 and 486 mean. Quality and testing Low ASR in VoIP: Why Answer Rates Drop and What to Check Low ASR on a VoIP route: why answer rates drop, how to tell traffic problems from route problems with NER and SIP codes, and a checklist to work through.