The 6 Most Common VoIP Call Quality Problems, and How to Fix or Report Them

The 6 Most Common VoIP Call Quality Problems, and How to Fix or Report Them

The 6 Most Common VoIP Call Quality Problems, and How to Fix or Report Them

by PBX.lu Editorial on July 28, 2026
⏱ 9 min read · Tutorial · Connectivity
Bad call quality is rarely one problem. It is usually one of six, and they have different fixes. This checklist walks you through the tests in the right order, from the cheapest and fastest to the most technical, so you can find the cause yourself before you open a support ticket. Work through it in about thirty minutes. You will finish with either a working phone system or a support ticket your provider cannot send back to you.
🎧 Short answer. Most VoIP quality problems come from the last few metres, not from the provider. Check in this order: the headset, then the single device, then the local network, then the internet connection, then the provider. Roughly seven out of ten cases are resolved before you reach step four. For the rest, step 5 gives you a template that tells your provider exactly what to look at.
About this list. These six categories appear on the troubleshooting pages of providers of every size, from small local SIP resellers to the large international platforms. They are ordered for navigation, not ranked by frequency: no provider publishes reliable figures on how often each one occurs. Find the one that matches what your colleague described and start there.

Before You Start: Three Questions That Narrow It Down

Answer these before touching any settings. They eliminate most of the checklist immediately.
Does it affect one person or everyone? One person points to a headset, a device, or a desk. Everyone points to the network or the connection.
Does it happen on every call or only some? Every call points to something local. Only calls to certain countries or mobile networks points to provider routing.
Did it start suddenly or get worse gradually? Sudden usually means a change: new hardware, a firmware update, a new device on the network, a construction site outside. Gradual usually means growth: more staff, more traffic, the same connection.
Call quality describes how faithfully your voice reaches the other person during a working call. It is a separate problem from an outage. If your phones show no service at all, this is not your checklist: see What Happens to Your VoIP Phone System If the Internet Goes Down?

Step 1: Reproduce the Problem on Purpose

You cannot diagnose a fault you cannot repeat. Make it happen deliberately before you change anything.
Call a colleague on a different network, ideally a mobile, and stay on the line for at least two minutes. Say a long sentence and let them read it back. Ask them to describe what they hear in their own words: cutting out, echo, delay, robotic, muffled, or silence. Their description matters more than yours, because half of all quality faults are only audible in one direction.
Write down the time, the number called, the device used, and both descriptions. You will need this in step 5.
🔁 If you cannot reproduce it
Intermittent faults are usually load-related. Try again at the busiest hour of your day, when the most people are on calls and the most data is moving. If the fault only appears then, treat it as a bandwidth or prioritisation problem and go straight to step 4.

Step 2: Run the Three Baseline Tests

These take ten minutes and rule out the majority of causes.
Test A: swap the headset. Move the affected person to a different headset, or to the handset on the desk phone. If the fault disappears, it was the headset. This is the single most common cause of echo and background noise, and the cheapest to fix.
Test B: swap the device. Have the same person make the same call from the mobile app on 4G or 5G, not on office Wi-Fi. If the call is clean, the problem is inside your building, not with your provider or your phone system.
Test C: swap the network path. Plug an affected softphone laptop into the network with an Ethernet cable instead of Wi-Fi. If the fault disappears, it is a Wi-Fi problem, not a bandwidth problem.
✅ Record what each test changed. Three results tell you which of the six problems below applies. Skip straight to the one that matches.

Step 3: Match the Symptom to the Problem

Find the description that matches what your colleague reported.

Problem 1. Echo: you hear your own voice back

Most likely cause: the far end's microphone is picking up their own speaker. Almost all echo originates at the other end of the call, not yours, even though only you can hear it.
What to check: speaker volume on the affected device, whether anyone is using a speakerphone in a hard-surfaced room, and whether a headset is sitting next to an open laptop speaker.
Also check the desk itself. Echo is not always acoustic. Phones, routers, power strips and computers stacked on top of one another create electrical interference that arrives as feedback on the line. Give the hardware some space before assuming the fault is in the network.
Fix order: lower the speaker volume, separate the equipment, switch from speakerphone to headset, then check the handset firmware. If echo appears only on calls to one specific destination, it is not your equipment: log it and move to step 5.

Problem 2. One-way audio: you hear them, they cannot hear you

Most likely cause: a firewall or router problem, not a hardware problem. Voice and call setup travel separately, and a firewall can allow one and block the other.
What to check: whether SIP ALG is switched on in your router. This feature is meant to help voice traffic and very often breaks it instead. Turn it off. Note that many routers switch it back on by themselves after a reset or a power cut, so check it again after any power interruption.
Fix order: disable SIP ALG, restart the router, retest. If it persists, your firewall needs the correct ports opened for your provider. Your provider will supply the exact list.
SIP ALG stands for Session Initiation Protocol Application Layer Gateway. It is a router feature that rewrites voice traffic as it passes through. On most consumer and small-business routers it does more harm than good, and disabling it is the standard first step in VoIP troubleshooting for systems of every size, from small local providers to the largest international platforms.

Problem 3. Robotic, choppy or metallic audio

Most likely cause: packet loss or jitter on your local network. Something is competing with voice traffic and winning.
What to check: what else is running when it happens. Large file uploads, cloud backups, video calls and software updates are the usual culprits. Backups scheduled during working hours are a frequent hidden cause.
Fix order: move backups outside working hours, then enable voice prioritisation on your router, then check whether your upload speed is the real constraint. Upload matters more than download for voice, and most business connections are advertised on their download figure.

Problem 4. Delay: you keep talking over each other

Most likely cause: latency, meaning the time your voice takes to arrive. Anything above roughly 150 milliseconds each way becomes noticeable in conversation.
What to check: whether the delay appears on all calls or only long-distance ones. Local calls with noticeable delay suggest a local network or VPN issue. Delay only on international calls suggests the route your provider uses.
Fix order: check whether affected staff are routing voice through a VPN, which adds a detour to every call. If not, this is a routing question for your provider, covered in the final section.

Problem 5. Calls drop mid-conversation

Most likely cause: for mobile app and softphone users, moving between Wi-Fi access points or between Wi-Fi and mobile data. For desk phones, a registration timeout or a failing cable or switch port.
What to check: whether drops happen at a consistent place in the building, or at a consistent time into the call. A consistent physical spot means Wi-Fi coverage. A consistent duration, such as always at the thirty-minute mark, means a session timeout on the router or firewall.
Fix order: map where drops happen, then address coverage or the timeout setting. Equipment options and costs are covered in How to Keep Your Cloud PBX Running When Things Go Wrong.

Problem 6. Muffled audio, low volume or background noise

Most likely cause: the headset or the room, not the network.
What to check: microphone position, headset age, and whether the affected desk sits near an air conditioning unit, an open-plan walkway or a window facing traffic.
Fix order: reposition the microphone two to three centimetres from the corner of the mouth, then test a headset with active noise cancellation. Whether desk hardware still earns its place at all is a separate question, covered in Do Businesses Still Need Desk Phones in 2026?

Step 4: Fix What Is Inside Your Control

If the symptom paths above did not resolve it, work through these five in order. All are free or low cost, and all are reversible.
Check for a double NAT. If your internet provider's box and your own router are both doing routing work, your traffic passes through two translation steps instead of one. This happens whenever an ISP modem sits in front of a separate business router or a mesh Wi-Fi unit. It degrades voice quality and makes every other fix harder to judge. Put the ISP box into bridge or modem-only mode so that one device handles routing.
Prioritise voice on your router. Quality of Service settings tell your router to let voice traffic go first when the connection is busy. Most business routers support it. Your provider can supply the exact port numbers to prioritise.
Check your upload capacity honestly. Each concurrent call needs roughly 100 kilobits per second of upload in each direction. Ten simultaneous calls need about one megabit of dedicated upload, plus headroom for everything else your business does at the same time.
Put desk phones on cable. Wi-Fi is convenient and adds a variable that wired connections do not have. Desk phones that never move should never be on Wi-Fi.
Update handset firmware. Out-of-date firmware causes echo, registration drops and codec mismatches. If your provider ships pre-provisioned handsets, this happens automatically. If you bought phones yourself, it does not.
⚠️ Change one thing at a time and retest after each. Changing three settings at once means you will never know which one mattered, and you will not be able to explain it to your provider if it comes back.

Step 5: How to Report It to Your Provider Properly

"The line is bad" produces a slow ticket. Specific data produces a fast one. Send this.
📋 Copy this escalation template
Subject: Call quality issue, [company name], [account or main number]
Symptom: [echo / one-way audio / robotic audio / delay / dropped calls / low volume], reported by the [caller / called party / both].
Scope: affects [one user / one department / all users]. Affects [all calls / calls to Belgium and France only / calls to mobile numbers only / internal calls only].
Started: [date]. [Sudden / gradual]. Changes made around that time: [new hardware, new internet line, office move, new software, none known].
Example calls:
  • [date and time], from [extension or number] to [number], duration [x], symptom: [description]
  • [date and time], from [extension or number] to [number], duration [x], symptom: [description]
Tests already run:
  • Different headset: [fault persists / fault disappears]
  • Mobile app on 4G instead of office Wi-Fi: [fault persists / fault disappears]
  • Wired Ethernet instead of Wi-Fi: [fault persists / fault disappears]
  • SIP ALG on router: [disabled / enabled / router does not have this setting]
  • Double NAT checked: [single router confirmed / ISP box in bridge mode / two routers present]
  • QoS for voice traffic: [enabled / not enabled / not supported]
Our setup: internet provider [name], connection type [fibre / cable / DSL], advertised speed [download / upload], router model [model], handsets [model], softphone or mobile app in use [yes / no].
What we are asking for: please check quality metrics on your side for the example calls above, confirm whether packet loss or jitter was recorded, and confirm the routing used for the affected destinations.
The last paragraph is the important one. It asks for something specific that only the provider can see, which makes it harder for the ticket to come back as a generic instruction to restart your router.

Step 6: Make the Fix Permanent

A resolved fault that nobody documented will return.
Record what the cause was and what fixed it, in one line, somewhere your team can find it. Schedule bandwidth-heavy tasks outside working hours as standing policy. Recheck SIP ALG after any power cut, since many routers silently switch it back on. Retest quality after any change to your network, your internet line or your office layout. Repeat the three baseline tests from step 2 once a quarter, which takes ten minutes and catches gradual decline before staff start complaining.
📊 What good looks like, in numbers
Ask your provider for these figures for your own account rather than accepting general assurances. Packet loss should sit below 1 percent. Jitter should stay under 30 milliseconds. One-way latency should stay under 150 milliseconds. These thresholds are consistent across the industry and are not provider-specific. A provider that cannot produce these numbers for a specific call you name is a provider that cannot diagnose your problem.

When the Problem Is Not on Your Side

If you have worked through every step and quality is still poor, the cause sits in your provider's network or in the route your calls take after they leave your building. That is outside your control, but it is not outside your influence, because it is a contractual matter rather than a technical one.
These are the questions worth putting to any provider, current or prospective.
Can you show me quality data for a specific call? Providers with proper monitoring can produce packet loss, jitter and latency figures for a call you name by time and number. Providers who cannot are guessing alongside you.
Do you use least-cost routing? Cheaper routes pass through more intermediaries, and each one adds a chance of delay or loss. This is the usual explanation when quality is fine domestically but poor to one specific country or mobile network. The mechanics are explained in Why Your Internet Connection Makes or Breaks Your Cloud PBX.
Where is my voice traffic actually processed? Every cloud phone system handles your audio at a physical location. The closer that location is to your office, the shorter the path your voice travels and the lower the delay. Platforms operate from different points of presence around the world, and a Luxembourg business connected to one far away carries a delay that no local change can remove. Ask where your calls are handled and whether a location in or near Luxembourg is available. This is the one quality factor you cannot fix from your own office.
Who is accountable when voice and internet come from different companies? This is the practical question behind a common piece of advice. Buying telephony and connectivity from one supplier gives you a single party to hold responsible, which shortens diagnosis considerably. Buying them separately keeps your options open and avoids one contract dictating the other. Neither is automatically right.
In Luxembourg you will find both patterns. POST Luxembourg, Cegecom, Visual Online and Mixvoip each sell connectivity and business telephony, so a single-supplier setup is available. Tango's Fix4Bizz is contractually tied to a Tango connection, which is worth knowing before signing rather than after. Bundling is not a weakness in itself; the point is to understand what you are agreeing to and what changing your mind later would cost. A side-by-side view of who offers what is in the provider comparison.
Do you supply pre-provisioned handsets, and do you manage the router? Phones configured and updated by the provider remove firmware mismatches. Where the provider also manages the router, the SIP ALG problem in Problem 2 largely stops recurring, because somebody owns the setting. Both are service-model questions rather than technology questions, and they cut across company size.
What does your SLA actually promise about quality? Most uptime guarantees cover whether the service is reachable, not whether it sounds acceptable. Ask what happens when calls connect but are unusable, and whether any compensation is automatic or requires you to claim it.
If you are collecting these answers as part of a wider evaluation, the Cloud PBX Buying Checklist and the Phone System Requirements Template turn them into something you can score across several quotes. The full range of capabilities to expect is set out in the features guide.
📚 Part of the connectivity and quality cluster on pbx.lu. Read next: Why Your Internet Connection Makes or Breaks Your Cloud PBX for what causes poor quality, or How to Keep Your Cloud PBX Running When Things Go Wrong for the equipment and costs of preventing it.
📅 Ready to explore Cloud PBX for your business?
A free 20-minute consultation, independent of any provider.
Book FREE consultation | Compare providers | Explore features