MOS Score: Measure & Improve Voice Call Quality

Have you ever been on a call where the audio cuts out or the voice on the other end sounds robotic?

It’s frustrating, and for businesses, it can be a disaster. To keep conversations clear, engineers need a reliable way to measure how good or bad a connection actually is.

That’s where the Mean Opinion Score (MOS) comes in.

It’s the industry standard for measuring call quality and ensuring your phone system performs when it matters most.

Let’s explore MOS in detail.

What is a MOS Score?

A Mean Opinion Score (MOS) is a numerical rating used to measure the perceived quality of voice calls and video meetings.

Ranging on a scale from 1 (unacceptable) to 5 (excellent), a MOS score gives IT teams a quick way to benchmark communication quality.

While originally based on subjective human feedback, modern MOS values are often calculated using software algorithms that analyze network health.

It specifically examines packet loss, jitter, and latency to predict the human experience without requiring actual listeners for every call.

To interpret MOS correctly, you first need to understand how the scoring scale works.

MOS Score Scale Explained

To really get a grasp on voice quality, you have to look at the numbers.

The telecom industry relies on a standard 1 to 5 rating scale (technically known as an absolute category rating). It is actually pretty intuitive; think of it like giving a star rating to your Uber driver or a restaurant.

Here is what those numbers actually feel like when you are on a call:

Basically, if your MOS value drops below 3.6, your team usually starts filing support tickets.

In practice, most businesses aim to stay comfortably above the “Fair” range to avoid noticeable call quality issues. The score of 4.0 to 4.5 is considered the ideal.

Now that the scale is clear, the next question is how this score is actually calculated.

How is MOS Score Calculated?

You might wonder how a computer puts a number on something as personal as “how good does this sound?” It turns out, there are two very different ways to get that number: asking real people or asking a machine.

1. Subjective Method

Originally, getting a MOS rating was a purely manual process, literally a subjective experiment. Companies would gather a panel of listeners or observers in a quiet, controlled room.

These people would listen to various audio clips processed through different connection types and give their individual ratings based on a 1-5 scale.

This method is still considered the most accurate representation of the human experience because it measures genuine perceived quality.

However, it has a massive downside: it is slow, expensive, and impossible to do in real-time. You can’t have a guy named Steve listening to every one of your customer support calls just to grade them.

To overcome these limitations, objective methods were introduced.

2. Objective Method

To solve the speed problem, engineers developed objective measurement methods. Instead of using human ears, MOS is calculated using complex algorithms that analyze the network data itself.

Software looks at technical factors like how much audio data is arriving late or getting lost. Then, using standards like PESQ (Perceptual Evaluation of Speech Quality) or POLQA, the system predicts what a human would likely rate that call.

This allows IT teams to monitor business communication quality automatically across thousands of voice calls without anyone needing to listen in.

Once MOS is calculated, several technical factors determine whether the score stays high or drops.

What Factors Affect MOS Score the Most?

When call quality drops, it is rarely just bad luck. Voice calls are divided into thousands of small digital fragments, which are transmitted and then patched together in milliseconds all over the internet.

The following are the key things that come in the way:

• Packet loss: When voice packets fail to reach the other end, you hear gaps or choppy audio. A single packet loss can significantly reduce the call quality.
• Latency: A large delay leads to clumsy pauses or talking over, which makes conversations more difficult to comprehend.
• Jitter: Due to uneven packet delivery, it causes choppy and distorted audio.
• Codec selection: Codecs compress audio to save bandwidth. Low-bitrate codecs spoil the quality, leaving the maximum posible MOS score.
• Bandwidth availability: Low bandwidth may slow down or cut audio packets, particularly when making several simultaneous calls or downloads.
• Hardware and Endpoints: Poor microphones, headsets, or slow devices can reduce perceived quality, no matter how stable the network is.

Once packet loss reaches 1 to 2 %, it may not affect Quality of Experience (QoE), but exceeding 5% makes conversations unintelligible. When MOS drops, checking these factors systematically helps identify the root cause faster.

Things to Check When Your MOS Score Drops

If your MOS value drops, it’s a sign that your voice calls or video meetings aren’t delivering the expected quality of experience.

Here’s a user-friendly guide to quickly identify and fix common issues affecting audio and video quality.

1. Check Your Network for Latency and Packet Loss

• High latency and packet loss are the most common reasons MOS decreases.
• Use monitoring tools to measure delays and packet drops during voice calls or video sessions.
• Even minor packet loss can noticeably degrade audio and video quality.

2. Review Codec and Audio/Video Settings

• Audio codecs and video codecs affect perceived quality.
• Ensure your phone system or conferencing software is using industry-standard codecs optimized for both audio and video quality.
• Misconfigured codecs can lower your MOS score even on a strong network.

3. Evaluate Hardware and Endpoints

• Poor headsets, microphones, or webcams can reduce voice quality and video quality.
• Test your session quality on multiple devices to isolate the weakest point.
• Upgrading hardware can improve MOS value without touching the network.

4. Verify Bandwidth and Network Stability

• Multiple downloads, cloud syncs, or heavy video streams can reduce communications quality.
• Consider dedicated network paths or QoS (Quality of Service) settings for voice calls and video meetings.

5. Examine Call Routing and Telecommunication Settings

• Long-distance or inefficient call routing can increase latency and packet reordering.
• Review telecommunication standards and network effects to ensure optimal telephone transmission quality.

6. Review Subjective and Objective Measurement Data

• If you’re using a subjective experiment or objective measurement methods to monitor MOS, check the individual ratings or software algorithms for anomalies.
• Compare audio and video quality scores to see if one modality is disproportionately affecting the Mean Opinion Score.

7. Conduct Regular On-Demand Demos and Tests

• Run on-demand demos to simulate voice calls and video meetings under real-world conditions.
• Track changes in MOS using your rating scale and note improvements or regressions.

While MOS is powerful, it’s not the only metric used to evaluate call quality. Discover how call routing improves communications quality. *Industry standards recommend one-way latency below 150ms to maintain natural, conversational VoIP call quality.

MOS Score vs Other Call Quality Metrics

MOS provides a high-level of perceived quality, but other metrics explain why quality changes.

Using MOS alongside technical metrics gives a clearer picture of both user experience and network health. To act on these insights, MOS needs to be measured continuously and in real time.

How to Measure MOS Score in Real Time?

Waiting for a user to complain about voice calls is the worst way to monitor quality. To stay ahead of issues, IT teams need to measure MOS values in real time. Here is the step-by-step process for getting real-time data.

1. Choose a calculation method:

You generally have two options here. Passive monitoring watches actual, live voice calls flowing through your network to spot problems as they occur.

Active monitoring, on the other hand, injects synthetic traffic, basically “fake” calls, into the network to stress-test the connection.

Active monitoring is great for troubleshooting when no one is on the phone, while passive gives you real human experience data.

2. Identify the core metrics:

To calculate a score without asking people for their opinion, your software needs raw data. It tracks latency (delay), jitter (variance), and packet loss. It also checks which audio codecs are being used, as a high-compression codec naturally lowers the maximum possible score.

3. Apply a computational model:

Once the data is collected, a mathematical formula takes over. Most systems first calculate the R-factor (a transmission rating from 0 to 100 based on the E-model). The software then uses a standard conversion table to translate that R-factor into the familiar 1-5 MOS value.

This allows you to turn technical network statistics into a simple quality grade.

4. Implementation tools:

You don’t do this math with a calculator. You use Network Performance Monitors (NPMs) or SD-WAN solutions that include these features.

Many modern tools and even open-source stacks can automatically display the MOS score on your dashboard, alerting you the moment audio quality begins to drop.

Therefore, real-time measurement turns raw network data into actionable quality insights.

Even with monitoring in place, certain issues consistently cause MOS scores to drop.

Common MOS Score Problems

Recurring MOS issues usually stem from technical limitations, configuration choices, or measurement constraints.

1. Network Performance Issues:

Network instability can still impact MOS, with factors like jitter, latency, and packet loss being key contributors. A jitter level >150ms is considered unacceptable as it degrades the call quality completely. The ideal level is below 30 ms.

Regularly monitor your network and bandwidth to maintain optimal communications quality.

2. Codec and Technical Factors:

Poor Codec Selection: Heavy compression can limit audio quality, lowering MOS even on a strong network.

Echo: Signal leakage between the mic and the speaker can be very distracting.

Packetization Delay: Slow processing of audio into packets adds lag to calls and video sessions.

3. Hardware and Endpoint Issues:

Low-quality Headsets/Microphones: No network tweaks can fix a bad source signal.

Unstable Connections: Spotty Wi-Fi or overloaded networks can reduce session quality.

Resource-intensive Software: PCs or phones running at full capacity may not process audio or video fast enough.

4. Limitations in Measuring MOS:

Subjectivity/Bias: Individual listener preferences can skew mean opinion scores in a subjective experiment.

Average Issues: A call may be perfect most of the time, but a brief dropout can be hidden in the average MOS.

Perfect 5.0 is Unrealistic: Due to compression and digital transmission, most systems rarely exceed 4.5.

Recognizing these common problems helps teams respond faster and more accurately.

The good news is that most MOS issues can be fixed with targeted improvements.

How to Improve MOS Score for VoIP Calls?

Knowing your MOS score is low is one thing; fixing it is another. Since voice calls and video meetings happen in real-time, they are incredibly sensitive to network hiccups.

To get your Mean Opinion Score back into the “Good” range (4.0+), you need to tackle the problem from three angles: the network, the hardware, and the software settings.

A. Optimize network configuration

Think of your network like a highway. If everyone is driving in the same lane, traffic jams happen.

• Prioritize Traffic with QoS: Quality of Service (QoS) is your traffic cop. It sets rules so that voice calls get the “VIP lane”, ensuring they don’t get stuck behind a large file download. This drastically reduces latency and packet loss.
• Use Dedicated VLANs: Segregate your VoIP traffic onto its own Virtual LAN. This keeps your phone system separate from regular computer data, minimizing congestion and network effects.
• Disable SIP ALG: Many routers come with a setting called SIP ALG enabled by default. While it is supposed to help, it often breaks VoIP calls by modifying data packets incorrectly. Turning this off is often the quickest fix for one-way audio or dropped calls.

B. Stabilize hardware and connectivity

Sometimes the issue isn’t the internet line, it’s the gear at the end of it.

• Switch to Wired Ethernet: Wi-Fi is convenient, but it is prone to interference. For the best MOS value, plug your desk phones and computers directly into the wall. A wired connection eliminates the random jitter that kills audio quality.
• Use Noise-Cancelling Headsets: If you’re using a cheap built-in laptop mic, your perceived quality will be low, no matter how fast your internet is. Professional headsets isolate the voice, improving the human experience on both ends.
• Update Firmware: Outdated router or phone firmware can introduce bugs that cause degradation. Keep your endpoints updated to ensure they handle audio codecs correctly.

C. Adjust Technical parameters

Tweak the software settings to match your available bandwidth.

• Choose the Right Codec: Audio codecs determine how your voice is packaged.

• G.711: It is the industry standard for uncompressed, high-quality audio. Use this if you have plenty of bandwidth.
• G.729: This is highly compressed. Use this if bandwidth is tight; it lowers the maximum MOS slightly but prevents dropped calls.
• Opus: It’s a modern, adaptive codec that adjusts audio and video quality dynamically based on connection speed.

• Implement Jitter Buffers: A jitter buffer intentionally delays incoming audio packets by a few milliseconds to ensure they are played in the correct order. It acts like a shock absorber for your voice quality, smoothing out the bumps in a choppy connection.

D. Active Monitoring

You can’t fix what you don’t watch.

• Track Key Metrics: Don’t just look at the final MOS. Keep an eye on latency, jitter, and packet loss individually.
• Use Monitoring Tools: Strict active monitoring tools can simulate calls 24/7. This lets you catch the communications quality issues before your users do, ensuring your quality of experience stays consistent.

When you apply these practices together, these improvements help maintain a consistently high MOS across voice calls and video meetings.

Conclusion

Ultimately, a MOS score isn’t just technical data; it is the heartbeat of your business communication.

If your voice calls are choppy or dropping out, you risk frustrating clients and losing revenue. By actively monitoring latency and packet loss, you can stop bad audio quality before it starts.

Don’t settle for “good though”; keep your communication quality sharp so every conversation counts.