Understanding the VoNR (Voice over New Radio) and its benefits

VoNR (Voice over New Radio), also known as Vo5G or Voice over 5G Standalone (Vo5GSA), is the native voice calling solution for 5G networks. It delivers voice (and often video/SMS) services entirely over the 5G New Radio (NR) access network and 5G Core (5GC), using IP Multimedia Subsystem (IMS) for session control, just as VoLTE does for 4G LTE.

Unlike earlier generations where voice used dedicated circuit-switched channels (2G/3G) or fell back to them, VoNR treats voice as prioritized Voice over IP (VoIP) packets carried over the packet-switched 5G infrastructure. This enables true end-to-end 5G connectivity without relying on legacy 4G elements for voice anchoring.

How VoNR Works: Architecture and Key Components

VoNR requires a 5G Standalone (SA) deployment:

5G RAN (gNodeB) handles the radio access.
5G Core (5GC) manages mobility, session, and policy via functions like AMF (Access and Mobility Management Function), SMF (Session Management Function), and UPF (User Plane Function).
IMS provides call control using SIP (Session Initiation Protocol) signaling. A dedicated QoS Flow (typically with 5QI = 1 for voice media and 5QI = 5 for signaling) ensures low-latency, guaranteed bit rate (GBR) treatment.
Voice packets use advanced codecs like EVS (Enhanced Voice Services) for superior audio quality, with options for wideband, super-wideband, or full-band audio. Video uses H.265 (and future enhancements).

Call Flow Overview (simplified):

UE registers with 5GC and IMS.
For an outgoing call, SIP signaling establishes the session.
A PDU Session is set up or modified with dedicated QoS flows for voice.
Media flows directly over 5G NR → UPF → IMS (or peer network).
Features like semi-persistent scheduling (SPS), configured grant, C-DRX, and TTI bundling optimize radio resources for voice efficiency and coverage.

This contrasts with EPS Fallback (a common transitional mechanism in early 5G SA), where a voice call attempt triggers the device to redirect or hand over to 4G LTE, where VoLTE handles the call. EPS Fallback allows quick 5G SA launches but temporarily disables high-speed 5G data during the call and requires tight 5G-LTE interworking (often via N26 interface).

VoNR vs. Related Technologies:

VoLTE (Voice over LTE): Uses 4G RAN + EPC + IMS. Mature, widely deployed, but limited by 4G latency and capacity. During VoLTE calls on non-SA 5G (NSA/EN-DC), data may use 5G while voice stays on 4G.
VoWiFi (Voice over Wi-Fi): Routes calls over any Wi-Fi via ePDG for secure tunneling to IMS. No cellular QoS guarantee; quality depends on Wi-Fi stability.
VoNR Advantages: Simultaneous high-speed 5G data + voice (no radio switch), lower end-to-end latency (faster call setup, ~0.25–2.5s vs. longer in legacy), better spectral efficiency via 5G features (beamforming, massive MIMO), and future-proofing for immersive services.

VoNR vs. VoLTE Comparison (key differences):

Aspect	VoLTE (4G)	VoNR (5G SA)
Network	LTE RAN + EPC + IMS	5G NR + 5GC + IMS
Latency / Call Setup	Low (but higher than 5G)	Ultra-low; faster setup and responsiveness
Voice Quality	HD (AMR-WB, EVS optional)	Superior (EVS mandatory; richer audio)
Data During Call	Often throttled or on 4G	Full 5G speeds simultaneously
Coverage Dependency	Mature LTE coverage	Requires robust 5G coverage (fallback to VoLTE common initially)
Fallback/Continuity	SRVCC to 2G/3G possible	Handover to VoLTE; eSRVCC (Rel-16) to 3G
Device/Radio Impact	Single radio often sufficient	Optimized for 5G but may increase power in weak coverage

Benefits and Implications

User Experience: Crystal-clear calls with natural sound, reduced background noise, faster connections, and seamless multitasking (e.g., streaming or using maps during a call). Lower latency improves conversational flow and supports emerging uses like real-time translation or AR-enhanced calls.
Network Efficiency: Better spectrum utilization, higher capacity for voice traffic, and reduced OPEX by eventually phasing out legacy circuit-switched or even full VoLTE reliance in mature SA networks.
Operator Perspective: Enables full 5G monetization (e.g., premium “5G Calling” features, network slicing for prioritized voice). Supports Rich Communication Services (RCS), video calling (ViNR), and integration with IoT/wearables.
Broader Ecosystem: Paves the way for immersive audio (e.g., IVAS in later releases), better emergency services (E911 with 5G location), and industrial applications requiring reliable low-latency voice.

Nuances: In weak 5G coverage, VoNR may still hand over to VoLTE for continuity. Power consumption can vary—5G voice may drain more battery in fringe areas due to higher-frequency bands, though optimizations like DRX help.

Deployment Status

5G SA deployment has accelerated significantly. Global 5G SA availability (sample share) reached ~17.6% in late 2025, with leaders in Asia-Pacific (China, India), the US (T-Mobile mature; AT&T/Verizon launching), GCC countries, and parts of Europe. At least 89 operators had launched public 5G SA services, with 181 investing across 73 countries. VoNR often follows SA rollout, sometimes starting in limited areas or alongside EPS Fallback.

Pioneers like T-Mobile (US) have widely deployed VoNR with uplink CA and low-band coverage.
Many operators still use hybrid approaches (EPS Fallback for reliability during transition).
Device support is growing but not universal—requires compatible chipsets, firmware, and IMS profiles. Major flagships increasingly enable it via updates.

By mid-2026, VoNR is transitioning from early/selective deployment to broader commercial use in mature SA markets, though full global coverage remains phased due to spectrum, infrastructure, and device ecosystem factors.

Challenges, Edge Cases, and Considerations

Coverage: 5G (especially mid/high bands) has shorter range than low-band LTE. Operators need low-frequency spectrum (<1 GHz) for deep indoor/rural voice coverage; otherwise, frequent fallbacks occur.
Device and Ecosystem: Not all UEs support full VoNR (or do so efficiently). Early devices often default to EPS Fallback. Power optimization and certification are ongoing.
Interworking/Mobility: Seamless handover to VoLTE is critical; poor configuration can cause dropped calls or degraded QoS. Rel-15 lacks direct 5G-to-3G CS continuity (added/enhanced in Rel-16).
QoS and Capacity: Voice requires dedicated GBR flows; high data loads or interference can impact quality. Operators must optimize scheduling, beam management, and admission control.
Regulatory/Emergency: Must support E911 with accurate 5G location; lawful interception and roaming interworking add complexity.
Migration Path: VoLTE remains a prerequisite for many SA launches. Long-term goal is VoNR dominance to simplify networks and decommission legacy elements.
Edge Cases: Indoor scenarios (favor Wi-Fi calling fallback), international roaming (IMS interworking tests needed), high-mobility (handover robustness), or mixed NSA/SA environments.

Implications include accelerated legacy network sunsetting in advanced markets, new revenue from differentiated voice/video services, and preparation for 5G-Advanced/6G features like immersive communications. However, uneven global rollout means hybrid VoLTE/VoNR experiences will persist for years.

VoNR: Architecture and Key Components

VoNR (Voice over New Radio) operates as the fully native voice service in a 5G Standalone (SA) network. It carries both voice signaling and media entirely over the 5G New Radio (NR) air interface and the 5G Core (5GC), with the IP Multimedia Subsystem (IMS) handling call control, session management, and media negotiation. This architecture eliminates reliance on 4G LTE for voice anchoring during calls, unlike non-standalone (NSA) deployments or transitional EPS Fallback mechanisms.

The design leverages 5G’s service-based architecture (SBA), cloud-native modularity, ultra-reliable low-latency communication (URLLC) capabilities, and advanced QoS framework. This results in faster call setup (often under 1 second in good conditions), superior audio quality via modern codecs, simultaneous high-speed data during calls, and better spectral efficiency compared to VoLTE.

High-Level Architecture of VoNR

The end-to-end VoNR system comprises three main domains:

5G Radio Access Network (RAN):
- gNodeB (gNB): The 5G base station that terminates the NR air interface. It handles Radio Resource Control (RRC), Packet Data Convergence Protocol (PDCP), Radio Link Control (RLC), Medium Access Control (MAC), and Physical (PHY) layers.
- Establishes Data Radio Bearers (DRBs) mapped to QoS Flows from the core.
- Supports voice-specific optimizations such as semi-persistent scheduling (SPS), configured grants (CG), Connected-mode Discontinuous Reception (C-DRX), TTI bundling, and robust header compression (RoHC) to minimize overhead and latency for small, periodic voice packets.
5G Core Network (5GC):
- A modular, service-based architecture with control-plane/user-plane separation.
- Key Network Functions (NFs) involved in VoNR:
  - AMF (Access and Mobility Management Function): Manages UE registration, authentication, mobility, and NAS (Non-Access Stratum) signaling. It coordinates between RAN and IMS, forwards SIP signaling, and handles reachability.
  - SMF (Session Management Function): Controls PDU Sessions, IP address allocation, and QoS enforcement. It interacts with PCF for policy decisions and triggers PDU Session Modification to establish dedicated QoS Flows for voice media and signaling.
  - UPF (User Plane Function): Routes user-plane traffic (RTP/RTCP media streams) between gNB and the IMS (or external networks). It performs packet inspection, QoS enforcement, and anchoring.
  - PCF (Policy Control Function): Provides policy rules for QoS, charging, and network slicing. It authorizes dedicated flows for voice based on IMS signaling (via Rx interface extensions).
  - UDM/UDR (Unified Data Management/Repository): Stores subscriber data and supports IMS interworking (extended from HSS in 4G).
- Other supporting elements include AUSF for authentication and NRF for NF discovery.
IMS (IP Multimedia Subsystem):
- The application-layer service platform (largely reused from VoLTE with 5G-specific extensions in 3GPP Rel-15 and beyond).
- Core elements: P-CSCF (Proxy-CSCF for SIP signaling entry point), I-CSCF, S-CSCF (Serving-CSCF for session control), and media gateways/handling functions.
- Uses SIP (Session Initiation Protocol) for signaling and SDP (Session Description Protocol) for media negotiation.
- Interfaces with 5GC via enhanced Rx (policy), Sh/Cx (subscriber data), and SIP headers carrying 5G-specific info (e.g., RAT type, location, charging ID).
- Supports features like T-ADS (Terminating Access Domain Selection) for routing decisions.

Media Plane: Voice is transported using RTP (Real-time Transport Protocol) / RTCP, typically with the EVS (Enhanced Voice Services) codec as the primary option. EVS supports narrowband to fullband (up to 20 kHz) audio, adaptive bitrates (5.9–128 kbps), channel-aware mode for packet loss resilience, and superior quality even in challenging radio conditions compared to AMR-WB. Video calls (ViNR) use codecs like H.265.

QoS Framework (Unique to 5G):

Voice uses QoS Flows within a PDU Session (instead of EPS bearers in 4G).
Standard 5QI values:
- 5QI = 5: Non-GBR, high-priority for SIP signaling (low latency, high reliability).
- 5QI = 1: GBR (Guaranteed Bit Rate) for conversational voice (priority ~20, packet delay budget ~100 ms, packet error rate 10⁻²).
- 5QI = 2: For video (if ViNR).
The gNB maps these QoS Flows to DRBs. Admission control and scheduling ensure prioritized treatment.

Network Slicing: Operators can use a dedicated slice for IMS voice (e.g., URLLC or eMBB with specific QoS profiles) for isolation and optimized performance.

Detailed VoNR Call Flow (Simplified End-to-End)

A typical Mobile-Originated (MO) VoNR call involves these high-level stages (similar to VoLTE but adapted to 5G procedures):

UE Registration and IMS Registration:
- UE camps on 5G NR cell, performs RRC Connection Setup with gNB.
- Registers with 5GC via AMF (5GMM procedures).
- Establishes a default PDU Session for IMS signaling (often “ims” DNN/APN).
- SIP REGISTER to P-CSCF (discovered via SMF or PCO), authenticated via AKA (using ISIM on UICC). Successful 200 OK completes IMS registration.
Call Initiation (SIP Signaling):
- UE sends SIP INVITE (with SDP offer including codec list, e.g., EVS) over the signaling QoS Flow (5QI=5).
- Signaling traverses gNB → UPF → P-CSCF → IMS core.
- IMS performs routing, authorization (via PCF), and alerts the terminating side (180 Ringing, etc.).
Resource Reservation and QoS Setup:
- During SIP negotiation (e.g., after 183 Session Progress), IMS triggers PCF → SMF to modify the PDU Session.
- SMF instructs UPF and gNB to establish a dedicated GBR QoS Flow (5QI=1) for RTP media.
- gNB configures DRB, performs RRC Reconfiguration with UE.
- UE confirms resource allocation (often via SIP PRACK/UPDATE for preconditions).
- Both sides must confirm radio resource reservation before ringing starts on the called party (to avoid “ghost ringing”).
Call Setup Completion:
- SIP 200 OK + ACK.
- RTP media flows bidirectionally over the dedicated QoS Flow.
- Call is active with full 5G data capability on the same or parallel PDU Sessions.
Call Termination:
- SIP BYE → resource release (QoS Flow deactivation, DRB release).

Mobile-Terminated (MT) calls follow a mirrored process, with paging via AMF and T-ADS for domain selection.

Key Optimizations in RAN for Voice Efficiency:

Semi-Persistent Scheduling (SPS) and Configured Grant: Pre-allocate resources for periodic voice packets, reducing PDCCH overhead.
RoHC: Compresses RTP/UDP/IP headers (from ~40+ bytes to a few bytes).
C-DRX: Allows UE to sleep between voice packets for battery savings.
Packet Loss Concealment and Jitter Buffer in EVS for robustness.
Beamforming and massive MIMO improve coverage and capacity for voice users.

VoNR vs. EPS Fallback: Transitional Context

Many early 5G SA deployments rely on EPS Fallback (or RAT Fallback) because full VoNR requires mature 5G coverage, device support, and network optimization:

In EPS Fallback, a voice attempt on 5G triggers redirection or inter-RAT handover to LTE (via N26 interface or without). The call then uses VoLTE on 4G, with data potentially throttled or switched.
VoNR avoids this radio switch, enabling true simultaneous 5G voice + data, lower latency (~20-50% faster setup in tests), and native 5G features.
Drawbacks of fallback: Increased setup delay, temporary loss of 5G data speeds, higher battery impact from dual-radio operation in some cases.
Handover from VoNR to VoLTE is supported for mobility (enhanced in Rel-16 with eSRVCC-like mechanisms to legacy if needed).

Nuances, Edge Cases, and Considerations

Coverage and Mobility: 5G mid/high-band cells have shorter range; low-band spectrum is critical for wide-area VoNR. Weak coverage triggers handover to VoLTE. High-mobility scenarios require robust measurement gaps and handover thresholds.
Power Consumption: Optimized features help, but 5G voice in fringe areas can consume more power than LTE due to higher frequencies and beam management. C-DRX and SPS mitigate this.
Device Requirements: UE must support VoNR (indicated in capabilities), proper IMS configuration, EVS codec, and 5G SA voice features. Not all devices enable it out-of-the-box; carrier provisioning and firmware updates are needed.
Roaming and Interworking: Requires tested IMS NNI (Network-Network Interface), consistent QoS mapping, and handling of 5G-specific SIP headers. Local Breakout vs. Home Routed adds complexity.
Emergency Calls (E911): Must support accurate 5G location (NR positioning) and fallback if needed. Regulatory compliance is mandatory.
Capacity and QoS Management: Voice GBR flows compete with data; operators use admission control, pre-emption, and slicing. High-load scenarios or interference can degrade quality if not tuned.
Video Calls (ViNR): Similar flow but with additional QoS for video (higher bandwidth).
SMS over 5G: Often uses SMS over NAS or IMS, independent of voice but sharing signaling paths.
Testing and KPIs: End-to-end latency, MOS (Mean Opinion Score) for voice quality, call setup success rate, drop rate, and resource utilization are key. Tools measure SIP/RTP, QoS Flow performance, and RAN metrics.
Future Enhancements: 5G-Advanced (Rel-18+) adds further optimizations like immersive voice (IVAS codec), better integration with network slicing, and AI-driven scheduling.

Implications:

For Users: More natural conversations, HD/super-HD audio, multitasking without quality trade-offs, and preparation for AR/VR-enhanced calls.
For Operators: Simplified network (eventual VoLTE decommissioning), higher capacity, new monetization via premium voice features or slices, but requires investment in SA infrastructure and optimization.
Ecosystem: Drives device chipset evolution (e.g., Snapdragon, Exynos, MediaTek with VoNR support) and IMS vendor alignment.

In practice, VoNR deployment is phased: initial use of EPS Fallback for reliability, transitioning to native VoNR as coverage densifies and ecosystems mature. Real-world performance depends on spectrum holdings, site density, backhaul, and configuration tuning.

VoNR Call Flow

Detailed VoNR Call Flow explores the end-to-end signaling and resource establishment for a native Voice over New Radio call in a 5G Standalone (SA) network. VoNR uses the IP Multimedia Subsystem (IMS) for session control, SIP for signaling, and 5G QoS Flows (instead of EPS bearers) for media transport. The flow leverages the service-based architecture of the 5G Core (5GC), with interactions among the UE, gNB, AMF, SMF, PCF, UPF, and IMS elements (primarily P-CSCF, I-CSCF, S-CSCF).

This is similar to VoLTE but adapted for 5G procedures: PDU Sessions replace PDN Connections, QoS Flows (identified by QFI) replace dedicated bearers, and interfaces use 5G-specific extensions (e.g., enhanced P-Access-Network-Info and P-Charging-Vector SIP headers carrying NR RAT type and 5G location info). The design prioritizes preconditions for resource reservation to ensure reliable voice quality before alerting.

VoNR supports both Mobile-Originated (MO) and Mobile-Terminated (MT) calls. The description below focuses primarily on an MO VoNR call (the most illustrative case), with notes on MT differences. Real-world flows include many more NG-AP, NAS, Diameter, and internal 5GC messages; the following provides a thorough, layered breakdown with key stages, messages, and rationale.

Prerequisites Before Any Call

A UE must complete these steps before initiating or receiving a VoNR call:

5G Cell Selection and RRC Connection:
- UE detects a suitable 5G NR cell (based on SIB1/SIB2 broadcast, including IMS voice support indicators).
- RRC Setup / Setup Complete (with NAS Registration Request, indicating IMS/VoNR capability via UE 5GMM capabilities and “IMS voice over PS session” preference).
5G Registration with 5GC:
- gNB forwards Initial UE Message (NAS) to AMF.
- AMF performs authentication/security (with AUSF/UDM), selects SMF if needed, and retrieves subscription data (including IMS voice support via UDM).
- AMF sends Registration Accept (with 5GS network feature support for IMS voice).
- UE sends Registration Complete.
- Outcome: 5GMM context established; UE is in RM-REGISTERED, CM-CONNECTED or IDLE state.
Default PDU Session Establishment for IMS Signaling:
- UE requests PDU Session Establishment (IMS DNN, requested 5QI=5 for signaling — non-GBR but high-priority).
- AMF selects SMF → SMF interacts with PCF for policy → UPF selection and N4 session setup.
- SMF accepts PDU Session; gNB establishes default DRB mapped to QoS Flow (QFI for 5QI=5).
- P-CSCF discovery via PCO (Protocol Configuration Options) or SMF-provided info.
- This PDU Session carries all SIP signaling.
IMS Registration:
- UE sends SIP REGISTER (via Gm interface to P-CSCF) over the IMS signaling QoS Flow.
- P-CSCF → I-CSCF → S-CSCF (with Cx/Sh interactions to UDM for authentication via AKA using ISIM).
- Successful 200 OK completes IMS registration.
- IMS now knows the UE’s contact (IP address, P-Access-Network-Info with 5G cell ID, RAT type).

These steps ensure the UE is “IMS voice capable” and ready. In practice, they occur at power-on or upon entering 5G coverage.

Detailed Mobile-Originated (MO) VoNR Call Flow

The flow divides into phases: call initiation, resource reservation (preconditions), alerting, media establishment, and active call. Typical end-to-end setup time targets <2 seconds in good conditions (faster than VoLTE due to 5G optimizations).

Phase 1: SIP Signaling Initiation (Call Setup Trigger)

UE (in RRC Connected if needed; may trigger Service Request from IDLE) sends SIP INVITE (with SDP offer).
- SDP includes: supported codecs (EVS mandatory/preferred; AMR-WB fallback), media descriptions (audio, RTP/AVP, ports), preconditions (a=des:qos mandatory local/remote sendrecv), ptime (usually 20 ms), and 5G-specific headers.
Signaling path: UE → gNB (over DRB for 5QI=5) → UPF → P-CSCF → IMS core (I/S-CSCF) → terminating network.
IMS performs routing, authorization, and charging correlation (using 5G Charging ID).
Early media or provisional responses (e.g., 100 Trying) may occur.

Phase 2: Resource Reservation and QoS Flow Setup (Critical for Reliability)

This is where VoNR differs most visibly from legacy flows — the network ensures dedicated resources before ringing.

Upon receiving INVITE (or during 183 Session Progress), P-CSCF sends policy request to PCF (via Rx or N5 interface) indicating IMS voice session needs.
PCF authorizes and pushes policy rules to SMF (policy association update).
SMF triggers PDU Session Modification:
- Instructs UPF to prepare for new QoS Flow (QFI with 5QI=1: GBR, priority ~20, PDB ~100 ms, PER 10⁻² for conversational voice).
- SMF sends PDU Session Modification Command (via AMF → gNB → UE) to establish the dedicated QoS Flow for RTP/RTCP media.
- gNB performs RRC Reconfiguration: adds/modifies DRB mapped to the new QoS Flow; configures voice optimizations (SPS, C-DRX, RoHC for header compression, packet duplication if configured).
- UE confirms with RRC Reconfiguration Complete and NAS PDU Session Modification Response.
UE and network confirm preconditions met via SIP:
- 183 Session Progress (with SDP answer) → PRACK (Provisional Response ACK) → 200 OK (PRACK).
- UPDATE / 200 OK (UPDATE) for resource confirmation (a=conf:qos and a=current:qos lines in SDP).
This “resource reservation” prevents “ghost ringing” (alerting the callee before resources are ready) and ensures low-latency, guaranteed-bit-rate treatment for voice packets.

Nuance: The dedicated QoS Flow can be within the existing IMS PDU Session (common) or a new one. Admission control at gNB/SMF may reject if resources are insufficient.

Phase 3: Alerting and Call Acceptance

Once preconditions are confirmed, IMS forwards the call to the terminating side.
Terminating UE receives INVITE → rings → sends 180 Ringing (provisional response) back through IMS to originating UE.
Originating UE may play ringback tone locally or via early media.
Called party answers → terminating side sends SIP 200 OK (final response with SDP answer confirming codec, usually EVS).
Originating side sends SIP ACK.

Phase 4: Media Establishment and Active Call

Bidirectional RTP/RTCP media streams start over the dedicated 5QI=1 QoS Flow.
- Voice packets use EVS codec (adaptive bitrate 5.9–128 kbps, channel-aware mode for robustness).
- gNB applies scheduling optimizations: semi-persistent scheduling (SPS) or configured grant for periodic small packets, reducing control channel overhead.
- UPF routes media to peer (or via IMS media handling if needed).
Full 5G data remains available simultaneously on other QoS Flows/PDU Sessions (no radio switch).
Supplementary services (hold, transfer, conferencing) use additional SIP re-INVITEs with possible further QoS modifications.

Call Setup Time KPI: Measured from SIP INVITE (MO) to 180 Ringing (or 200 OK). VoNR often achieves faster times than VoLTE due to lower air interface latency and efficient 5G procedures.

Mobile-Terminated (MT) VoNR Call Flow Differences

IMS receives INVITE for the called UE.
P-CSCF queries PCF/SMF for terminating access domain (T-ADS considers 5G registration, location, preferences).
If UE is idle, AMF pages the UE (via gNB) → UE performs Service Request to re-establish NAS signaling connection.
Similar PDU Session Modification for dedicated QoS Flow (5QI=1) on the terminating side.
SIP 180 Ringing → 200 OK follows the same precondition/resource reservation logic.
The flow mirrors MO but with paging and terminating-side policy triggers.

Call Termination / Release Flow

Either party can end the call:

UE sends SIP BYE.
IMS propagates BYE → 200 OK.
PCF/SMF triggers PDU Session Modification or Release to deactivate the dedicated 5QI=1 QoS Flow and associated DRB (resources freed).
gNB releases radio resources; UE returns to default signaling flow.
Charging records finalized.

Clean releases avoid dangling resources; abnormal releases (e.g., radio link failure) trigger context release and IMS session timeout.

Key Optimizations and 5G-Specific Enhancements in the Flow

RAN Optimizations: SPS/CG for voice periodicity, RoHC (robust header compression reduces RTP/UDP/IP overhead), C-DRX for UE power saving between talk spurts, TTI bundling/repetition for coverage.
QoS Differentiation: Strict mapping — 5QI=5 (signaling: low latency, high reliability), 5QI=1 (voice: GBR), 5QI=2 (video if ViNR).
Network Slicing: Voice can use a dedicated slice with tailored QoS/URLLC profiles.
IMS Transparency: IMS sees minimal changes (mainly extended headers for 5G info); core impact is higher on 5GC (SMF/PCF handling QoS Flows instead of bearers).
Mobility During Call: Handover to another gNB uses make-before-break with QoS Flow continuity. Handover to LTE (if coverage lost) uses enhanced mechanisms (Rel-16 improvements).

Comparison: VoNR vs. EPS Fallback Call Flow

In early deployments, many networks use EPS Fallback instead of native VoNR:

During MO/MT, when dedicated QoS Flow (5QI=1) setup is attempted, the network rejects it with “IMS voice fallback” cause.
Triggers redirection (RRC Release with redirect info) or inter-RAT handover to LTE (via N26 if available).
UE moves to 4G → performs TAU/Attach → sets up VoLTE dedicated bearer (QCI=1).
Call then proceeds as VoLTE; data may be limited during the call.
Impact: Adds delay (extra radio switch, ~500 ms–2 s), temporary loss of 5G data speeds, but higher reliability in immature 5G coverage.
VoNR avoids this entirely once coverage and optimization are sufficient.

Edge Cases and Nuances:

Weak Coverage: Frequent handover to VoLTE; operators tune thresholds carefully.
Roaming: Requires IMS NNI testing, consistent QoS mapping, and handling of local breakout vs. home-routed scenarios.
Emergency Calls: Similar flow but with prioritized QoS and accurate NR positioning for E911.
Video Calls (ViNR): Additional QoS Flow (higher bandwidth) and SDP for video codecs (e.g., H.265).
Power/Performance: Optimizations reduce impact, but high-band 5G may increase drain in fringe areas compared to LTE.
Failure Handling: SIP retries, QoS rejection fallback to VoLTE, or call drop with re-attempt.
SMS Integration: Can use SMS over IMS or SMS over NAS (independent but shares signaling paths).
Testing Considerations: KPIs include call setup success rate, drop rate, MOS score (via EVS quality), latency (SIP and RTP), and resource utilization. Tools analyze full protocol traces (RRC, NAS, NG-AP, SIP, RTP).

Implications and Broader Context

This detailed flow enables superior user experience: faster setup, richer audio (EVS fullband), seamless multitasking, and preparation for future features like immersive voice (IVAS in later releases) or AI-enhanced calls. For operators, it simplifies the network long-term (phasing out VoLTE) but requires dense 5G coverage, low-latency backhaul, and careful tuning of admission control/scheduling.

In mature SA networks, native VoNR dominates where coverage allows, with EPS Fallback as a robust safety net. Real performance varies by spectrum (low-band for coverage, mid/high for capacity), device chipset support, and operator configuration.

RTP (Real-time Transport Protocol)

RTP (Real-time Transport Protocol) and RTCP (RTP Control Protocol) form the core media transport and monitoring layer for real-time applications like voice and video in packet-switched networks, including VoLTE and VoNR (Voice over New Radio / Vo5G). Defined primarily in IETF RFC 3550 (with extensions in other RFCs), they operate above UDP to deliver time-sensitive media streams while providing feedback for quality adaptation.

In the context of IMS-based mobile voice services, RTP carries the actual encoded voice payload (e.g., from AMR-WB or EVS codecs), while RTCP monitors transmission quality, enabling dynamic adjustments to maintain conversational clarity. Both protocols are essential once SIP/SDP negotiation completes and dedicated QoS resources (QCI=1 in VoLTE or 5QI=1 in VoNR) are established.

RTP: Real-time Transport Protocol – The Media Carrier

RTP provides end-to-end delivery of real-time data such as audio (voice) or video over IP networks. It does not guarantee delivery, ordering, or timeliness on its own — those rely on the underlying UDP transport and application-level mechanisms (jitter buffers, error concealment in the codec, etc.). Instead, RTP adds critical real-time metadata to each packet.

Key Features of RTP:

Timestamping: Each packet includes a timestamp (32-bit) reflecting the sampling instant of the first octet in the payload. This enables synchronization and jitter calculation at the receiver.
Sequence Numbering: A 16-bit sequence number increments per packet, allowing detection of lost, duplicated, or out-of-order packets.
Payload Type (PT): 7-bit field identifying the codec/format (e.g., dynamic PT for EVS or AMR-WB negotiated via SDP).
Synchronization Source (SSRC): 32-bit identifier uniquely identifying the source of the RTP stream within a session (helps in multicast or multi-party scenarios).
Contributing Source (CSRC): Optional list for mixers (e.g., conference calls).
Marker Bit: Indicates significant events, such as the start of a talk spurt in voice (useful for discontinuous transmission/DTX).
Padding and Extension: Flags for optional padding or header extensions.

RTP Packet Structure (simplified):

Fixed header (12 bytes minimum): Version (2 bits, usually 2), Padding (1), Extension (1), CSRC Count (4), Marker (1), Payload Type (7), Sequence Number (16), Timestamp (32), SSRC (32), optional CSRC list.
Followed by the payload: Encoded voice data (e.g., EVS frame at 5.9–128 kbps, typically 20 ms packetization interval).
Optional header extensions (e.g., for absolute capture time or other 3GPP-specific info).

In VoNR/VoLTE:

RTP packets carry the output of the voice codec (EVS preferred in VoNR for super-wideband/fullband quality; AMR-WB common in VoLTE).
Transported over UDP (even port number; RTCP uses the next odd port).
In the 5G user plane: UE → gNB (over DRB mapped to 5QI=1 QoS Flow) → UPF → IMS/media gateway or peer endpoint.
Packetization is typically 20 ms for voice to balance latency and efficiency.

Typical RTP Packet Size (Uncompressed):

Headers: IP (20/40 bytes IPv4/IPv6) + UDP (8) + RTP (12) = ~40–60 bytes overhead.
Payload: Small for voice (~20–50 bytes depending on codec/bitrate) → high overhead ratio without optimization.

RTCP: RTP Control Protocol – The Quality Feedback Mechanism

RTCP runs alongside RTP (on a separate port) to provide out-of-band control and monitoring. It does not carry media but periodically sends reports that participants use to assess and adapt to network conditions.

Key RTCP Packet Types:

Sender Report (SR): Sent by active senders; includes transmission statistics (packets/octets sent, timestamps) plus reception report blocks.
Receiver Report (RR): Sent by receivers (or non-active senders); reports reception quality for one or more SSRCs.
Source Description (SDES): Provides textual info like CNAME (canonical name) for source identification.
BYE: Indicates session termination.
APP: Application-specific packets.

Key Metrics in RTCP Reports (from Receiver/Sender Reports):

Fraction Lost: Packet loss rate since the last report.
Cumulative Number of Packets Lost.
Interarrival Jitter: Estimated variance in packet arrival times (critical for voice; high jitter causes choppiness).
Last SR Timestamp (LSR) and Delay Since Last SR (DLSR): Used to calculate round-trip time (RTT).
Extended Highest Sequence Number Received.

These statistics allow:

Senders to adapt (e.g., lower EVS bitrate or switch to channel-aware mode during congestion).
Receivers to adjust jitter buffer size.
Overall Quality of Experience (QoE) monitoring by the network/IMS (e.g., for Mean Opinion Score estimation or troubleshooting).

RTCP Transmission Interval:

Dynamically calculated based on session bandwidth (typically 5% of RTP session bandwidth reserved for RTCP).
Aim: ~5 seconds average interval per participant in large sessions, more frequent in small point-to-point calls.
In mobile voice, operators may tune this for faster feedback.

RTP/RTCP in VoNR vs. VoLTE: Similarities, Differences, and Optimizations

Both technologies use RTP/RTCP identically at the application layer (IMS media plane), but the underlying 5G vs. 4G infrastructure introduces nuances in performance, efficiency, and integration.

Similarities:

Media transport after SIP SDP negotiation (codec, ports, IP addresses agreed).
QoS prioritization: Dedicated bearer (QCI=1 in VoLTE) or QoS Flow (5QI=1 in VoNR) for conversational voice — GBR, low packet delay budget (~100 ms), low error rate.
UDP-based for low latency; no retransmissions at transport layer (RLC UM mode often used in RAN for voice DRBs to avoid delay from ARQ).
Support for EVS and AMR-WB codecs; payload formats defined in 3GPP TS 26.XXX series.
RTCP for bidirectional quality feedback.

Differences and VoNR Advantages:

Latency and Responsiveness: 5G NR air interface + 5GC offers lower end-to-end latency than LTE/EPC. This reduces jitter and improves conversational flow, making RTCP-derived adaptations (e.g., bitrate switching) more effective. VoNR call setup and media path can feel more “immediate.”
Simultaneous Data: In native VoNR (5G SA), high-speed data continues on parallel QoS Flows without radio fallback. VoLTE (or EPS Fallback from 5G) may throttle or switch radio during voice.
Header Overhead Handling: Both use Robust Header Compression (RoHC) at PDCP layer (3GPP TS 38.323 for NR; TS 36.323 for LTE). RoHC compresses IP/UDP/RTP headers dramatically — from ~40–60 bytes to as low as 1–4 bytes after context establishment (Profile 0x0001 for RTP/UDP/IP is key for voice).
- Critical in VoNR due to small voice payloads and need for spectral efficiency.
- Initial packets send full headers to build context; subsequent ones are highly compressed.
RAN Optimizations in VoNR:
- Semi-Persistent Scheduling (SPS) or Configured Grant for predictable voice packet arrivals → fewer scheduling grants.
- C-DRX (Connected Discontinuous Reception) for UE power saving between talk spurts.
- RoHC is mandatory/emphasized in VoNR profiles for efficiency.
- EVS codec integration: Better packet loss resilience (channel-aware mode) works synergistically with RTP sequence numbers and RTCP reports.
Protocol Stack in VoNR:
- UE: Codec → RTP/RTCP → UDP/IP → PDCP (RoHC + ciphering) → RLC (often UM) → MAC → PHY (NR).
- Network: gNB (terminates PDCP/RLC/MAC) → UPF (user plane routing) → IMS.
Packet Sizes: VoNR RTP packets (with RoHC) typically 40–100+ bytes depending on EVS mode/bitrate. Without optimization, overhead dominates; RoHC makes voice highly efficient.
Interworking: In mixed VoNR ↔ VoLTE calls, RTP/RTCP flows transparently, but codec may negotiate down (EVS AMR-WB IO mode avoids transcoding). Handover from VoNR to VoLTE preserves media continuity where possible.

Edge Cases and Considerations:

Packet Loss/Jitter Handling: RTP sequence + timestamps + RTCP reports feed into jitter buffer and codec concealment. EVS excels here compared to older codecs. High loss (>1–2%) triggers adaptation or noticeable degradation.
Silence Suppression (DTX/VAD): Reduces packets during silence; marker bit and RTCP help manage comfort noise.
Multicast/Conferencing: RTP supports SSRC/CSRC; RTCP scales reports.
Security: Often secured via SRTP/SRTCP (Secure RTP) with keys from IMS signaling.
Power and Efficiency: In weak 5G coverage, frequent retransmissions or fallback to VoLTE increase drain. Optimizations (SPS, RoHC, C-DRX) are vital.
Monitoring/KPIs: Operators use RTCP-derived metrics (loss, jitter, MOS estimates) plus RAN counters (BLER, scheduling stats) for QoE. Tools like Wireshark decode full RTP/RTCP traces.
Future Enhancements: 5G-Advanced and beyond introduce immersive codecs (IVAS) that build on RTP with spatial audio extensions; better integration with network slicing and AI-driven adaptation.
Roaming/Interworking: Consistent RTP/RTCP behavior requires tested NNI (Network-Network Interface); QoS mapping must preserve priorities.

Implications for Voice Quality and Network Design

RTP/RTCP enable the shift from circuit-switched voice (fixed channels, no feedback) to adaptive, high-quality IP voice. In VoNR, they fully leverage 5G’s strengths — lower latency, massive capacity, and advanced scheduling — to deliver superior experiences: richer audio (via EVS + fullband), fewer artifacts, and seamless multitasking.

Challenges include:

Ensuring RoHC context synchronization (especially during handovers or radio issues).
Balancing RTCP overhead (too frequent wastes bandwidth; too infrequent delays adaptation).
Admission control on GBR flows to prevent congestion impacting voice.
Device support: UEs must implement RTP/RTCP payload handling, RoHC profiles, and jitter buffers efficiently.

In mature 5G SA deployments, RTP/RTCP over NR (with 5QI=1 treatment) provides a noticeable upgrade over VoLTE equivalents, particularly in good coverage. In transitional scenarios (EPS/RAT fallback), the media plane temporarily reverts to LTE behavior.

VoNR codec details

VoNR codec details center primarily on the Enhanced Voice Services (EVS) codec, which serves as the flagship and preferred (often effectively mandatory) voice codec in Voice over New Radio (VoNR / Vo5G) deployments. EVS, standardized by 3GPP in Release 12 (with ongoing enhancements in later releases), represents a major advancement over the AMR-WB (Adaptive Multi-Rate Wideband) codec that dominated VoLTE. While VoNR supports fallback to AMR-WB (and AMR-NB) for interoperability, its architecture and GSMA profiles (e.g., NG.114) emphasize EVS to deliver superior audio quality, efficiency, and robustness.

EVS is defined in key 3GPP specifications:

TS 26.441: General overview.
TS 26.445: Detailed algorithmic description (the core ~600+ page reference for encoder/decoder behavior).
TS 26.446: AMR-WB backward-compatible functions.
Related: TS 26.447 (error concealment), TS 26.448 (jitter buffer management), TS 26.449/26.450 (CNG/DTX), and RTP payload format in RFC 4867 extensions plus 3GPP-specific handling.

Core Technical Characteristics of EVS

EVS is a hybrid speech/audio codec that combines advanced techniques for speech (e.g., Linear Prediction, Code-Excited Linear Prediction) with music-like signal handling. It operates on 16-bit PCM input samples at one of four sampling rates:

8 kHz → Narrowband (NB): ~20–4000 Hz (or up to ~3400 Hz effective telephony).
16 kHz → Wideband (WB): ~20–8000 Hz (or ~50–7000 Hz).
32 kHz → Super-Wideband (SWB): ~20–16000 Hz (or ~50–14000/16000 Hz).
48 kHz → Fullband (FB): ~20–20000 Hz (approaching CD-quality for voice, capturing more natural harmonics, consonants, and background sounds).

Bitrate Flexibility (switched every 20 ms frame): EVS supports source-controlled variable bitrate (SC-VBR) at the lowest rate and constant bitrate (CBR) otherwise. Key operating points include:

Bandwidth	Supported Bitrates (kbps)	Notes
NB	5.9 (SC-VBR), 7.2, 8, 9.6, 13.2, 16.4, 24.4	Lowest rates for basic telephony; 5.9 always uses VBR + DTX.
WB	5.9, 7.2, 8, 9.6, 13.2, 13.2 channel-aware, 16.4, 24.4, 32, 48, 64, 96, 128 (plus 6.6–23.85 for AMR-WB IO)	Channel-aware at 13.2 kbps adds forward error correction-like redundancy.
SWB	9.6, 13.2, 13.2 channel-aware, 16.4, 24.4, 32, 48, 64, 96, 128	Significant quality leap starts here at low-to-mid rates.
FB	16.4, 24.4, 32, 48, 64, 96, 128	Highest fidelity; useful for music-on-hold, mixed speech/audio, or premium “Super HD” calls.

AMR-WB Interoperable (IO) Mode: EVS includes all 9 AMR-WB bitrates (6.6, 8.85, 12.65, 14.25, 15.85, 18.25, 19.85, 23.05, 23.85 kbps) in a fully backward-compatible mode. No transcoding is needed when interoperating with legacy AMR-WB devices/networks — packets can be decoded directly by AMR-WB decoders.
Discontinuous Transmission (DTX), Voice Activity Detection (VAD), and Comfort Noise Generation (CNG): Supported across relevant rates. DTX reduces transmission during silence, saving power and radio resources; CNG (frequency-domain or LP-based) generates natural background noise.
Packetization: Typically one frame per RTP packet (20 ms). Header formats include compact (minimal overhead) and full modes. In EVS Primary mode, compact headers are common; AMR-WB IO may use full headers in some cases.

Key Algorithmic Innovations:

Channel-Aware Mode (CAM): Available at 13.2 kbps for WB and SWB. It adds partial redundancy (forward error correction-like) to improve resilience against packet loss and jitter. This is especially valuable in challenging radio conditions (e.g., cell edge, high mobility). CAM can be activated/deactivated dynamically via Codec Mode Request (CMR) or RTCP feedback. Offset values (e.g., 2, 3, 5, 7) control redundancy level.
Error Concealment: Advanced frame erasure handling maintains acceptable quality even with >15% packet loss (far better than AMR-WB’s ~3% threshold before noticeable artifacts).
Jitter Buffer Management (JBM): Sophisticated adaptation to network delay variations.
Source-Controlled Operation: The encoder internally classifies input (speech, music, mixed, noise) and adapts coding tools in real time.
Bandwidth Extension: Optimized techniques extend lower-bandwidth signals to higher ones without excessive bitrate penalty.

Subjective quality (POLQA/MOS scores) shows EVS outperforming AMR-WB at equivalent or lower bitrates, with particularly large gains in SWB/FB and impaired conditions. For example, EVS at 13.2 kbps SWB with CAM often exceeds AMR-WB at 23.85 kbps in clean channels and maintains clarity under loss.

SDP Negotiation and Configuration in VoNR

Codec selection occurs during SIP/SDP offer/answer (in INVITE, 183 Session Progress, 200 OK, PRACK/UPDATE for preconditions). The media type is audio/EVS (per IANA/RFC).

Common SDP parameters (fmtp attributes):

br (or br-send / br-recv): Bitrate range, e.g., br=5.9-24.4 or br=13.2 (single value). 5.9 denotes SC-VBR average.
bw (or bw-send / bw-recv): Bandwidth, e.g., bw=nb-swb (narrow to super-wide), bw=swb, bw=fb.
cmr: Codec Mode Request handling (-1 disabled, 0 default, 1 present in every packet).
ch-aw-recv / ch-aw-send: Channel-aware mode config (0 = off at start; positive values for offset/redundancy).
channels: Usually 1 (mono); stereo support in some extensions.
mode-set or other flags for header format, DTX, etc.

GSMA NG.114 (VoNR Profile) Recommended Configurations (examples for SDP offers):

A1: br=5.9-13.2; bw=nb-swb
A2: br=5.9-24.4; bw=nb-swb
B0: br=13.2; bw=swb
B1/B2: Variants emphasizing SWB at 9.6–24.4 kbps.

In VoNR, UEs typically prioritize EVS (with SWB/FB options) in the SDP offer. The IMS (P-CSCF/PCF) authorizes QoS (5QI=1 GBR flow) based on the negotiated bitrate/bandwidth. Dynamic adaptation during the call (bitrate/bandwidth switching, CAM on/off) is possible via re-INVITE/UPDATE or in-band CMR, though networks often lock the initial negotiation for stability.

VoNR vs. VoLTE: Codec Usage Nuances

VoLTE (IR.92 Profile): AMR-WB is the baseline (mandatory for HD Voice). EVS is optional and enabled in many mature networks, but not universally prioritized. Many deployments default to AMR-WB for simplicity and legacy compatibility.
VoNR (NG.114 Profile): EVS is strongly recommended/required in compliant profiles. Devices must include at least one EVS payload type in SDP offers (often with SWB support). This positions VoNR for “Ultra HD” or “Super HD” voice as a differentiator. AMR-WB remains for interworking (VoNR ↔ VoLTE calls often negotiate down or use IO mode to avoid transcoding).
Interworking: In mixed calls, the network may force a downgrade (via bearer modification if QoS doesn’t match) or use transcoding (resource-intensive). TRFO (Transcoder-Free Operation) prefers IO mode. Handover from VoNR to VoLTE preserves media where possible, potentially triggering codec renegotiation.
Video Calls (ViNR): EVS pairs with video codecs (H.265 preferred in 5G); voice negotiation is independent but shares the same QoS principles.

Additional Features and Optimizations in VoNR Context

Robustness Synergy: EVS CAM + 5G RAN features (SPS/CG scheduling, RoHC for RTP/UDP/IP headers, low-latency air interface) yields excellent performance even in moderate coverage. RTP sequence numbers, timestamps, and RTCP reports (loss, jitter) feed into adaptation.
Power and Efficiency: Lower bitrates for equivalent (or better) quality reduce radio resource usage. DTX/CNG saves battery. However, higher-bandwidth modes (SWB/FB) increase processing slightly.
Edge Cases:
- Mixed Bandwidth: If encoder/decoder differ, bitrate/bandwidth adjusts (e.g., fallback to WB if peer doesn’t support SWB).
- High Packet Loss: CAM activates dynamically; concealment handles residuals.
- Roaming/Inter-operator: Requires consistent SDP handling and tested NNI; often falls back to AMR-WB.
- Emergency Calls: Reliable codec with accurate 5G location; EVS robustness helps in marginal coverage.
- Music/Audio Signals: EVS handles non-speech better than AMR-WB (e.g., music-on-hold, background music).
- Device Variability: Not all UEs enable full FB or aggressive SWB; carrier provisioning and firmware matter. Flagships increasingly support rich configurations.

Quality Implications: In strong 5G coverage, EVS SWB/FB delivers noticeably richer, more natural conversations (better presence, less muffling). Under impairment, it sustains higher MOS than AMR-WB. Operators can differentiate with premium EVS profiles (e.g., always SWB).

Future Outlook: Later 3GPP releases (5G-Advanced) build toward immersive voice/audio services (IVAS), extending EVS concepts for spatial audio, higher fidelity, and tighter integration with network slicing/AI adaptation.

EVS Channel-Aware Mode (CAM)

EVS Channel-Aware Mode (CAM) is one of the standout robustness features of the Enhanced Voice Services (EVS) codec, specifically designed to improve voice quality under challenging network conditions such as high packet loss, jitter, burst errors, or poor radio link quality in mobile environments like VoNR (5G SA) and VoLTE. It operates exclusively at a fixed bitrate of 13.2 kbps for Wideband (WB) or Super-Wideband (SWB) audio bandwidths and is detailed in 3GPP TS 26.445 (the core algorithmic specification for EVS).

Unlike standard EVS operation at the same bitrate (which uses core coding tools for high efficiency), CAM introduces partial redundancy — a form of lightweight, in-band forward error correction (FEC) — without increasing the overall transmitted bitrate. This makes it particularly valuable in real-world packet-switched networks where RTP packets can be lost, delayed, or corrupted due to radio impairments, congestion, or mobility.

How Channel-Aware Mode Works: Core Mechanism

CAM leverages source-controlled coding and partial redundancy to protect critical speech information:

Frame Classification and Redundancy Selection:
- The encoder analyzes each 20 ms speech frame using advanced signal classification (speech vs. music, transient detection, energy patterns, etc.).
- It identifies “important” previous frames (typically the one or two most recent frames) whose loss would cause noticeable artifacts.
- Spare bits are created by slightly reducing the quality of the primary (current) frame encoding. These spare bits are used to transmit a partial copy (redundant version) of selected prior frame(s).
- The redundancy is partial, not full: It includes key parameters such as Linear Prediction (LP) coefficients, excitation information, or bandwidth extension data — enough to enable better reconstruction but not a complete duplicate.
Transmission in RTP Payload:
- The RTP packet for the current frame carries:
  - The primary encoded data for the current frame (at a reduced internal quality to accommodate redundancy).
  - Embedded partial redundant data for one or more previous frames.
- This maintains the constant 13.2 kbps bitrate (including headers in the payload format).
- Redundancy can span a configurable offset (in frames), allowing the redundant data to cover losses that occurred a few packets earlier.
Decoder-Side Recovery:
- If the primary frame arrives intact, the decoder uses it normally and may ignore or store the redundant data for potential future use.
- If the primary frame is lost or late (detected via RTP sequence numbers and timestamps), the decoder reconstructs it using:
  - The partial redundant copy received in a subsequent packet.
  - Advanced error concealment tools (already strong in EVS, handling >15% loss gracefully).
- This hybrid approach significantly reduces artifacts like clicks, muffling, or choppiness compared to pure concealment.

Key Technical Details (from TS 26.445 and related specs):

Offset Values: Configurable redundancy depth, typically 2, 3, 5, or 7 frames. Higher offsets provide protection against longer bursts but may introduce slight additional delay or complexity in reconstruction.
Activation: Can be enabled at session start or dynamically during the call.
Bandwidth Restriction: Only WB (up to 8 kHz) or SWB (up to 16 kHz). Not available for NB or FB at this bitrate.
DTX Compatibility: Works with Discontinuous Transmission (DTX) and Comfort Noise Generation (CNG).
Bit Allocation Impact: In CAM, a portion of the 13.2 kbps is allocated to redundancy, so the primary encoding uses a modified quantization scheme to minimize quality degradation in clean conditions.

The mode is described in TS 26.445 section 4.4.5 (coder overview) and detailed further in encoder/decoder functional descriptions, with specific bit allocation tables in section 7.4.

SDP Negotiation and Configuration in VoNR/VoLTE

During SIP/SDP offer-answer (in INVITE, 183 Session Progress, etc.), CAM is signaled via the audio/EVS media type with specific fmtp parameters:

ch-aw-recv and ch-aw-send: Control receive/send direction.
- -1: Disabled.
- 0 (or absent): Not used at session start (can be activated later).
- Positive values (2, 3, 5, 7): Offset for partial redundancy at start.
br=13.2 and bw=wb or bw=swb (or nb-swb range that includes it).
cmr (Codec Mode Request): Allows in-band or RTCP-based requests to toggle CAM or change offset.

In GSMA NG.114 (IMS Profile for VoNR), EVS is strongly recommended, and configurations often include modes that support 13.2 kbps CAM (e.g., B0: br=13.2; bw=swb). Devices must support at least basic EVS configurations, with CAM as an enhancement for robustness.

Dynamic adaptation is possible via RTCP-APP messages or in-band CMR, allowing the system to enable CAM when RTCP reports indicate rising packet loss or jitter.

Performance Benefits and Comparisons

CAM provides a clear advantage in impaired conditions while maintaining near-transparent quality in clean channels:

Packet Loss Resilience: EVS with CAM handles 5–15%+ frame erasure rates (FER) with minimal degradation. Standard EVS or AMR-WB typically shows noticeable artifacts above 3%.
MOS Gains: Independent tests (POLQA, P.800) show EVS CAM at 13.2 kbps SWB outperforming AMR-WB at higher bitrates (e.g., 23.85 kbps) under 2–10% loss. In bursty loss scenarios, the gap widens significantly.
Clean Channel Impact: Negligible or very small quality penalty compared to non-CAM 13.2 kbps EVS, thanks to intelligent frame selection and modified quantization.
Jitter Handling: Complements EVS jitter buffer management (JBM); redundant data helps recover from late arrivals without excessive buffering delay.

Real-World Implications in VoNR:

In 5G networks, even with advanced RAN features (SPS, RoHC, low-latency air interface), cell-edge, high-mobility, or loaded scenarios can cause packet loss. CAM ensures sustained clarity without frequent codec downgrades or call drops.
Synergy with 5QI=1 QoS Flow: Prioritized treatment reduces baseline loss, but CAM acts as the last line of defense.
Compared to VoLTE: VoNR deployments more aggressively enable EVS + CAM as a 5G differentiator (“Ultra HD Voice” under stress).

Nuances and Edge Cases:

Trade-offs: Slight increase in encoder/decoder complexity and processing (minor battery impact on devices). Higher offsets may add marginal reconstruction delay.
Interworking: In mixed VoNR ↔ VoLTE calls, negotiation may fall back if the peer lacks CAM support (often using AMR-WB IO mode).
Activation Strategy: Operators can configure thresholds based on RTCP metrics (fraction lost, jitter). Not always on by default to optimize clean-channel capacity.
Burst vs. Random Loss: Excels in bursty patterns common in wireless (e.g., fading channels), where redundancy covers consecutive losses.
Music/Silence: Less effective for non-speech; EVS source classification helps route music to frequency-domain tools instead.
Future Extensions: Proposals for additional CAM bitrates (beyond 13.2 kbps) exist in research, but the standardized version remains fixed at 13.2 kbps for WB/SWB.

Broader Considerations in VoNR Context

Network Optimization: CAM works best with RoHC (header compression), proper jitter buffer sizing, and admission control on GBR flows. In weak 5G coverage, it reduces handover triggers to VoLTE.
Testing and KPIs: Evaluated via frame erasure simulations, POLQA MOS under various packet loss profiles (random/bursty), and end-to-end latency. Tools analyze RTP payload for redundancy usage.
Device and Ecosystem: Requires chipset/firmware support; flagships increasingly enable rich EVS configurations including CAM. GSMA profiles ensure interoperability.
Power and Efficiency: Redundancy uses existing bitrate efficiently; DTX still applies during silence.
Regulatory/Emergency: Enhances reliability for E911 calls in marginal coverage, where accurate 5G positioning pairs with robust media.

In summary, EVS Channel-Aware Mode transforms 13.2 kbps into a highly resilient operating point by intelligently embedding partial redundancy, making VoNR voice calls far more tolerant to real-world impairments than previous generations. It delivers clearer, more consistent conversations in challenging conditions with minimal clean-channel penalty — a key reason EVS is positioned as the native codec for 5G voice. This feature, combined with 5G’s lower latency and advanced scheduling, contributes significantly to the superior user experience in mature VoNR deployments.

Comparision of VoNR with VoLTE codecs

VoNR (Voice over New Radio / Vo5G) and VoLTE (Voice over LTE) both deliver packet-switched voice services over IMS using SIP/SDP for session control and RTP/RTCP for media transport. The core voice codecs available to both are largely the same set defined by 3GPP, but there are important differences in mandatory support, typical/default usage, bandwidth options, robustness features, bitrate efficiency, and real-world performance influenced by the underlying radio access and core network.

Primary Codecs in VoLTE and VoNR

AMR-NB (Adaptive Multi-Rate Narrowband): Legacy narrowband codec (up to ~3.4 kHz audio bandwidth). Provides basic quality; rarely used as primary in modern HD Voice deployments but remains for fallback/interoperability.
AMR-WB (Adaptive Multi-Rate Wideband): The foundation of “HD Voice.” Supports wideband audio (up to 7 kHz). Bitrates: 6.6–23.85 kbps. Standardized earlier and widely deployed in VoLTE.
EVS (Enhanced Voice Services): The advanced codec family introduced in 3GPP Release 12. Designed for both VoLTE and VoNR. It is backward-compatible with AMR-WB via an AMR-WB IO (Interoperable) mode, allowing seamless switching or transcoding-free interworking with legacy AMR-WB devices/networks.

Key Nuance: Both technologies can negotiate EVS or AMR-WB via SDP offer/answer during call setup (in SIP INVITE, 183 Session Progress, 200 OK, etc.). The IMS core ultimately decides the codec based on UE capabilities, network policy, peer support, and QoS authorization. However, implementation differs significantly between VoLTE and VoNR.

Codec Support and Mandates

VoLTE (4G):
- AMR-WB is the primary/mature codec for HD Voice.
- EVS is optional (though increasingly supported and enabled by many operators for better quality).
- Older or basic VoLTE deployments often default to AMR-WB only.
- Mandatory support typically includes AMR-NB and AMR-WB; EVS requires additional device and network configuration.
VoNR (5G SA):
- EVS is strongly recommended or effectively mandatory per GSMA NG.114 guidelines and many operator profiles for Vo5G.
- Devices compliant with 5G voice often prioritize EVS in SDP offers (e.g., EVS configurations with super-wideband or fullband).
- AMR-WB remains supported for interoperability (e.g., calls to VoLTE-only users or roaming).
- This shift enables “Ultra HD” or “Super HD” voice as a native 5G differentiator.

In mixed calls (VoNR ↔ VoLTE), the network may negotiate down to AMR-WB or use EVS AMR-WB IO mode to avoid transcoding.

Detailed Codec Comparison: EVS vs. AMR-WB (the Main Practical Difference)

Aspect	AMR-WB (Primary in VoLTE)	EVS (Primary/Advanced in VoNR; Optional in VoLTE)
Audio Bandwidth	Wideband (WB): up to 7 kHz (50–7000 Hz)	Narrowband (NB), Wideband (WB), Super-Wideband (SWB: up to 14 kHz), Fullband (FB: up to 20 kHz)
Bitrate Range	6.6 – 23.85 kbps (fixed modes)	5.9 – 128 kbps (highly flexible; switches every 20 ms)
Quality at Same Bitrate	Good HD Voice; natural but limited high frequencies	Superior at equivalent or lower bitrates; richer, more natural sound (especially SWB/FB)
Efficiency	Higher bitrates needed for good quality	Same or better quality at lower bitrates → better spectral efficiency and capacity
Robustness (Packet Loss/Jitter)	Basic packet loss concealment	Advanced: Channel-Aware Mode (CAM) at 13.2 kbps (WB/SWB); source-controlled VBR; better error concealment and jitter buffer handling
Modes	Fixed bitrates; no channel-aware	Multiple modes including Channel-Aware; adaptive; supports music-like signals better
Backward Compatibility	N/A (legacy)	Full AMR-WB IO mode (no transcoding needed when interoperating)
Typical Use in Practice	Default in many VoLTE networks; MOS ~4.0–4.5 in good conditions	Default/preferred in VoNR; enables “Super HD” with MOS approaching 4.5–4.8+; closer to source audio quality
Impact on Network	Higher resource use for equivalent quality	More efficient; allows higher capacity or better quality under same radio conditions

EVS Bandwidth Modes (sampling rates):

NB: ~8 kHz sampling
WB: 16 kHz
SWB: 32 kHz (captures more consonants and naturalness)
FB: 48 kHz (approaches CD-quality for voice, better for mixed speech/music or background sounds)

Subjective and objective tests (POLQA MOS) consistently show EVS outperforming AMR-WB, especially under network impairments. For example, EVS at 13.2 kbps SWB with channel-aware mode maintains higher MOS than AMR-WB at higher bitrates when packet loss increases.

Codec Negotiation and SDP Handling

In both VoNR and VoLTE:

UE includes supported codecs in SDP offer (e.g., payload types for EVS with parameters like br=5.9-24.4; bw=nb-swb or specific configurations A1/B0 etc. per GSMA).
Network (P-CSCF/PCF/SMF) authorizes QoS based on negotiated bitrate/bandwidth.
Answer selects the common codec; dynamic changes possible via re-INVITE or UPDATE.

Differences in Practice:

VoNR UEs often list EVS (with SWB/FB options) as the highest-priority codec in SDP.
VoLTE UEs may prioritize AMR-WB.
In VoNR, the dedicated 5QI=1 QoS Flow is configured with GBR suitable for the selected EVS bitrate.
Fallback logic (e.g., from EVS to AMR-WB) may trigger PDU Session Modification or bearer adjustment.
Transcoding is avoided where possible using AMR-WB IO mode.

Voice Quality and Performance Implications

Clean Channel Conditions: EVS (especially SWB/FB) delivers noticeably richer, clearer audio with better presence and reduced muffling. Users often describe it as “more natural” or “FaceTime-like.”
Impaired Conditions (packet loss, jitter, poor radio): EVS channel-aware mode provides superior robustness. Studies show VoNR with EVS maintains higher MOS than VoLTE with AMR-WB as SNR drops or interference increases. 5G’s lower latency, better scheduling (SPS/CG), and advanced RAN features further amplify this advantage.
Latency: VoNR inherently offers lower end-to-end media latency due to 5G air interface and core, enhancing conversational flow regardless of codec. Combined with EVS, the experience feels more immediate.
Capacity/Efficiency: EVS allows operators to deliver higher quality using fewer radio resources, which is particularly valuable in dense 5G deployments.
Power and Device Impact: Higher-bandwidth EVS modes may slightly increase processing/battery use, but optimizations in 5G (C-DRX, RoHC) help mitigate this.

Real-World Nuances:

Not all VoLTE networks enable EVS even when devices support it — many stick with AMR-WB for simplicity and legacy compatibility.
VoNR deployments more aggressively enable EVS as a 5G selling point (“5G HD Voice+” or “Ultra HD”).
In roaming or inter-operator calls, negotiation often falls back to AMR-WB unless both sides support EVS.
Video calls (ViNR / ViLTE) pair voice codecs with video codecs (H.265 preferred in VoNR vs. H.264 in older VoLTE), but voice negotiation is independent.

Broader Considerations and Edge Cases

Interworking: EVS AMR-WB IO mode ensures smooth transitions without quality loss from transcoding.
Emergency Calls: Both must support reliable codecs; EVS robustness helps in marginal coverage.
Future Enhancements: Later 3GPP releases introduce Immersive Voice and Audio Services (IVAS) building on EVS concepts for spatial/immersive audio in 5G-Advanced/6G.
Operator Variability: Actual experienced quality depends on configuration — some VoLTE networks with EVS enabled can match or approach basic VoNR quality, while mature VoNR with full EVS FB can feel markedly superior.
Testing: Quality is measured via POLQA/MOS, packet loss resilience, and end-to-end latency. EVS consistently scores higher in independent evaluations, especially under stress.

In summary, while the codec toolbox overlaps significantly, VoNR is positioned to deliver a superior voice experience primarily through preferential and more capable use of EVS. This yields wider audio bandwidth, better efficiency, stronger robustness to impairments, and overall higher perceptual quality — especially when combined with 5G’s lower latency and advanced radio optimizations. VoLTE relies more heavily on AMR-WB, providing solid HD Voice that remains excellent in mature 4G networks but generally falls short of VoNR’s potential “Super HD” ceiling.

The practical difference is most noticeable in good-to-moderate 5G coverage (richer sound, fewer artifacts) and challenging conditions (sustained clarity). As 5G SA matures and EVS becomes ubiquitous, the gap will widen, accelerating the long-term migration from VoLTE. For specific operator deployments, device models, or quantitative MOS comparisons in particular scenarios, real-network traces or POLQA test results would provide even finer granularity.

Benefits and Implications of VoNR (Voice over New Radio / Vo5G)

Benefits and Implications of VoNR (Voice over New Radio / Vo5G) represent a significant evolution in mobile voice services, shifting from reliance on legacy 4G LTE (VoLTE) or circuit-switched fallbacks to a fully native, packet-switched solution within the 5G Standalone (SA) architecture. By leveraging the 5G New Radio (NR) air interface, 5G Core (5GC) with its service-based architecture, advanced QoS Flows (e.g., 5QI=1 for voice), and IMS for session control, VoNR unlocks superior performance while enabling long-term network simplification.

This analysis explores benefits from multiple angles — user experience, technical/network performance, operator economics, and ecosystem implications — alongside nuances, edge cases, real-world considerations, and broader strategic outcomes. It draws on architectural strengths (lower latency, better spectral efficiency, EVS codec prioritization) while acknowledging transitional realities like coverage dependencies and hybrid deployments.

1. User Experience Benefits

VoNR delivers a noticeably enhanced conversational experience compared to VoLTE or earlier technologies:

Superior Voice Quality: Prioritizes the EVS (Enhanced Voice Services) codec with support for super-wideband (up to 14 kHz) or fullband (up to 20 kHz) audio, delivering richer, more natural sound with reduced distortion and better background noise handling. In good conditions, this approaches “near-real-life” or “Super HD” quality, outperforming typical AMR-WB HD Voice in VoLTE. Robustness features like EVS channel-aware mode maintain clarity under packet loss or jitter.
Faster Call Setup and Lower Latency: Signaling and media paths benefit from 5G optimizations, often achieving setup times under 1–2 seconds (vs. 2–3+ seconds in VoLTE) and end-to-end media latency reductions (e.g., ~50 ms vs. higher in LTE). This improves conversational flow, reduces “dead air,” and supports real-time applications like live translation or responsive emergency calls.
Seamless Multitasking: True simultaneous high-speed 5G data and voice without radio switches or throttling. Users can download files, use maps, stream, or browse at full 5G speeds during calls — a major improvement over EPS Fallback or legacy scenarios where data degrades.
Reduced Interruptions: No forced handover to LTE for voice anchoring in native SA coverage, leading to more stable connections and fewer drops in well-optimized areas.

Nuances: Quality gains are most pronounced in strong 5G coverage. In fringe areas, frequent handovers to VoLTE can negate some advantages, and subjective improvements depend on device, peer endpoint, and network tuning.

2. Technical and Network Performance Benefits

VoNR fully exploits 5G capabilities for efficiency and scalability:

Optimized Resource Utilization: RAN features like semi-persistent scheduling (SPS), configured grants, C-DRX, TTI bundling, and robust header compression (RoHC) make voice traffic highly efficient. Voice packets benefit from beamforming, massive MIMO, and advanced interference management, improving capacity and coverage in dense deployments.
Advanced QoS and Slicing: Dedicated GBR QoS Flows ensure predictable low-latency treatment. Network slicing allows isolated, prioritized voice slices (e.g., for mission-critical or premium services), with flexible policy control via PCF.
Future-Proof Foundation: Native support for video calls (ViNR with H.265), Rich Communication Services (RCS), and integration with immersive technologies (AR/VR-enhanced calls, spatial audio via future IVAS codecs). Lower latency and higher throughput prepare the network for convergence of voice, data, and multimedia.
Security Enhancements: Inherits 5G SA improvements, such as concealed subscriber identifiers (SUPI), reducing risks like IMSI catchers.

Edge Cases: In high-mobility or weak coverage (especially mid/high-band 5G), performance may degrade, triggering VoLTE handovers. Capacity benefits shine in loaded networks but require careful admission control to protect GBR flows from data congestion.

3. Operator and Economic Benefits

For mobile network operators (MNOs), VoNR supports strategic transformation:

Network Simplification and OPEX Reduction: Enables eventual decommissioning of legacy circuit-switched infrastructure and reduced reliance on VoLTE/EPS Fallback. Consolidates voice and data under a unified 5G/IMS architecture, lowering long-term maintenance costs and spectrum fragmentation.
Spectrum and Capacity Efficiency: Better utilization of 5G spectrum (including low-band for coverage) allows higher voice capacity without dedicated resources. This frees spectrum for data services and supports denser urban deployments.
Monetization Opportunities: Differentiated “5G Calling” or “Ultra HD Voice” offerings, premium slices for enterprises (e.g., low-latency voice in private 5G), and integration with advanced services like AR communications or mission-critical push-to-talk. Supports new revenue from converged multimedia and IoT/wearable voice extensions.
Accelerated 5G SA Maturity: VoNR acts as a key driver for full SA rollout, unlocking broader 5G use cases (URLLC, massive IoT) and competitive differentiation.

Implications: Market growth projections reflect strong potential, with VoNR contributing to 5G service expansion. However, high initial infrastructure and spectrum costs can delay ROI, particularly in regions with uneven 5G coverage.

4. Battery, Device, and Ecosystem Implications

Power Consumption: Optimizations (C-DRX, SPS, RoHC) help in stable conditions, and eliminating LTE fallback can save power by avoiding radio switches. However, real-world tests during mobility show higher drain for VoNR in some scenarios due to 5G beam management and higher-frequency usage in fringe areas. Mature implementations and low-band spectrum mitigate this.
Device Ecosystem: Requires compatible chipsets, firmware, and IMS profiles. Flagship devices increasingly support VoNR via updates, but broader adoption depends on carrier provisioning. Not all UEs enable it efficiently initially.
Roaming and Interworking: Seamless handover to VoLTE ensures continuity, but full benefits require widespread SA coverage and tested NNI (network-network interfaces). Emergency services (E911) gain from accurate 5G positioning.

Broader Ecosystem Implications:

Legacy Sunsetting: Accelerates phase-out of 2G/3G and eventually full VoLTE in mature markets, simplifying operations but requiring careful migration to avoid service gaps.
Innovation Enablement: Paves the way for 5G-Advanced features (immersive communications) and 6G convergence. Supports enterprise/private 5G with native voice (e.g., factories, campuses).
Societal/Regulatory: Improved accessibility (clearer calls for hearing-impaired via better codecs), support for public safety, and compliance with evolving QoS/emergency standards.
Global Variability: Leading deployments (e.g., T-Mobile US with broad VoNR reach) show rapid call share growth, while others rely on hybrid EPS Fallback. As of 2026, VoNR is transitioning from selective/limited to broader use in SA-mature regions, with coverage densification as the key enabler.

Challenges, Trade-offs, and Edge Considerations

Coverage Dependency: VoNR performs best with robust low-band 5G; mid/high bands offer capacity but limited range, leading to more handovers in rural/suburban or indoor scenarios.
Transitional Impacts: Early deployments use EPS Fallback for reliability, temporarily sacrificing some 5G data advantages during calls. Optimization of handover thresholds is critical to minimize disruptions.
Implementation Costs: High capex for SA infrastructure, backhaul, and tuning; security/management overhead adds complexity.
Quality Variability: Benefits are conditional on network load, device capabilities, peer support (mixed VoNR-VoLTE calls may negotiate down), and configuration. Real-world MOS gains and setup improvements are measurable but context-dependent.
Power and Mobility Trade-offs: Higher consumption in dynamic scenarios requires ongoing device/RAN refinements.

Long-Term Strategic Implications: VoNR positions operators for a simplified, all-IP 5G core, reducing complexity while enabling differentiated services. For users, it elevates everyday calling toward more immersive, reliable communication. Economically, it supports efficiency gains and new revenues but demands sustained investment in coverage and optimization. In a hybrid world persisting for years, the experience blends VoNR’s strengths with VoLTE’s maturity.

ViNR (Video over New Radio / Video over 5G)

ViNR (Video over New Radio / Video over 5G) extends the native IMS-based multimedia capabilities of 5G Standalone (SA) networks to support high-quality bidirectional video calls alongside voice. It builds directly on the VoNR architecture, using the same 5G New Radio (NR) access, 5G Core (5GC) with service-based elements (AMF, SMF, UPF, PCF), and IP Multimedia Subsystem (IMS) for session control via SIP/SDP.

While VoNR focuses on conversational voice (typically with EVS codec over a dedicated GBR QoS Flow with 5QI=1), ViNR adds synchronized video media, enabling face-to-face calling with richer visual experience. This is analogous to ViLTE (Video over LTE) in 4G, but ViNR leverages 5G’s lower latency, higher capacity, advanced QoS Flows, and better radio optimizations for superior performance — especially in mature SA deployments where full 5G data speeds remain available simultaneously.

ViNR is defined in key specifications:

3GPP TS 26.114: IMS Multimedia Telephony (MTSI) media handling and interaction, covering video codecs, RTP/RTCP transport, and robustness features.
GSMA NG.114: IMS Profile for Voice, Video, and Messaging over 5GS — the primary interoperability profile for ViNR, extending VoNR requirements to include mandatory/optional video support.
Related: 3GPP TS 23.501/23.502 for 5G QoS and procedures; RTP payload formats for video.

As of 2026, ViNR is commercially feasible and verified in vendor interworking tests. It is increasingly activated in 5G SA networks where coverage and device support allow, though many operators still rely on hybrid approaches or EPS Fallback for reliability in early deployments.

Architecture and Key Components for ViNR

ViNR reuses the VoNR end-to-end stack with extensions for video:

5G RAN (gNB): Establishes multiple Data Radio Bearers (DRBs) mapped to QoS Flows. Video requires additional radio resources and scheduling tailored to higher, more variable bitrates.
5GC:
- AMF handles mobility and paging.
- SMF manages PDU Session modifications for dedicated QoS Flows.
- UPF routes RTP/RTCP media streams.
- PCF authorizes policies based on IMS signaling (via enhanced Rx/N5).
IMS: SIP signaling for session setup, SDP negotiation of audio + video codecs/ports, and supplementary services (e.g., video ringback tones in later releases). T-ADS (Terminating Access Domain Selection) decides routing.
Media Plane:
- Audio: EVS (preferred in ViNR, with WB/SWB/FB options and Channel-Aware Mode for robustness).
- Video: Primarily H.265 (HEVC – High Efficiency Video Coding) as the advanced option; H.264 (AVC) remains for backward compatibility.
- Transport: RTP/RTCP over UDP, with Robust Header Compression (RoHC) at PDCP to reduce overhead.

QoS Differentiation (Critical for ViNR):

5QI=5: Non-GBR, high-priority for SIP signaling (low latency, high reliability).
5QI=1: GBR for conversational voice (PDB ~100 ms, PER 10⁻²).
5QI=2: GBR for conversational video (live) — priority ~40, PDB ~150 ms, PER 10⁻³. This ensures video gets guaranteed bitrate treatment while allowing slightly higher delay tolerance than voice.
Additional non-GBR flows (e.g., 5QI=6–9) for best-effort data or buffered streaming if needed.

The gNB maps these to appropriate DRBs (often RLC UM mode for low latency). Network slicing can isolate ViNR traffic for premium or enterprise use cases.

Detailed ViNR Call Flow (High-Level, Building on VoNR)

The flow mirrors a VoNR call but includes video in SDP negotiation and an extra dedicated QoS Flow for video media. Preconditions ensure both audio and video resources are reserved before alerting.

Prerequisites: Same as VoNR — 5G registration, default PDU Session for IMS (5QI=5 signaling), IMS registration.
SIP Initiation (MO Call):
- UE sends SIP INVITE with SDP offer including audio (EVS parameters: br, bw, ch-aw-recv, etc.) and video (m=video line with H.265/H.264 payload types, fmtp attributes for profile/level, bitrate ranges, resolution/framerate preferences, e.g., 720p/1080p at 30 fps).
- SDP may use preconditions (a=des:qos mandatory) for both media types.
IMS Processing and Policy:
- P-CSCF forwards to IMS core; PCF is triggered for policy rules covering combined voice + video session.
- SMF performs PDU Session Modification to establish two dedicated GBR QoS Flows:
  - 5QI=1 for audio (RTP/RTCP voice).
  - 5QI=2 for video (RTP/RTCP video).
Resource Reservation and SDP Negotiation:
- gNB configures DRBs via RRC Reconfiguration.
- SIP provisional responses (183 Session Progress with SDP answer) + PRACK/UPDATE confirm preconditions for both audio and video.
- Codec negotiation: Audio often settles on EVS (possibly SWB with CAM); video prefers H.265 Main Profile for efficiency, falling back to H.264 Constrained Baseline if needed.
Call Acceptance and Media:
- 180 Ringing → 200 OK (final SDP).
- Bidirectional RTP streams start: Audio over 5QI=1, Video over 5QI=2.
- RTCP provides feedback on loss, jitter, and quality for dynamic adaptation (e.g., bitrate reduction on congestion).
Termination: SIP BYE triggers release of dedicated QoS Flows/DRBs.

MT Call Differences: Involves paging (if idle), T-ADS domain selection, and mirrored resource setup on the terminating side.

Optimizations:

RAN: SPS/CG for periodic packets, C-DRX for power saving, beamforming/massive MIMO for capacity.
Video-specific: Region-of-Interest (ROI) signaling, dynamic rate adaptation, and robustness improvements from TS 26.114 (e.g., error concealment, reference picture selection).

Video Codecs in ViNR: H.265 vs. H.264 (and Comparison to ViLTE)

Video negotiation occurs independently of audio but shares the session.

H.264 (AVC – Advanced Video Coding): Mandatory baseline in many profiles (Constrained Baseline Profile Level 1.2+). Widely supported, lower encoding/decoding complexity. Good for 720p/1080p but requires higher bitrates (e.g., 4–8 Mbps for 1080p).
H.265 (HEVC – High Efficiency Video Coding): Strongly recommended or preferred in NG.114 and TS 26.114 for ViNR. Up to ~50% better compression efficiency than H.264 at equivalent quality — same visual fidelity at roughly half the bitrate. Better support for higher resolutions (1080p/4K), improved motion compensation, larger Coding Tree Units (up to 64×64 vs. 16×16 macroblocks in H.264), and advanced tools like better intra/inter prediction.

ViNR vs. ViLTE Codec Comparison (practical differences):

Aspect	ViLTE (4G)	ViNR (5G SA)
Primary Video Codec	H.264 (mandatory)	H.265 (preferred); H.264 fallback
Efficiency	Baseline; higher bitrates needed	~50% better compression; lower bitrates for same quality
Resolution/Framerate	Strong for 720p/1080p; limited 4K	Better for 1080p+; more feasible 4K with lower bandwidth
Audio Pairing	AMR-WB dominant; EVS optional	EVS prioritized (richer sync with video)
QoS	QCI=2 for video (similar parameters)	5QI=2 GBR; tighter integration with 5G scheduling
Adaptation	Good, but limited by LTE latency	Superior dynamic rate adaptation + lower end-to-end latency
Typical Bitrate (1080p)	Higher (e.g., 2–5+ Mbps combined with voice)	Lower overall due to H.265 + 5G efficiency

H.265 requires more processing power (higher CPU/GPU on devices), but modern chipsets handle it efficiently with hardware acceleration. In mixed ViNR ↔ ViLTE calls, negotiation often falls back to H.264 to avoid transcoding.

SDP examples include parameters like profile-level-id, level-asymmetry-allowed, packetization-mode, and bitrate constraints.

Benefits and Implications of ViNR

User Experience: Crystal-clear, synchronized audio-video with natural feel. Faster setup, lower latency (improved conversational flow), and seamless multitasking (full 5G data during calls). Richer features like video ringback tones/announcements (enhanced in Rel-16+).
Quality and Efficiency: H.265 + EVS delivers higher perceptual quality at lower resource use. 5G RAN optimizations (beamforming, etc.) improve coverage/capacity for video users.
Operator Advantages: Monetization via premium “5G Video Calling” tiers or slices. Network simplification (unified SA architecture). Higher spectral efficiency supports more simultaneous calls.
Ecosystem: Prepares for immersive services (AR-enhanced calls, spatial audio/video in 5G-Advanced). Better integration with RCS and data channels.

Nuances and Edge Cases:

Coverage: Video is more bandwidth-hungry and sensitive to radio conditions than voice. Mid/high-band 5G excels for capacity but needs low-band for wide-area reliability; frequent handovers to ViLTE/VoLTE possible in fringe areas.
Device Support: Requires capable chipsets, cameras, displays, and firmware. Not all 5G phones enable ViNR out-of-the-box; carrier provisioning is key. Battery impact higher due to video encoding/decoding and sustained high data rates.
Interworking/Roaming: Tested NNI essential; mixed calls may downgrade codecs or trigger fallback. Emergency video (if supported) needs accurate 5G positioning.
Capacity/QoS Management: Video GBR flows compete with other traffic; admission control and pre-emption are critical in loaded cells. High-motion scenes or poor lighting increase bitrate needs.
Power and Heat: Video calls drain battery faster; optimizations like C-DRX and adaptive bitrate help, but edge cases (weak signal + high resolution) exacerbate this.
Regulatory: Must support lawful interception, emergency services with video (where mandated), and accessibility features.

Deployment Status: ViNR is ready for activation in many 5G SA networks and smartphones, with vendor interworking verified. Leading operators in mature SA markets (e.g., parts of Asia, US, GCC) offer or trial it as a 5G differentiator. However, EPS Fallback or hybrid VoNR/ViLTE remains common for reliability during transition. Full native ViNR grows with 5G coverage densification and low-band spectrum availability.

Broader Implications: ViNR accelerates the shift from OTT video apps (e.g., WhatsApp, FaceTime) to carrier-grade, QoS-guaranteed calling — more reliable, integrated billing, and better quality in cellular coverage. It positions 5G for converged multimedia services and future evolutions like immersive communications. Challenges include uneven global rollout, device ecosystem maturity, and optimization tuning, but benefits in user satisfaction and network efficiency are compelling in high-capacity areas.

How VoNR Works: Architecture and Key Components

Benefits and Implications

Deployment Status

Challenges, Edge Cases, and Considerations

VoNR: Architecture and Key Components

High-Level Architecture of VoNR

Detailed VoNR Call Flow (Simplified End-to-End)

VoNR vs. EPS Fallback: Transitional Context

Nuances, Edge Cases, and Considerations

VoNR Call Flow

Prerequisites Before Any Call

Detailed Mobile-Originated (MO) VoNR Call Flow

Phase 1: SIP Signaling Initiation (Call Setup Trigger)

Phase 2: Resource Reservation and QoS Flow Setup (Critical for Reliability)

Phase 3: Alerting and Call Acceptance

Phase 4: Media Establishment and Active Call

Mobile-Terminated (MT) VoNR Call Flow Differences

Call Termination / Release Flow

Key Optimizations and 5G-Specific Enhancements in the Flow

Comparison: VoNR vs. EPS Fallback Call Flow

Implications and Broader Context

RTP (Real-time Transport Protocol)

RTP: Real-time Transport Protocol – The Media Carrier

RTCP: RTP Control Protocol – The Quality Feedback Mechanism

RTP/RTCP in VoNR vs. VoLTE: Similarities, Differences, and Optimizations

Implications for Voice Quality and Network Design

VoNR codec details

Core Technical Characteristics of EVS

SDP Negotiation and Configuration in VoNR

VoNR vs. VoLTE: Codec Usage Nuances

Additional Features and Optimizations in VoNR Context

EVS Channel-Aware Mode (CAM)

How Channel-Aware Mode Works: Core Mechanism

SDP Negotiation and Configuration in VoNR/VoLTE

Performance Benefits and Comparisons

Broader Considerations in VoNR Context

Comparision of VoNR with VoLTE codecs

Primary Codecs in VoLTE and VoNR

Codec Support and Mandates

Detailed Codec Comparison: EVS vs. AMR-WB (the Main Practical Difference)

Codec Negotiation and SDP Handling

Voice Quality and Performance Implications

Broader Considerations and Edge Cases

Benefits and Implications of VoNR (Voice over New Radio / Vo5G)

1. User Experience Benefits

2. Technical and Network Performance Benefits

3. Operator and Economic Benefits

4. Battery, Device, and Ecosystem Implications

Challenges, Trade-offs, and Edge Considerations

ViNR (Video over New Radio / Video over 5G)

Architecture and Key Components for ViNR

Detailed ViNR Call Flow (High-Level, Building on VoNR)

Video Codecs in ViNR: H.265 vs. H.264 (and Comparison to ViLTE)

Benefits and Implications of ViNR

Share this:

Like this:

Related

Leave a Reply Cancel reply