VoIP Call Quality Problems: Fixes for AU Business

The Three Factors That Determine VoIP Call Quality

This guide diagnoses and fixes the most common VoIP call quality problems on Australian NBN connections: choppy audio, echo, one-way audio, dropped calls, and crackling. Written by an independent editorial team with direct experience troubleshooting VoIP on Australian NBN connections. Unlike generic troubleshooting articles, this covers the specific Australian factors that cause quality issues: FTTN upload limitations, ISP CVC congestion during peak hours, SIP ALG on Telstra smart modems, and QoS misconfiguration on consumer-grade routers.

NBN Call Quality Benchmarks by Connection Type

Not all NBN connections deliver the same VOIP performance. The connection type between your premises and the fibre backbone determines how much latency and packet loss variability you are exposed to. Here is what each NBN technology means for VOIP call quality in Australia:

For VOIP call quality, the acceptable thresholds are: one-way latency under 150ms (under 80ms for good quality), jitter under 30ms, and packet loss under 1%. Most FTTP and FTTC connections stay well within these parameters. FTTN connections on longer or impaired copper runs can have poorer performance. Busy-hour congestion, when present, may arise in the NBN access network or the provider's network; CVC was an aggregate point-of-interconnection construct, not contention at the node.

SIP ALG: The Most Common Cause of AU VOIP Problems

SIP ALG (Session Initiation Protocol Application Layer Gateway) is a feature built into many consumer and business routers that was designed to help VOIP calls pass through NAT. In practice, it causes more problems than it solves for modern hosted cloud PBX deployments. SIP ALG inspects and modifies SIP packets as they pass through the router, which frequently corrupts the signalling and causes symptoms including:

• Calls connecting but with one-way audio (you can hear the other person, they cannot hear you, or vice versa)
• Calls that ring but cannot be answered
• Calls that drop after exactly 30 seconds or 60 seconds
• Registration failures where phones cannot register to the SIP server
• Intermittent call quality that varies by time of day or resets after a router reboot

How to disable SIP ALG on common Australian routers:

• TP-Link (Archer series, Deco): Advanced > NAT Forwarding > ALG > disable SIP. On older firmware: Advanced > Security > disable SIP ALG.
• Netgear (Nighthawk, Orbi): Advanced > Advanced Setup > WAN Setup > disable SIP ALG checkbox. Some Netgear firmware hides this, check the router admin panel at 192.168.1.1.
• ASUS (RT series): Advanced Settings > WAN > NAT Passthrough > set SIP Passthrough according to your VoIP provider's guidance.Telstra Smart Modem (Sagemcom, Technicolor): Telstra-branded modems have limited admin access. Your business phone company may be able to work around SIP ALG using SIP over TLS for signalling, or a VPN tunnel. Note SRTP encrypts the audio (RTP) stream, not SIP signalling, so enabling SRTP alone does not fix an ALG-related signalling problem.
• Ubiquiti (UniFi, EdgeRouter): SIP ALG is not enabled by default on Ubiquiti hardware. If someone has enabled it manually, it can be disabled in the firewall ALG settings.

For a deeper dive into threats and hardening steps, see our VoIP security basics guide.

QoS Configuration for VOIP on Australian Business Routers

QoS (Quality of Service) prioritises VOIP traffic on your network so that a large file download or video stream does not degrade call quality. QoS may be unnecessary on a lightly loaded connection, but its usefulness depends on actual utilisation-especially upload saturation-and router queuing, not the nominal NBN speed tier alone. It becomes relevant when: your available upload bandwidth is under 10 Mbps (FTTN on a long run), you have multiple staff on calls simultaneously, or you run other bandwidth-intensive applications (video conferencing, large file transfers) alongside VOIP.

QoS settings for common scenarios:

• DSCP marking: Tag VOIP packets with DSCP EF (Expedited Forwarding, value 46) for voice and DSCP AF41 (value 34) for video where your provider and network equipment support it, confirm the exact marking with your provider rather than assuming it. DSCP/QoS markings are mainly useful once configured and preserved across your own LAN/WAN, they are not guaranteed to survive across the public internet.
• Port-based QoS/firewall rules: Follow the exact SIP and RTP port requirements published by your specific VoIP provider or PBX vendor, for example SIP signalling on UDP 5060 and RTP audio on a provider-defined range such as UDP 10000-20000, since ranges vary by provider. Do not expose or port-forward broad SIP/RTP ranges to the public internet unless your provider or PBX architecture specifically requires it.
• Bandwidth reservation: Budget roughly 100 kbps in each direction per G.711 call as a practical baseline, then add capacity headroom rather than reserving the bare minimum. For a 5-seat office with potentially 3 concurrent calls: reserve at least 300-500 kbps upload as a conservative allowance. On an NBN 25 plan (5 Mbps upload), this is a small fraction of available upload, negligible. On an FTTN connection with poor sync speed (1-2 Mbps upload), it matters.

Diagnosing VOIP Call Quality Problems: Is It Your ISP or Your Provider?

When call quality is poor, the problem is in one of four places: your ISP connection, your local network (router, switches, Wi-Fi), your business phone company's platform, or the carrier network the other party is on. Blaming the wrong party wastes time. Here is how to narrow it down:

Step 1: Test your internet connection quality. Use a VOIP-specific speed test (not a general speed test, you need latency, jitter, and packet loss, not just bandwidth). Tools: PingPlotter (free, Windows/Mac), Cloudflare speed test (speed.cloudflare.com) includes latency and jitter. Run the test while call quality is poor, not when things are working fine. Treat high latency or packet loss as warning signs, not attribution: 150ms is commonly a one-way voice-design limit, while browser speed tests generally report round-trip latency, and either result can reflect the local network, test path or destination as well as the ISP.

Step 2: Test via wired vs wireless. Plug a laptop directly into the router via Ethernet and make a test call via softphone. If quality improves, your Wi-Fi is the issue, not your ISP or provider. Business VOIP phones should always be on wired Ethernet where possible; Wi-Fi VOIP is workable but introduces jitter from wireless interference.

Step 3: Test with a different provider's platform. Some Australian cloud phone system companies offer a trial or demo account; availability varies by provider. Equally poor quality on a second platform suggests that a shared factor such as the local network or internet path should be investigated, but it does not by itself prove where the fault lies. If quality is better on the trial, your current provider's platform or routing may be the issue.

Escalate to your ISP when: latency above 150ms on a wired connection, packet loss above 1% on a wired connection, quality degrades consistently during the busy evening period (7pm-11pm), or quality degrades during rain or wet weather (FTTN copper fault indicator).

Escalate to your business phone company when: one-way audio persists after SIP ALG is disabled, registration failures on specific handsets or softphones, calls drop at a consistent time interval (30s, 60s), or quality is poor only when calling specific number ranges (interstate, mobile, international).

How Australian NBN Connection Types Affect Call Quality

NBN connection type is one important variable in Australian VoIP quality, alongside the local network, provider network, endpoint configuration and individual line condition. FTTP (Fibre to the Premises) delivers the most consistent low-latency performance because the fibre runs all the way to your building. FTTN (Fibre to the Node) relies on copper from the node to your premises, and the quality of that copper run has a direct impact on latency and packet loss. If your business phone stopped working recently after an NBN switchover, see Business Phone Stopped Working After NBN: What to Do for immediate steps.During the busy evening period (7pm-11pm) on congested NBN infrastructure, even FTTP connections can show elevated latency. For business users who need consistent quality during business hours, NBN performance during business hours is generally more reliable than residential peak hours.

How to Measure Your Connection Quality for VoIP

Some speed tests report only throughput, but others-including Cloudflare's-also report latency, jitter and packet loss. To assess whether your connection can support VoIP well, you need to measure latency, jitter, and packet loss. Tools for this:Ping test to your business phone company's SIP server (ask them for the server IP or hostname). Most routers have a built-in ping tool, or use the Windows or macOS command prompt. Run 100 pings and look at average, minimum, maximum, and any lost packets. For a more comprehensive test, use Cloudflare's speed.cloudflare.com which measures latency, download speed, and jitter. Run tests during your business hours, not late at night when the network is quieter.

Codec Selection: G.711 vs G.722 vs G.729

VoIP audio is compressed using a codec (coder-decoder). The codec affects both call quality and bandwidth usage. For Australian NBN connections, the recommended choices are:

QoS: Prioritising Voice Traffic on Your Network

Quality of Service (QoS) configuration on your router ensures that VoIP packets are prioritised over other network traffic. Without QoS, a large file upload or a software update download can introduce enough network congestion to degrade active calls. QoS becomes useful when voice traffic is competing with other traffic for constrained network capacity, particularly upload bandwidth, rather than being required purely because of staff count. A lightly used fast connection may run fine without it; a heavily used or bandwidth-constrained connection can benefit substantially. See our business VoIP setup guide for a full walkthrough.Most business-grade routers support QoS. The general approach: follow your provider's documented SIP and RTP port ranges (commonly UDP 5060 for SIP signalling and a provider-specific RTP range such as UDP 10000-20000) and mark that traffic with the highest priority queue on your own network. Do not treat these port numbers as universal or safe to expose directly to the internet without checking your provider's documentation first. This ensures voice packets are processed ahead of less time-sensitive traffic even during periods of high network utilisation.

Diagnosing Specific Call Quality Problems

See our NBN VoIP setup guide for step-by-step instructions on checking your NBN line quality and configuring your router.

When to Call Your ISP Versus Your business phone company

One of the most frustrating aspects of VoIP troubleshooting is determining who is responsible for a call quality issue. Both your ISP and your business phone company will naturally lean toward attributing the problem to the other party. Having a clear framework for isolating the issue saves time and prevents the blame-shifting loop that can leave a problem unresolved for weeks.

If you are unsure how your NBN connection type affects VoIP, our NBN compatibility guide breaks it down by technology type.

The ISP is responsible for everything from the NBN connection point to your router WAN port. This includes: the quality and stability of the NBN line itself (line sync rate, error rates on FTTN copper, HFC coaxial noise), the latency from your premises to the ISP's point of interconnect, packet loss at the NBN or ISP network level, and congestion on the ISP's internal network during peak hours. Identify the ISP-side hop carefully: from a LAN device, traceroute's first hop is normally your own router, while the first hop beyond it may be an ISP device and may not answer probes reliably. Elevated results to an ISP-side hop justify further path testing, but do not conclusively assign responsibility because routers may rate-limit or deprioritise ICMP responses.

Your business phone company is responsible for everything from their SIP platform to your phone registration. This includes: SIP server availability and response time, codec configuration on your account, media server routing quality, and the accuracy of your account configuration. To isolate provider responsibility, run a ping test directly to your provider's SIP server address (ask them for it). Compare the results to your ISP gateway ping. A clean gateway ping and poor SIP-server ping justify further investigation, but do not isolate the provider: the server may treat ICMP differently, and the actual RTP media path may use different hosts. Your router and local network (including SIP ALG, QoS, and firewall rules) are your own responsibility to configure correctly before escalating to either party.

Call Quality During Peak Hours on Australian NBN

A specific pattern affects Australian VoIP users on the NBN: call quality that is consistently good during business hours but can degrade noticeably during the busy evening period (7pm-11pm). This is peak hour congestion on the NBN, and it is a known characteristic of the network, particularly for residential-grade NBN plans.

The cause is congestion on the ISP's own network, upstream of the NBN access network itself, sometimes still referred to by the older term CVC (Connectivity Virtual Circuit) contention, the capacity an ISP provisions to carry its customers' traffic. During peak hours, many customers are simultaneously streaming video, gaming, and making video calls. An ISP with insufficient provisioned capacity on a given link experiences congestion at that network layer, which shows up as increased latency and packet loss across all traffic, including VoIP calls. Congestion severity varies significantly by ISP and by the specific access network in your area, not by a single universal rule.

For businesses whose staff work office hours and whose customers call during business hours (9am to 6pm), evening congestion may never be a material issue. For businesses with staff who work from home in the evenings, or whose customers call after hours, it is worth testing call quality during the busy evening window specifically. If you consistently observe degraded quality during this period, the most effective options are: checking whether your plan or provider includes any enhanced service level or traffic-priority features (these are optional and provider/plan-specific, a "business" plan name does not itself guarantee better congestion performance), switching to an ISP with a track record of lower peak-time congestion (this varies significantly between providers, check independent speed-test data for your area), or routing after-hours calls to a mobile rather than relying on the home internet connection during peak congestion. Confirm the actual SLA/eSLA and inclusions of your specific plan rather than assuming the word "business" guarantees it.

Long-Term Call Quality Monitoring

Reactive troubleshooting, where you only investigate call quality when a problem is already affecting customers, is less effective than proactive monitoring. Some cloud phone system platforms provide per-call quality reporting or estimated MOS data; availability and calculation methods vary by platform and licence. Understanding how to use this reporting turns call quality management from a reactive exercise into a proactive one.

MOS (Mean Opinion Score) is the standard measure of voice call quality. MOS originally refers to the mean of subjective listener ratings on a 1-to-5 scale; platform reports usually show an estimated MOS calculated by a model using factors such as delay, packet loss and codec characteristics. A score above 4.0 is considered toll-quality (equivalent to PSTN). A score between 3.5 and 4.0 is acceptable for business voice. Below 3.5, quality is noticeably degraded. Below 3.0, calls are difficult to conduct. Your provider's call quality reports will show MOS scores per call, often with the underlying metrics (latency, jitter, packet loss) available for deeper investigation.

Review your call quality reports weekly during the first month after go-live. Look for patterns: are low-MOS calls concentrated at specific times of day (suggesting congestion), on specific extensions (suggesting a local network issue with that phone or switch port), or on calls to specific destinations (suggesting a routing issue on the provider's network)? Patterns in the data point you toward the right fix. A provider without customer-facing call-quality reporting offers less proactive visibility, although provider-side and external monitoring can still be used to diagnose quality. This should be a requirement, not a nice-to-have, when evaluating providers.

What Most People Get Wrong About VoIP Quality

1. Assuming it must be your business phone company without checking first. Call quality issues can originate in your local network, your internet/access path, your business phone company's platform, or the carrier/interconnect handling the other party's call. In practice, local network and internet-path issues are a common cause, but diagnose using the steps in this guide before assigning blame.2. Running a speed test and assuming everything is fine. A throughput-only test is insufficient for VoIP, so use a test that also reports latency, jitter and packet loss. Use a VoIP-specific test that reports jitter (under 30ms needed) and packet loss (under 1% needed).3. Upgrading the NBN plan without fixing the real problem. Going from NBN 50 to NBN 100 will not fix call quality if the problem is SIP ALG, QoS, or CVC congestion during peak hours. Diagnose before spending more.

Your Next Steps

1. Check for SIP ALG on your router first. It is a common cause of one-way audio and registration failures with some VoIP platforms and routers, so it is worth ruling out early.
2. Enable QoS for voice traffic. Follow your provider's documented SIP/RTP port ranges (commonly SIP on port 5060) and prioritise that traffic on your own network.
3. Test during the busy evening period (7pm-11pm). This is when NBN/ISP congestion is most likely to show up.
4. Check your upload speed. Budget roughly 100 kbps upload per concurrent G.711 call as a practical baseline, then add headroom, for example at least 1 Mbps upload for 5 concurrent calls as a conservative allowance, not a bare minimum. FTTN connections may struggle.
5. Consider a dedicated VLAN for voice. A voice VLAN separates voice from data for management and policy purposes, but preventing congestion also requires correctly configured QoS and sufficient capacity. Our setup guide walks you through this.
6. If problems persist, contact your ISP about congestion. If quality is consistently poor during the busy evening period specifically, ask your ISP about congestion on your access network.Still having issues? Get a free recommendation and we will help you assess whether your connection can support VoIP or if you need to consider alternatives.

Power outages are a call quality issue that most businesses do not plan for until the first blackout takes the phones offline. For the full picture on keeping calls up during outages, see our guide to NBN battery backup for VOIP, including which UPS devices work with NBN equipment and what backup-power options exist for your specific NBN technology.

Power Outages: When Call Quality Drops to Zero

The most overlooked VoIP call quality issue is not jitter or packet loss. It is a complete outage. On NBN, your VoIP system depends on mains power for the equipment involved, and when the power goes out, call quality does not degrade gracefully. It stops.

Unlike the old copper PSTN (which carried its own power from the exchange), the equipment a VoIP service depends on during an outage varies by your specific NBN technology and voice setup. Depending on your connection, this can include the NBN connection box (NTD) on FTTP/HFC/FTTC, a VDSL modem on FTTN/FTTB, your router, switches and Wi-Fi access points, and your VoIP phones, ATA or PBX, plus upstream network infrastructure the provider does not control.

Mitigation options:

• UPS (uninterruptible power supply): A basic UPS (~$150-250 AUD, check current pricing) on your local equipment can extend operation during an outage. Runtime depends on battery capacity and load, and does not guarantee the wider access network or your provider's network remains operational, so treat it as extending local resilience rather than a guaranteed number of hours of service.
• Mobile failover: Some cloud phone system companies (including Maxotel) can automatically redirect calls to mobile numbers during outages, so callers still reach your team.
• 4G/5G backup: A mobile broadband failover router keeps your internet (and VoIP) running when NBN drops, though this only helps if the mobile network itself is unaffected by the same outage, and call quality still depends on mobile signal strength at your location.

Call-recording availability varies by provider, licence and configuration; where available, it may be used for quality monitoring or dispute resolution. Before enabling it, review your obligations under Australian law. See Call Recording Laws in Australia for a full guide.

If calls fail to connect or are rejected only when several people are already on the phone, the likely cause is an undersized SIP trunk count relative to concurrent call demand, a call-capacity issue rather than an audio-quality one. It does not normally make an already-connected call sound choppy or distorted. Our guide to how many SIP trunks you need walks through the concurrent call calculation, seasonal peak adjustments, and the correct sizing formula for Australian SMBs using cloud phone system or on-premise systems.

Voicemail delivery problems are sometimes mistaken for call quality issues , missed voicemail notifications look like missed calls in the call log. If your team is experiencing unreliable voicemail delivery, our guide to voicemail to email setup for Australian businesses covers the configuration steps and common causes of email delivery failure, including SMTP relay, SPF/DKIM, and provider-specific settings.

Does Better Call Quality Cost More?

Call quality is not directly tied to plan price, a lower-cost cloud phone plan on a good NBN connection can deliver better call quality than a more expensive plan on a congested or incompatible connection. Quality also depends on the local network, ISP and access path, provider media routing and endpoint configuration; a plan upgrade helps only when capacity or plan-specific service characteristics are contributing. Australian cloud phone system pricing varies widely by provider, call inclusions, licence level, support and SLA, so treat any headline monthly figure as an example rather than a guaranteed market rate, and check whether a stated SLA actually covers voice quality or uptime specifically before assuming a higher price buys better call quality.

A 4G/5G failover setup can improve resilience while the NBN service is unavailable, but it does not repair the primary connection and still depends on mobile coverage, power and network availability. Our guide to VoIP over 4G and 5G failover for Australian businesses covers the hardware required, the automatic failover configuration, and what a redundant setup costs so calls stay live even when NBN goes down.

Why do my VoIP calls sound good in the morning but bad in the afternoon?

This pattern can indicate time-dependent congestion, but the congested point may be in the local network, the provider's network or an NBN access segment; testing is needed before attributing it to FTTN, HFC, a node or a pillar. Solutions include: requesting an NBN upgrade to FTTP (where available), using QoS to maximise priority of voice traffic on your internal network, or contacting your NBN provider to report congestion.

How much bandwidth does VoIP use?

Each active VoIP call uses approximately 80-100 kbps of bandwidth (with G.711 or G.722 codecs). Ten simultaneous calls use less than 1 Mbps. Voice usually uses modest bandwidth, but constrained or heavily utilised upload and download links can still limit call quality. Even an NBN25 connection can support multiple simultaneous calls. Latency and jitter are far more important than raw bandwidth speed.

Will a better NBN plan improve my VoIP quality?

A higher NBN speed tier may not help when capacity saturation is not the cause, but it can help when added upload or download capacity reduces local congestion. However, upgrading your NBN technology type (e.g. from FTTN to FTTP, where FTTP upgrade is available in your area) can significantly improve quality because it replaces the variable copper run with consistent fibre. Contact your NBN provider or check NBN Co's website for upgrade availability at your address.

Is my business phone company responsible for call quality problems?

Your business phone company is responsible for quality on their network from their data centre outward. They are not responsible for the quality of your NBN connection or your internal network. Call quality problems can originate in the internal network, NBN or ISP path, phone platform, or carrier and interconnect networks, so responsibility should be established through testing. Start troubleshooting at your router and NBN connection before escalating to your business phone company.

If you have worked through this guide and still have call quality issues, the problem is diagnosable. A good business phone company with Australian support staff should be able to work through this with you. Troubleshooting your call quality is part of what you are paying them for, not an extra. If your provider cannot or will not help you diagnose the issue, that tells you something useful too.

My calls have an echo. What causes it and how do I fix it?

Echo heard by one party is often generated toward the other end of the call, but causes include acoustic coupling, hybrid impedance mismatch and echo-cancellation faults. The person you are speaking with has either a handset with a poorly sealed earpiece (allowing sound to leak from the speaker into the microphone) or speaker phone is active and the room acoustics are causing their microphone to pick up your voice from the speaker and send it back to you. Ask the other person to check their handset seal or switch from speakerphone to handset. If echo occurs on all calls, investigate the local handset or headset, analogue interfaces, gateway echo cancellation and endpoint configuration rather than assuming a sidetone firmware setting.

What does "one-way audio" mean and how is it fixed?

One-way audio means you can hear the other person but they cannot hear you, or vice versa. It is caused by a routing asymmetry in the audio path, most commonly from a firewall or NAT device blocking the return path of RTP (audio) packets. SIP ALG is one possible cause of one-way audio because it can alter SIP addressing and media information; firewall, NAT and provider-side media-routing faults are other possibilities. Disable SIP ALG on your router, check that your firewall is not blocking the RTP port range your provider uses (typically UDP 10000-20000, confirm with your provider), and test again. If one-way audio persists, the provider should trace the full signalling and RTP path, including the endpoint, firewall or NAT device, PBX or SBC, and provider media routing.

How much internet bandwidth do I need for VoIP?

A single simultaneous VoIP call using the G.711 codec requires approximately 100kbps of upload and download bandwidth. G.729 (a compressed codec) reduces this to around 32kbps per call at the cost of slightly lower audio quality. In practice, the bandwidth requirement for VoIP is very low. A four-person office where all four people are on calls simultaneously needs 400kbps of upload capacity for voice, which virtually every NBN plan exceeds. The more important measure is latency and packet loss on your upload path, not raw speed. A 25Mbps connection with low latency and no packet loss delivers excellent VoIP quality. A 100Mbps connection with high jitter and 2 percent packet loss delivers poor quality.

Related reading:

• NBN VoIP setup guide
• NBN compatibility for VoIP