VoLTE vs VoNR: How Voice Call Testing Differs Between LTE and 5G Core

VoLTE and VoNR testing differ mainly in the call-flow path being validated: EPS fallback and anchored LTE signaling for VoLTE versus native 5G SA voice bearers for VoNR. That architectural split changes which KPIs you measure, how call setup time and MOS get captured, and most critically, what network mode must actually be present for the test to be valid at all. Get the mode wrong and you will test VoLTE while believing you tested VoNR, and no amount of post-processing fixes that.

VoLTE vs VoNR: what's actually different under the hood

The real difference between VoLTE and VoNR comes down to which core network carries the call. VoLTE carries voice as IMS-based VoIP over the LTE Evolved Packet Core (EPC), using SIP signaling over a dedicated bearer tagged with QCI 1 to guarantee low latency and priority handling. VoNR does the same job but skips the EPC detour entirely: voice is carried natively over the 5G NR air interface and the 5G Core (5GC), with SIP signaling riding on 5G QoS flows instead of EPS bearers.

That sounds like a minor plumbing difference until you hit the NSA/SA wall. VoNR is only possible on a genuine 5G Standalone (SA) deployment; on Non-Standalone (NSA) 5G, voice calls fall back to LTE regardless of what the data session is doing. Section 5 covers exactly why that happens.

Why does a test engineer need to care this much? Because testing 'VoNR' on an NSA network will silently produce VoLTE results. The call connects, the audio sounds fine, the KPIs look normal, and none of it tells you anything about native 5G voice performance. Confirming network mode before the first test call is not optional.

Call flow and signaling: IMS registration, SIP, and EPS fallback vs native SA voice path

Call setup on VoLTE follows a predictable sequence: IMS registration over LTE, a SIP INVITE/180 Ringing/200 OK exchange, dedicated bearer establishment, and finally an RTP media path carrying the actual voice packets. Every one of those steps adds latency, and every one shows up in a drive test log as a discrete signaling event.

VoNR's call flow is leaner. IMS registration happens over 5G NR, SIP signaling is carried over NR QoS flows, and there's no bearer-to-EPS mapping step to wait for. Fewer moving parts, in theory, means faster call setup, assuming the UE actually stays on SA throughout.

What EPS fallback adds to the picture

EPS Fallback (EPSFB) happens when a UE sitting in 5G SA coverage initiates or receives a call, but the network decides to redirect it to LTE for the voice leg. This redirection inserts a measurable delay before SIP signaling even starts, since the UE has to complete an RRC reconfiguration to LTE first.

If you're logging a VoNR test campaign, capture the RRC reconfiguration or redirection message, timestamp the fallback trigger, and calculate the delta in call setup time versus a native SA voice call. Without that delta, you can't tell whether a slow call was a native VoNR call with a slow core, or a fallback call that never got the chance to be VoNR at all.

KPI breakdown: what you measure for VoLTE vs VoNR

The KPI names stay the same between VoLTE and VoNR testing. What changes is what each KPI is actually capturing underneath.

The practical consequence: you cannot directly compare raw CST or drop rate numbers across a VoLTE campaign and a VoNR campaign without noting, for every single call, the RAT, the NSA/SA mode, and the fallback status. Blend those together and your averages mean almost nothing.

How MOS and POLQA scoring actually works for each technology

POLQA (ITU-T P.863) is a perceptual audio quality algorithm. It compares a known reference signal against the received, degraded signal and outputs a MOS-equivalent score. This is the same algorithm used for both VoLTE and VoNR MOS testing, the methodology doesn't change between technologies, only the path the audio travels.

In practice, VoLTE/VoNR MOS testing means playing a known reference audio sample at one end of a call, capturing the received signal at the other end, and running both through POLQA post-processing. Because the test is end-to-end, it captures codec performance, transcoding artifacts, jitter buffer behavior, and any RAT-specific impairments, not just raw radio signal quality.

This is where engineers sometimes conflate two very different things. MOS is a service-level voice quality score. SINR and RSRQ are radio signal quality measurements. A cell can report excellent SINR and still produce a poor MOS score if there's jitter buffer underrun, RTP packet loss, or a codec mismatch somewhere in the chain. Good radio does not guarantee good voice.

Dingli's VoLTE/VoNR MOS Testing solution is built around this end-to-end measurement approach, giving mobile voice service providers a way to measure voice quality rather than just radio conditions.

5G NSA vs SA: why this decides whether you're testing VoNR at all

As noted earlier, an NSA network anchors its control plane to LTE, so voice always routes through VoLTE no matter which radio carries the data. There is no native VoNR path to test here, full stop.

On SA, the UE has a direct connection to the 5GC and can establish a native VoNR call. But EPS fallback can still happen depending on coverage gaps, network configuration, or UE capability limits. SA coverage is not a guarantee of a VoNR call.

Should I keep VoNR on or off?

This is a configuration and device-setting question for network planners, not a testing decision. From a validation standpoint, engineers should test both paths: the native VoNR path and the fallback path, because real subscribers will experience both depending on where they stand relative to SA coverage. Testing only one path gives you half the picture.

Before assuming any VoNR call in your logs was actually a VoNR call, confirm the network mode and the IMS/VoNR support flag on the device. Skipping this check is the single most common cause of mislabeled VoNR test results.

Testing methodology differences: drive test, lab test, and automated call generation

Drive testing for VoLTE is a well-established workflow: mobile-originated and mobile-terminated call generation along defined routes, with logging of RAT, cell ID, signaling, and RTP KPIs at every point. Dingli's LTE drive test solutions measure coverage, signal quality, throughput, handover and service experience along these defined test routes.

Drive testing for VoNR adds a layer of complexity. It requires SA-capable test devices and actual SA coverage along the route, plus explicit fallback detection logic. A route planned as a 'VoNR test route' may contain SA coverage holes that silently turn calls into VoLTE calls without any obvious signal in the summary report.

Lab validation before the drive test budget gets spent

Before committing to an expensive field campaign, voice call flows can be validated in controlled lab conditions to catch signaling or interoperability issues early. Dingli's Laboratory Automation solutions support verifying 4G/5G performance and reliability before deployment, which is the logical first gate before field testing.

For ongoing benchmarking as VoNR coverage expands, continuous call generation lets engineers track network behavior over time rather than relying on a single snapshot. Dingli's Autonomous Measurement solution supports this kind of continuous monitoring and validation of network performance over time.

Troubleshooting: diagnosing poor MOS, dropped VoNR calls, and fallback failures

• Poor MOS with good radio KPIs: look at jitter buffer behavior, RTP packet loss, and codec negotiation before blaming RSRP or SINR. The radio link can be clean while the voice path is not.
• Dropped VoNR calls: check whether the drop coincides with a handover out of SA coverage. That pattern points to a fallback failure, not a radio link failure, and the fix is entirely different.
• EPS fallback failures: the UE requests fallback but LTE redirection or bearer setup fails. In signaling logs this shows up as a stalled RRC reconfiguration before SIP signaling ever resumes.
• What are the disadvantages of VoLTE? Compared to VoNR, VoLTE's bearer-based QoS model and its reliance on EPS fallback when migrating from 5G coverage can introduce extra call setup delay. That's a known architectural trade-off, and engineers should account for it when setting CST benchmarks rather than expecting VoLTE and VoNR numbers to converge.
• Segment first, diagnose second: tag every anomaly by RAT and fallback status before concluding it's a radio problem versus a core or IMS problem. Skipping this step is how teams waste weeks chasing the wrong layer.

VoLTE vs VoNR test parameters at a glance

Use this as a quick reference when scoping a test plan or briefing a field team on what to expect from each technology.

If you only remember one row, make it the equipment requirement row. A test device that can't confirm SA registration and VoNR support will never tell you the difference between a native VoNR call and a fallback call.

Building a voice test strategy that covers both technologies

As networks transition from LTE to 5G SA, test campaigns should expect a realistic mix of native VoNR calls, native VoLTE calls, and fallback calls, all in the same route or the same day. Report results segmented by call type rather than as one blended average, otherwise the numbers will flatten out real problems.

The fullest picture comes from layering three stages: pre-launch lab validation to catch signaling and interop issues early, drive testing to capture real-world RF and fallback behavior, and ongoing automated monitoring to track how VoNR coverage and performance evolve as SA footprint expands.

Dingli's 5G Private Network Testing and Multiple Network Benchmarking solutions, alongside VoLTE/VoNR MOS Testing, LTE/5G drive test tools, and autonomous measurement products, give engineering teams the components needed to validate voice quality across both network generations rather than just one.

Engineers planning a VoLTE/VoNR validation campaign can review the relevant testing solutions and product documentation on Dingli's support portal before scoping equipment and test routes.

Frequently asked questions

Should I keep VoNR on or off?

This is a device/network configuration decision, not a testing choice. For validation purposes, engineers should test both the native VoNR path and the EPS fallback path, since real subscribers will experience both depending on their position relative to SA coverage at any given moment.

What are the disadvantages of VoLTE?

VoLTE's bearer-based QoS model and its role as the fallback target when 5G SA voice isn't available can add extra call setup delay compared to native VoNR. This is an architectural trade-off engineers should factor into call setup time benchmarks rather than expecting parity with VoNR.

Which is better, VoLTE or LTE?

This compares two different things. LTE is the radio access network; VoLTE is the voice service that runs over it via IMS and SIP signaling instead of the legacy circuit-switched fallback. VoLTE is the modern standard for voice on LTE networks, not a competing alternative to LTE itself.