Transcript Voice Quality

Voice Quality (Special Topic)
ISSUE1.0
Fixed Network Technical
Support Dept.
Introduction
HUAWE

Please enumerate the voice quality faults during the
deployment
maintenance
troubleshooting.
and
describe
briefly
the
References
HUAWE

Support website—engineers documentation
Guide to NGN Voice, Fax and Modem Configuration--20040729--B.doc
NGN Bearer Network Grade and Definition of VoIP Quality
HUAWE
Definition
of bearer network quality grade
Network Grade
Unidirectional Delay (ms)
Packet Loss Ratio
Jitter (ms)
Good (customized)
0~40
0~0.1%
0~10
Relatively poor*
40~100
0.1%~1%
10~20
Bad*
100~400
1%~5%
20~60
* Indicates that a grade is defined based on the communication industry standard YD/T 1071--2000 Technical
Requirements for IP Telephone Gateway Equipment
Definition
Grade
of VoIP Voice Quality
MOS
Users’ Satisfaction
Superior
4.0--5.0
Very good, able to be heard clearly, low delay, smooth communication.
Good
3.5--4.0
Slightly poor, able to be heard clearly, low delay, not smooth communication, some noise
Medium
3.0--3.5
Medium, unable to be heard clearly, with certain delay; communication is possible.
Poor
1.5--3.0
Poor, unable to be heard clearly, great delay, repeated communication.
Bad
0--1.5
Bad, unable to be understood, great delay, not smooth communication.
According to the communication industry standard YD/T 1071--2000 Technical Requirements for IP Telephone
Gateway Equipment
Service Performance in Different Network Conditions
HUAWE
Service Type
Modem
Transparent
Voice
Video
Two-stage
dialing
Message
Performance
Bad Network Service
Poor Network Service Performance
Performance
Available
Unavailable
Unavailable
Available
Unavailable
Unavailable
T.38
Available
Available
Unavailable
G.711 a/u
Superior
Good
Medium
G.729 a/b
Good
Good
Poor
G.723
Good
Close to good
Medium
384K
No deterioration, available
Slight deterioration, available
Available, good voice quality
Available, voice quality close to
Available, as poor voice
or better
good
quality as medium
Available
Available
Available
transmission
Transparent
Fax
Good Network Service
transmission
RFC2833
Apparent deterioration,
unavailable
The red part with oblique lines as the background is the service performance
unable to satisfy operation requirements
Key Factor Which Affects Voice Quality—Network Delay
HUAWE
Network delay:
Delay means that the less time difference for end-to-end packet transmission, the better.
Unit: ms.
Generally, if delay exceeds 100ms, we will feel in our conversation that the peer does not
speak naturally and reacts slowly; if delay exceeds 250ms, we will feel unbearable in our
conversation.
Delay results from the following:
1) End-to-end voice delay = encoding delay+compression & packetization delay+network
transmission delay+ unpacking & decompression delay +decoding delay.
2) Buffer (jitterbuf) set to eliminate network jitter.
Key Factor Which Affects Voice Quality—Network Jitter
HUAWE
Network jitter:
Jitter refers to delay offset. Unit: ms.
Network jitter refers to change of IP packet transmission time.
Jitter indicates how steady packet transmission is. The transmission mode of IP network is
packet-based store-and-forward and packets are transmitted at a fixed rate. Besides, voice is
sensitive to delay and those voice packets with great jitter will generally be discarded. In this
case, jitter may bring about voice distortion.
Network jitter is normally distributed. The jitter value we generally call is the root-mean-square
value of a normal distribution curve and the maximum actual jitter is much greater than the rootmean-square value. Therefore, if the network jitter reported via RTCP is 60ms, voice quality will
not be improved until jitter buffer is increased to 120ms and good voice quality will not be
achieved until jitter buffer is increased to 160~180ms.
Key Factor Which Affects Voice Quality—Network Packet Loss
HUAWE
Network packet loss ratio:
Packet loss ratio refers to the ratio of lost packets to all packets and is represented with x%
When any packet loss occurs to network, the lost-packet compensation algorithm will be
employed to avoid any influence brought about by packet loss. According to the correlation
between previous and subsequent voice information, lost frames are reconstructed during the
decoding to ensure acoustic effects of received voice; it is stipulated in ITU—T that the
maximum compensation time of PLC should be 60ms.
If packet loss exceeds the compensation range of a device, noise, voice distortion or call drop
will occur.
HUAWE
Key Factor Which Affects Voice Quality—Network
Bandwidth
Network bandwidth:
It is an index used to measure network flow.
If different encoding/decoding methods are used, bandwidth occupied will be different. If
there is heavy call traffic, but insufficient network bandwidth in the system, delay and jitter
will occur and the packet loss index will sharply decrease. This will have a big, even severe,
influence upon voice quality.
Key Factor Which Affects Voice Quality—Selected CODEC
HUAWE
CODEC:
One of the key technologies of VoIP service is the voice coding & compression technology.
The encoding/decoding algorithm and compression technology will have a direct influence
upon voice quality of VoIP service (voice encoding & compression is implemented in a
gateway. In this process, digital coding is first needed and converted into PCM codes. Then, a
special DSP chip will be used for data compression. Finally, IP packet data are formed so as
to adapt to transmission bandwidth over IP network.
Relationship between encoding/decoding rate and voice quality: the higher rate, the better
voice quality.
Key Factor Which Affects Voice Quality--EC
HUAWE
EC:
Echo indicates that a speaker’s voice goes through network equipment and loops back to himself.
Features of echo: echo affects the perception quality of the speaker instead of that of the receiver;
If IP is interconnected with PSTN and either of both communication parties uses a two-wire telephone,
which involves a 2-/4-wire conversion circuit with coils, echo will be heard; echo can be perceived by
people in two conditions: strength: some signals should be fed back (leakage, sound wave reflection) and
there should be certain amplitude; delay: there should be enough delay. ITU--T G.131 and ITU--T G.161
point out that when the echo delay exceeds 25ms, echo can be perceived by people. Therefore, an Echo
Controller (EC) should be equipped. Generally based on the DSP technology, the EC employs an Finite
Impact Response (FIR) filter to simulate an echo path and generate echo signals, so as to subtract actual
echo signals and offset echo.
Handling of the Above-mentioned Key Factors
HUAWE
Countermeasures for network bandwidth:
1. Give consideration to equipment capacity in network design, select proper encoding/decoding and as great
bandwidth as possible at the egress.
2. When there are both voice service and data service on an IAD, the IAD may be configured with the QoS
strategies of voice service to ensure voice service bandwidth.
3. Adopt Voice Activity Detection and Comfort Noise Generation on the device to reduce bandwidth occupancy.
Countermeasure for delay:
1. The bearer network itself reduces network delay.
Countermeasure for jitter:
1. The bearer network itself reduces network jitter; 2. Set the jitterbuf on the device
Countermeasure for packet loss:
1. Employ the bearer network to reduce packet loss; 2. Equip the device with the lost-packet compensation
algorithm.
Questions
HUAWE

What factors will affect voice quality of NGN?

Give examples of how to check the quality of a bearer
network.
Summary
HUAWE

Different kinds of service (voice, modem, and fax) present different
requirements for network. In handling a corresponding problem, we should
know clearly about the requirements of various kinds of service for network
quality.