As LTE networks start to mature across the world, more and more carriers are looking to introduce voice services on LTE networks using a technology called Voice over LTE (VoLTE). However, making sure the VoLTE user experience is as good as—or preferably better than—legacy 2G/3G voice services is very challenging.
This SlideShare presentation cover the following:
- How VoLTE is different from legacy wireless voice services
- The VoLTE user experience, and how to measure it
- Key elements of a successful test and verification strategy
3. 3
Sources
Facts
Q1 2013: 6.4 billion mobile
subscribers, with smartphones
comprising 50% of new mobile
device sales
Voice and SMS represent
approximately 70% of all global
wireless revenues
Voice-over-LTE (VoLTE)
revenues estimated to reach $2
billion by 2016
Trends
Highly competitive market,
ARPU no longer increasing
HD Voice is one way to seek
competitive advantage
(73 networks so far)
Voice and other services
moving to more efficient all-IP
LTE networks
• Ericsson Mobility Report – On the Pulse of the Networked Society – June 2013
• Global Mobile Suppliers Association – Mobile HD Voice: Global Update Report – June 4th, 2013
4. 4
• How can I compare the VoLTE voice experience offered
by different suppliers and technology options?
• How can I know that new infrastructure will deliver
high-quality VoLTE services, before it goes live?
• How can I ensure that my devices offer the VoLTE voice
experience my customers expect?
• How can I ensure that my network offers the VoLTE
voice experience my customer expect?
• How does the VoLTE Voice experience on my network
and devices compare to my competitors?
VoLTE: Better or Worse?
8. 8
QoS RAN IMS Codec
Dedicated vs.
Non-Dedicated
Bearers
Semi-persistent
scheduling
resource
allocation
Session Initiation
Protocol (SIP)
HD Voice
(Wideband AMR)
Quality of Service
Class Identifier
(QCI)
Transmission
Time Interval
(TTI) Bundling
Policy and
Charging Rules
Function (PCRF)
Dynamic
scheduler in
eNodeB
Robust Header
Compression
(RoHC)
Real-time
Transport
Protocol (RTP)
Real-time
Streaming
Protocol (RTSP)
Key Enablers of VoLTE Fall Into Four Categories
10. 10
QCI Sets Specific Packet Loss & Delay Targets
Quality of Service (QoS) Class Identifier indicates max delay &
packet error rates for each LTE bearer
Various network functions rely on the QCI to make prioritization
& resource allocation decisions
11. 11
eNB scheduler dynamically
allocates resource blocks
(RBs) across all users.
QoS delay budgets are a key
factor in allocations
eNB Dynamic Scheduler Grants Resources
12. 12
Semi-Persistent Resource Allocation for VoLTE
Semi-persistent scheduling reduces the signaling overhead for RB
allocation by granting periodic use of a set of RBs
VoLTE sends short packets on a regular basis: SPS ensures
resources are available which match the periodicity of VoLTE
13. 13
16QAM Signal Constellation
(4 bits per symbol)
64QAM Signal Constellation
(6 bits per symbol)
Link adaptation manages the packet
loss rate by adjusting the rate at which
bits are transmitted up or down based
on radio conditions (SNR).
MoreBits/s&HigherSNR
Link Adaptation is Key to Packet Loss
15. 15
IP Multimedia Subsystem
Session Initiation Protocol (SIP) Session
Bit rate
Packet size
Packet transport frequency
RTP payload
Bandwidth adaptation
Policy and Charging Rules Function (PCRF)
Policy rules (bandwidth, quality class, IP packet filters)
PDN Gateway (P-GW)
Interpret rules and establish EPC dedicated bearers for
voice
16. 16
HD Voice (WB-AMR Codec)
Type of Channel Channel Bandwidth Type of Service
Narrowband 50Hz – 3.8KHz Narrowband Voice
Wideband 50Hz – 7.5KHz HD Voice (including VoLTE)
Super-wideband 20Hz – 14KHz HD Voice (including VoLTE)
Channel bandwidths for different voice services
17. 17
THE VOLTE USER EXPERIENCE
The VoLTE User Experience: Better or Worse?
18. 18
End User Experience Key Performance
Indicators (KPIs)
Ability to make and maintain calls Call Initiation Rate (%), Call
Drop Rate (%)
Time it takes for a phone to start ringing Call Setup Time (s)
Speech quality during a call Mean Opinion Score (MOS)
Mouth-to-ear delay or latency variations
during a call
Mouth-to-ear delay (s),
Latency, jitter
What to Measure?
19. 19
Packet Loss & Delay Drive QoE for VoLTE
VoLTE packet loss directly impacts
speech quality
Source: “Validating voice over LTE end-to-end”, Ericsson Review, January 2012.
Speech Quality vs. Frame Error Rate
Packet loss rate of < 1%
results in good speech
quality on avg.
20. 20
Packet Loss & Delay Drive QoE for VoLTE
VoLTE packet delay impacts mouth-
to-ear delay, and packet delay
variability (jitter) impacts both
speech quality & mouth-to-ear delay
Speech Quality
vs. Mouth to
Ear Delay
Source: “Validating voice over LTE end-to-end”, Ericsson Review, January 2012.
ITU recommended end-
to-end mouth-to-ear
delay is <200 ms
21. 21
UE-Specific Delays Dominate Mouth-to-Ear Delay
Packet delay is
strongly
impacted by LTE
scheduling &
HARQ:
• Average E2E
transport delay
• Jitter buffer
delay
Breakdown of mouth-to-ear delay for VoLTE lab
and field tests performed by Ericsson
Source: “Validating voice over LTE end-to-end”, Ericsson Review, January 2012.
23. 23
PESQ
(ITU-T P.862)
POLQA
(ITU-T P.863)
Codecs AMR
EFR
AMR
AMR-WB
EFR
EVRC
EVRC-B
EVRC-WB
Reference
Speech
Material
(sampling
frequency)
8 kHz 8 kHz
48 kHz
Applications POTS
VoIP
3G
HD Voice
Voice
Enhancement
Devices
POLQA has more robust quality
predictions for …
Cross-technology quality
benchmarking (e.g., GSM vs. CDMA)
Noise reduction and voice quality
enhancement
Time-scaling, unified
communication and VoIP
Filtering and spectral shaping
Recordings made at an ear
simulator
Perceptual Evaluation of Speech Quality (PESQ) vs.
Perceptual Objective Listening Quality Analysis (POLQA)
24. 24
Downlink MOS Uplink MOS
Device Device
A
Device
B
Device
C
Device
D
Device
A
Device
B
Device
C
Device
D
Average 3.09 3.34 3.16 3.62 3.46 3.81 3.44 3.31
Standard
Deviation
0.42 0.08 0.39 0.29 0.22 0.03 0.27 0.12
Maximum
Score
3.56 3.45 3.43 3.84 3.65 3.85 3.83 3.45
% MOS less
than 3.0
33% 0% 17% 0% 0% 0% 0% 0%
POLQA MOS Score Comparison for VoLTE
Devices
Source: Spirent Testing on Live Network with Nomad HD
25. 25
Call Initiation and Setup Time Comparison for
VoLTE Devices
Source: Spirent Testing on Live Network with Nomad HD
26. 26
Call Drop Comparison for VoLTE Devices
Source: Spirent Testing on Live Network with Nomad HD
29. 29
Measuring device performance across multiple OS’s
and technologies
Testing in both live and simulated network
environments
Evaluating multiple devices simultaneously
Managing extensive testing projects from a
centralized location
Measuring how a device’s performance will
impact subscribers
Comparing VoLTE voice quality to circuit switched voice: Is it as good
or better?
Common VoLTE Test Challenges
30. 30
Metrics that focus on what the end-users experience (including speech
quality, the ability to make and maintain calls, and mouth-to-ear latency)
Use the same voice service measurement systems in the field and in the
lab
Measurement systems that can test any device, on any network,
anywhere in the world and still provide one central location for
results collection and analysis
Lab test solutions that provide simple interfaces for LTE and VoLTE
configuration while also enabling fast creation of automated VoLTE tests
Coverage of relevant compliance tests originating from operators and
standards organizations
Characteristics of Ideal Test Strategy
32. 32
Spirent Believes in “Better”
How can we help you improve VoLTE
quality and time-to-market?
33. 33
For More Information
Spirent’s VoLTE and HD Voice web page:
• Application Notes for CS8 and Nomad HD
Voice Quality testing
• Video demonstrations
http://www.spirent.com/go/VoLTE
LTE air-interface is a key contributor to end-to-end packet loss and delayThe level of jitter requires an increased jitter buffer size – that increases the delay and may lead to dropped packets for excessively delayed packets
LTE air-interface is a key contributor to end-to-end packet loss and delayThe level of jitter requires an increased jitter buffer size – that increases the delay and may lead to dropped packets for excessively delayed packets