- VoIP carries voice as data packets over a broadband connection, so call quality depends on the network conditions of that connection rather than a dedicated voice circuit.
- Jitter, the variation in packet arrival timing, and packet loss are the main technical causes of choppy or broken VoIP audio, as described in ITU guidance on voice over IP.
- Each concurrent VoIP call uses a modest amount of bandwidth, but the figure varies with the codec chosen, so total demand scales with the number of simultaneous calls.
- Quality of Service settings on a router can prioritise voice packets over other traffic, reducing the impact of congestion on calls.
- As Openreach completes the all-IP migration in 2027, landline calls increasingly run over broadband, making these quality factors relevant to ordinary phone users as well as businesses.
VoIP call quality is driven by jitter, packet loss, latency, available bandwidth and codec choice; prioritising voice traffic with router Quality of Service settings and a stable connection improves it most.
Last reviewed: June 2026
Why VoIP quality depends on the network
Voice over Internet Protocol, or VoIP, converts speech into digital packets and sends them across a broadband connection rather than over a dedicated voice circuit. That design brings flexibility and lower call costs, but it also means the audio shares the same connection as everything else on the network. When that connection is congested or unstable, the voice packets suffer alongside other traffic, and the caller hears the result.
This dependence on network conditions is the single most important thing to understand about VoIP quality. A copper telephone line gave each call its own steady channel, whereas a VoIP call competes for capacity and is sensitive to timing. As Openreach completes the all-IP migration in 2027, more ordinary landline calls run this way, so these factors matter to home users as well as to businesses.
It also explains why a VoIP call can sound flawless one minute and degrade the next on the same connection. Voice is unusually demanding not because it needs much bandwidth, but because it needs that bandwidth delivered in a steady, low-delay stream. A file download tolerates packets arriving in bursts and out of order, because the file is reassembled before use; a live conversation cannot wait, so any packet that arrives late or not at all is heard immediately as a glitch. The factors covered below all come back to that one principle: voice quality is about the consistency of delivery, not just the raw speed of the line.
Jitter, packet loss and latency
Jitter is the variation in the time it takes successive voice packets to arrive. Speech is reconstructed by playing packets back in a smooth stream, so when their arrival timing fluctuates, the audio can sound choppy or robotic. A small buffer smooths out minor jitter, but heavy variation overwhelms it and the call degrades. ITU guidance on voice over IP describes jitter as one of the core impairments that determine perceived quality.
Packet loss occurs when packets fail to arrive at all, usually because of congestion. Lost packets create gaps, clipped words or dropouts, since the missing audio cannot be recovered. Latency, the one-way delay between speaking and being heard, is a third factor: high latency does not garble audio but causes awkward talk-over, where both parties start speaking because the delay disrupts the natural rhythm of conversation. Together these three impairments account for most VoIP quality complaints.
These three rarely arrive on their own, which is why diagnosis matters. Congestion on a busy connection commonly drives latency and jitter up and starts to cause loss at the same time, so a single underlying cause can produce several symptoms at once. The jitter buffer that smooths arrival timing does so by holding packets briefly before playback, which itself adds a little delay, so there is a genuine trade-off between smoothing jitter and keeping latency low. ITU work on one-way transmission time sets out why conversation begins to feel unnatural once total delay grows beyond a modest threshold, which is the reason a connection that tests fast on raw speed can still feel awkward to talk over if its delay is high.
Bandwidth and codec choice
Each VoIP call uses a modest amount of bandwidth in both directions, but the exact figure depends on the codec, which is the method used to compress and encode the audio. A higher-quality codec uses more bandwidth per call but sounds clearer, while a more compressed codec uses less at the cost of fidelity. Because demand is per call, a site with many simultaneous calls needs its bandwidth planned around peak concurrency.
The practical implication is that available bandwidth must comfortably exceed the combined demand of all concurrent calls plus the rest of the network's usage. A connection that looks fast on a speed test can still produce poor calls if voice traffic has to compete with large downloads or video streams. Matching the codec to the connection, and ensuring headroom above peak call volume, are both part of keeping quality consistent.
The upload direction deserves particular attention on home and small-office connections. Many broadband packages are asymmetric, offering far more download than upload capacity, yet a voice call sends as much data outward as it receives. A connection that downloads quickly can still produce broken outbound audio if its upload is saturated by a cloud backup, a large email attachment or someone else's video call. Where the codec can be selected, choosing a more efficient one reduces the per-call demand and buys headroom, though it is no substitute for an upload path that can carry every concurrent call with room to spare.
VoIP quality factors and typical fixes
The table below summarises the main quality factors, the symptom each produces and the practical fix that usually addresses it.
| Factor | Typical symptom | Typical fix |
|---|---|---|
| Jitter | Choppy or robotic audio | Enable QoS and a jitter buffer |
| Packet loss | Dropouts and clipped words | Reduce congestion, use wired connection |
| Latency | Talk-over and delay | Choose a lower-latency route or connection |
| Insufficient bandwidth | Quality drops during busy periods | Increase capacity or limit competing traffic |
| Codec mismatch | Muffled or overly compressed audio | Match codec to available bandwidth |
Router configuration and Quality of Service
The router is where many quality problems can be addressed. Quality of Service, usually shortened to QoS, is a feature that lets the router prioritise voice packets ahead of less time-sensitive traffic such as file downloads or backups. When QoS is configured to recognise and protect voice, calls hold up better even when the connection is busy, because the voice packets are not left waiting behind bulk data.
Other router-level steps help too. A wired Ethernet connection for the handset or adaptor avoids the variability of Wi-Fi, which can introduce jitter and loss. Keeping firmware current, avoiding double-NAT configurations and placing the VoIP device close to the router on the network path all reduce the chance of impairment. Where a business runs many calls, separating voice onto its own network segment further insulates it from other traffic.
It is worth understanding why Wi-Fi causes so many avoidable problems, because it is the most common culprit in a home setup. A wireless link is a shared, variable medium: interference from neighbouring networks, distance from the access point and obstructions such as walls all cause packets to be retransmitted, and each retransmission adds delay and jitter to a stream that cannot tolerate either. A phone or adaptor that roams between access points, or sits at the edge of coverage, will produce intermittent quality that is hard to diagnose because it changes with conditions. Moving the device onto a wired connection removes that whole class of fault, which is why it is usually the first step a support team will suggest.
How to test and diagnose VoIP quality
Diagnosing a quality problem starts with measuring the connection under realistic conditions. Online VoIP test tools report jitter, packet loss and latency, and running them at quiet and busy times reveals whether congestion is the cause. If quality drops only when the office is busy or a large download is running, that points firmly at bandwidth or QoS rather than the VoIP service itself.
It also helps to isolate variables methodically: test on a wired connection to rule out Wi-Fi, try a second device to rule out a faulty handset, and check whether the problem affects all calls or only certain destinations. Persistent loss on a connection that is otherwise lightly used may indicate a line fault worth raising with the broadband provider. Keeping a simple record of when problems occur makes the pattern clearer and speeds up any support conversation.
When to escalate to your provider
Some quality problems sit outside anything that can be fixed at the router or the handset, and recognising that boundary saves wasted effort. If methodical testing shows steady packet loss or high latency on a wired connection that is lightly loaded, the issue is more likely to lie in the broadband line itself or the wider network than in the local setup, and it is then reasonable to raise a fault with the provider. Bringing dated test results, a note of the affected destinations, and the times the problem occurs gives the provider concrete evidence to investigate rather than a general complaint.
It also helps to separate the broadband provider from the VoIP service provider, since they may be different companies. A fault that affects every internet activity, not just calls, points to the broadband line; a fault that affects only calls to particular numbers or only one VoIP account points to the voice service or its configuration. Ofcom publishes information on broadband speed and performance expectations and on how complaints are handled, which sets the backdrop for what a provider should investigate once a clear, evidenced fault has been reported.
Frequently Asked Questions
Why is my VoIP call quality poor?
Poor VoIP quality usually comes from network conditions on the broadband connection rather than the phone service itself. Jitter, packet loss, high latency or insufficient bandwidth during busy periods are the most common causes. Testing the connection at quiet and busy times helps identify which factor is responsible, and a wired connection rules out Wi-Fi as the source.
What is jitter in VoIP?
Jitter is the variation in the time it takes successive voice packets to arrive at their destination. Because speech is reassembled into a smooth stream, inconsistent arrival timing makes audio sound choppy or robotic. A jitter buffer smooths out small fluctuations by holding packets briefly before playback, but heavy jitter overwhelms it and degrades the call.
How much bandwidth does VoIP need?
Each VoIP call uses a modest amount of bandwidth in both directions, with the exact figure depending on the codec used to compress the audio. Total demand scales with the number of simultaneous calls, so capacity should comfortably exceed peak concurrent usage plus other traffic. A connection that is fast on a speed test can still struggle if voice competes with heavy downloads, and on asymmetric lines the upload path is often the limiting factor.
Does latency affect VoIP quality?
Latency is the one-way delay between speaking and being heard, and high latency causes awkward talk-over rather than garbled audio. When the delay grows, both parties start speaking at once because the natural rhythm of conversation is disrupted. ITU guidance on one-way transmission time explains why conversation feels unnatural beyond a modest delay threshold, so choosing a lower-latency connection or route reduces this effect.
How do I improve VoIP call quality?
Enabling Quality of Service on the router to prioritise voice traffic is one of the most effective steps, alongside using a wired connection instead of Wi-Fi. Ensuring bandwidth comfortably exceeds peak call demand, matching the codec to the connection and reducing competing traffic all help. Testing for jitter, loss and latency identifies which fix will have the most impact, and a clearly evidenced fault can then be escalated to the provider.