Voice Over Satellite — Making It Work for Professional Radio Users

For radio network designers, IT managers, and operations leads deploying satellite-linked voice communications.

Voice quality over satellite has a reputation problem. Anyone who has experienced the half-second lag of a GEO satellite phone call, the clipping and artefacts that characterise overloaded VSAT links, or the frustration of double-talking because neither party can hear the other is speaking, carries that experience as a reference point. And it is a fair one; poor satellite voice quality is a real and longstanding problem in professional communications.

But the reference point is increasingly outdated. LEO satellite technology, combined with modern voice codec design and quality-of-service engineering, is capable of delivering satellite voice communications that are functionally indistinguishable from a well-engineered terrestrial radio link. Understanding how to achieve this and what the engineering trade-offs involve is essential for anyone designing or procuring satellite-linked professional radio systems.

Why GEO Voice Was Difficult

The voice quality problems associated with traditional satellite communications were largely a consequence of GEO orbital geometry. A round-trip signal delay of 560 milliseconds or more is at the outer edge of what human conversation can accommodate. Beyond approximately 300 milliseconds of end-to-end delay, conversational flow degrades, users begin to talk over each other, and the cognitive effort of managing the delay reduces the effectiveness of the communication.

Echo is a compounding problem. At high latency, even small amounts of acoustic echo from handset speakers, room acoustics, or interface impedance mismatch become highly perceptible. GEO satellite systems require aggressive echo cancellation to manage this, and when the echo cancellation fails or is misconfigured, the result is the characteristic degraded audio that has defined satellite voice for many users.

What LEO Changes

LEO satellite latency in the 20-to-60-millisecond range changes the voice engineering problem fundamentally. At 40 milliseconds of one-way delay, the round-trip is 80 milliseconds — well within the 150-millisecond threshold that ITU-T G.114 defines as acceptable for conversational voice. Echo at this latency is manageable with standard echo cancellation without the aggressive suppression that was required for GEO. And double-talking, the key operational problem in professional radio is no longer a latency-driven pathology.

The result is that LEO satellite voice, when properly engineered, can meet the same quality standards as a well-designed IP radio network. This is not marketing language; it is a direct consequence of the physics of a much shorter signal path.

Codec Selection and Its Impact

For professional radio applications, codec selection is one of the most important voice quality decisions. A codec is the algorithm that converts analogue voice into digital data for transmission and back again. Different codecs make different trade-offs between bandwidth consumption, computational complexity, and voice quality.

For satellite links where bandwidth may be constrained, low-bitrate codecs such as AMBE+2 (used in DMR and P25 digital radio) and OPUS (widely used in modern IP communications) are well-suited. OPUS in particular is an adaptive codec that can vary its bitrate based on available bandwidth while maintaining good voice quality — a valuable property on satellite links that may experience variable throughput. G.711, the uncompressed PCM codec widely used in enterprise telephony, consumes 64 kbps per channel and is less efficient on satellite capacity, though its audio quality is excellent when bandwidth is available.

For a satellite-linked radio system carrying multiple simultaneous voice channels, codec efficiency directly affects the number of channels that can be supported on a given satellite bandwidth allocation. This is a real engineering and commercial trade-off.

Quality of Service Engineering

On any shared IP link and most satellite backhaul links are shared at some level, voice traffic must be prioritised over data to maintain acceptable quality. This is achieved through Quality of Service (QoS) configuration at the router level, using mechanisms such as DSCP marking to identify voice packets and scheduling algorithms that ensure voice traffic is transmitted with minimal queuing delay.

For satellite-linked radio networks, QoS engineering is not optional; it is the difference between a system that works and one that delivers acceptable voice quality only when the data load is low. A properly configured QoS policy will protect voice traffic against the throughput variability that satellite links can exhibit, particularly in congested or weather-affected conditions.

Practical System Design Recommendations

For anyone designing a satellite-linked professional radio voice system, the following principles apply. First, select a LEO satellite service rather than GEO wherever real-time voice is a requirement; the latency difference is not an edge case, it is foundational. Second, specify codec requirements explicitly in any procurement; do not assume that a satellite service provider’s default voice configuration is optimised for professional radio applications. Third, implement and test QoS configuration before operational deployment, simulating high data load while monitoring voice quality metrics, which will reveal problems before they affect operations. Fourth, plan for jitter buffering satellite links that can exhibit burst latency variation that requires a jitter buffer at the receiving end to smooth packet arrival timing. The buffer size is a trade-off between added delay and audio continuity.

Finally, test under realistic operational conditions. Office-environment voice testing does not replicate the acoustic conditions of a site cabin, a vehicle cab, or an outdoor environment where professional radios are actually used. Test where it matters.

The Achievable Standard

A satellite-linked professional radio network, designed to the principles above, can deliver voice quality that meets or exceeds the Mean Opinion Score thresholds for professional-grade communications. This is not a theoretical claim — it is demonstrated by operational deployments across the energy, logistics, and public safety sectors. The question for businesses evaluating satellite voice is not whether acceptable quality is achievable. It is whether they are working with partners who know how to achieve it.

https://www.linkedin.com/company/yesway-communications