Voice over LTE (VoLTE)

1. Voice in the LTE All-IP Network

LTE (Long Term Evolution) was designed as an network. Unlike previous generations (2G and 3G), the EPC does not contain a separate circuit for traditional voice calls. Voice in LTE therefore needs an IMS service or a transition to an earlier network.

Previous generations relied on a technology called for voice calls. In a 2G or 3G call, the network reserved a circuit between the two parties for the duration of the call. LTE uses packet transmission for the user session, so voice has to be described and handled as an IP service.

Early 4G deployments therefore used a transitional mechanism known as Circuit-Switched Fallback (CSFB) to connect voice calls to the legacy circuit-switched domain.

2. Circuit-Switched Fallback (CSFB)

Circuit-Switched Fallback was a transitional solution. An LTE-capable phone made and received traditional calls by moving to an older radio technology with a circuit-switched domain for the duration of the call.

How CSFB Works

A CSFB procedure for a phone using LTE data follows this sequence of events:

The MME receives an SGs paging request from the MSC for an incoming voice call and pages the UE on LTE.
The MME and eNodeB release or redirect the UE so that it can leave the LTE radio connection for the circuit-switched service.
The device radio moves to the operator's older 3G (UMTS) or 2G (GSM) network, which has circuit-switched infrastructure for voice calls.
Once connected to the 2G/3G network, the call is established and speech can begin.
During the call, the device remains on the 2G/3G network. Internet services may be suspended or operate at a rate determined by the selected network and device.
When the call ends, the phone disconnects from the 2G/3G network and re-attaches to 4G LTE to resume data services.

In 3GPP CSFB, the MME requests the service through the SGs interface toward the MSC. The eNodeB then releases or redirects the UE to GERAN or UTRAN, where the circuit-switched call is controlled by the MSC. The exact signaling and data behavior depend on the deployed 2G or 3G interworking option.

The Downsides of CSFB

CSFB functioned as a transition mechanism, but it introduced several technical drawbacks:

Longer Call Setup Times: The process of moving to another network and establishing the call can take longer than a call set up within one technology, delaying the first ring after dialing.
No Simultaneous Voice and LTE Data: During a CSFB call, the LTE bearer is released or suspended. Maps, web pages, and email therefore cannot use that LTE session while the circuit-switched call is active; packet data over the older radio technology may still be available when the terminal and network support it.
Inefficient Use of Network Resources: Maintaining both a modern 4G network and legacy 2G/3G networks purely for fallback is expensive and complex for operators. It also requires constant signaling between the different network cores.

The All-IP LTE network therefore needed an IMS-based voice service. That service is VoLTE.

3. VoLTE Voice Service in 4G

VoLTE stands for Voice over Long Term Evolution. It is an IMS-based voice service that carries signaling and speech media as IP packets over LTE. The phone does not need to move to an older technology when the operator offers VoLTE and the device is configured for it.

Technical Properties of VoLTE

Handling voice within the LTE architecture gives VoLTE the following technical properties:

Higher Call Quality (HD Voice): VoLTE can provide a wider speech band than traditional 2G/3G telephony when both endpoints and the network select a suitable . One supported codec is AMR-WB (Adaptive Multi-Rate Wideband). Narrowband telephony commonly covers about $300$ Hz to $3.4$ kHz, while AMR-WB can carry a speech band of about $50$ Hz to $7$ kHz.
Shorter Call Setup: The phone does not perform an inter-RAT fallback, so VoLTE can shorten call setup compared with CSFB. The actual time depends on IMS registration, signaling, and network load.
Simultaneous Voice and High-Speed Data: The phone remains on the 4G network during the call, so data transmission can continue in parallel. This allows navigation, web browsing, or an email attachment download while the voice session is active.
Possible Power Savings: CSFB requires the device to search for and maintain a second radio technology. Staying on LTE can reduce that extra work, although the result depends on radio conditions and device configuration.
Shared IMS Platform: IMS can also host RCS services such as messaging, file sharing, and video calling. Activating VoLTE alone does not enable RCS, because those services require a separate operator deployment and profile.

4. VoLTE Architecture and Operation

Delivering voice as packets requires session control, codec negotiation, and quality-of-service rules. VoLTE uses the , which separates call signaling from media transport.

IMS Session Control in VoLTE

IMS is a layer of network components above the standard LTE data network (the EPC). It manages multimedia sessions, including VoLTE calls, with SIP (Session Initiation Protocol). When a call is initiated, the device sends an "INVITE" message to IMS, which routes the session to the called party.

Quality of Service (QoS) in VoLTE

QoS is one difference between a carrier VoLTE profile and a typical over-the-top (OTT) VoIP app. An OTT app usually sends media over the public internet on a "best-effort" basis, where packets compete with other traffic. VoLTE uses IMS and EPC policy rules, but quality still depends on coverage, load, and the operator's configuration.

VoLTE uses IMS and EPC policy rules. For voice media, the network can establish a dedicated EPS bearer with a specific .

IMS signaling, including SIP messages, uses the characteristics of QCI = 5, usually on the signaling bearer.
Voice media, meaning the RTP packets carrying speech, uses the GBR characteristics of QCI = 1, often on a dedicated EPS bearer.

QCI = 1 describes voice media as GBR with a 100 ms packet-delay budget and a target packet-error loss rate of 10^-2 in the LTE profile. It does not give absolute priority in every situation: the scheduler still assigns radio resources, and channel conditions can degrade a call. QCI 5 has a different characteristic and is used for IMS signaling.

QCI identifies a profile carried in policy rules; it is not a throughput guarantee on every segment. The PCRF and PCEF bind the media flow to a bearer, while the eNodeB schedules that bearer under current radio conditions. A correct QCI configuration therefore does not replace coverage or radio capacity.

The default EPS bearer carries IMS registration and SIP signaling, while a dedicated bearer is established after policy control authorizes the media flow. The PCRF supplies rules, the PCEF enforces them in the gateway, and the eNodeB applies the radio scheduling decision.

RTP carries encoded speech frames, while SIP remains the session-control channel. Separating the two flows allows codec or media parameters to change without mixing them with call setup and termination messages.

SDP carried in SIP describes the codec parameters and media endpoints. RTP then transports the encoded frames after the session is accepted, so signaling and media remain separate during the call.

VoLTE Call Flow

A simplified sequence of events for a VoLTE call is:

IMS Registration: When VoLTE is provisioned and enabled, a capable phone registers with the IMS core after connecting to LTE. This establishes its identity and readiness to make and receive VoLTE calls and sets up a default EPS bearer for signaling.
Call Initiation: Dialing a number causes the VoLTE application to create a SIP INVITE message. It contains the called number and the audio codecs supported by the device.
Signaling Path: The SIP INVITE is sent over the QCI 5 signaling bearer to the IMS core.
Bearer Setup: The IMS core passes the service requirements to policy functions and the EPC, which can work with the eNodeB to set up a dedicated bearer with QCI 1 for voice media.
Call Alerting:The IMS network routes the SIP INVITE to the called party. The device receives the message, starts ringing, and returns a SIP "Ringing" message.
Call Answered: The called party answers the call. The device sends a SIP 200 OK message to the IMS network, which forwards it to the calling device.
Media Flow: Both phones are ready. Speech is digitized, compressed by the negotiated codec, for example AMR-WB, placed into RTP (Real-time Transport Protocol) packets, and sent over the voice bearer with the QCI 1 characteristics.
Call Termination: When one person hangs up, their phone sends a SIP BYE message. The network then tears down the QCI 1 dedicated bearer, freeing up the resources, and the call is terminated.