Introduction to Voice technologies
1
Voice over IP introduction • VoIP = Voice + IP • VOICE Traditionally, voice was transmitted using a separate dedicated infrastructure and it is still in place i.e. PSTN The first network that was put in place was for voice ONLY. Based on TDM
2
Voice over IP introduction (contd..) • VoIP = Voice + IP • TCP/IP based Data Networks Most common data network implementations are based on TCP/IP. Internet and most business networks are also based on TCP/IP. The purpose of data networks is to transfer & share computer data between users 3
Voice & Data Network infrastructure
• VOICE
• DATA
Circuit Switching
Packet Switching
Phones/terminals
Data Terminals
Signaling
Signaling
Routing
Routing
Transmission facilities
Transmission facilities 4
What is meant by Data? • Computer Data • Voice • Video • What is common in all of them? They can all be represented as bits i.e. these are all different forms of information As all can be represented as digital data making Voice/Video/Data integration possible 5
Voice technologies
• Voice in PSTN (TDM Based) • Voice over Packet (VoIP, VoFR or VoATM) 6
Voice over IP (contd..) • Transport voice traffic using IP • Voice over the Internet? Interconnected networks Applications: e-mail, file transfer, e-com
• The greatest challenges Voice quality and bandwidth Control and prioritize the access
• Internet: best-effort transfer The next generation VoIP != Internet telephony
7
IP (Internet Protocol) • A packet-based protocol Routing on a packet-by-packet base
• Packet transfer with no guarantees May not receive in order May be lost or severely delayed
• TCP/IP Retransmission Assemble the packets in order Congestion control Useful for file-transfers and e-mail
8
Voice over IP Protocols Presentation
G.729(A)/G.723(.1)/G.711
Session
H.323/MGCP/SIP
Transport
RTP/UDP/RSVP
Network
IP/WFQ/IP-prec
Link Physical
MLPPP/FR/ATM AAL1 –––
9
Why VoIP? • Why carry voice? Internet supports instant access to anything “Dot-com” Many new services and applications However, voice services provide more revenues
• Why use IP for voice? Circuit-switching is not for datacom IP-based Packet switching: Equipment cost, integrated access, less bandwidth, and widespread availability 10
Lower Equipment Cost – PSTN switch Proprietary – hardware, OS, applications High operation and management cost Training, support and feature development cost – Mainframe computer – The IP world Standard hardware and mass-produced Application software is quite separate – IN does not match the openness and flexibility of IP A few highly successful services 11
Voice/Data Integration – Click to talk application Personal communication E-commerce CTI – Computer Telephony Integration – Web collaboration Shop on-line with a friend at another location – Video conferencing – IP-based PBX – IP-based call centers 12
Enterprise Voice Over IP Applications
• Toll bypass Most common application
• PBX extension Saves costs by reducing maintenance costs and overhead
• H.323 interoperability Supports voice-enabled Web applications
13
Cisco “Voice over” Applications
14
Connection Types
• Local • On-net • Off-net • PLAR • PBX-to-PBX • On-net to Off-net 15
Local Connections
555-4001
Between two FXS Stations
555-4002 16
On-net Connections
Site B
Site A IP Router Gateway
Router Gateway
Calls within an enterprise
17
On-net Connections (contd..)
Branch A
Branch B Soft Phone
192.168.1.1 192.168.1.254
172.16.1.254
Internet
IP Phone
18
Off-net Connections Dial Access code: 9 Then PSTN number
Branch A
Branch B
192.168.1.1 192.168.1.254
172.16.1.254
FR/ATM
PSTN
19
Tie Line Trunks
PBX
PBX
Router Gateway
IP,FR ATM
Router Gateway
20
On to off-net Connections
Branch A
Branch B
PSTN
192.168.1.1 192.168.1.254
172.16.1.254
Internet
21
Toll Bypass Using 3600
PBX PBX
PSTN
3620
V
4 to 12 Analog ports
QoS WAN (Intranet) 3640
Branch Office
V
Headquarters 22
Introduction to PSTN Legacy Voice Infrastructure
23
Addressing in Telephone Systems • Numbering is never flat, it is always hierarchical • E.163 Standard (replaced by E.164) • E.164 ITU-T standard for ISDN numbers • In switching terminology the numbers are termed as DNs or (Directory Numbers) 24
Dialing Types
• Pulse Each digit is represented as a series of pulses.
• Touch Tone (DTMF) Each digit represented as a pair of frequencies 25
Pulse Dialing Scheme
Make = Circuit Closed
Off-Hook
Dialing Inter-Digit Delay Next Digit
Break = Open Circuit 700 ms
Pulse Period (100 ms)
Supported on Cisco routers
26
DTMF Dialing Supported on Cisco routers Dual Tone Multifrequency (DTMF) 1209
1336
1477
1633
697
1
2
3
A
770
4
5
6
B
852
7
8
9
C
941
*
0
#
D
27
Types of circuit switched calls
Call on same switch a calling b
Call established through multiple switches c calling d End-Office Central Office Local Exchange CLASS 5 switch
Tandem for calls within city Transit for calls out of city 28
Introduction to Signaling The main purpose of Signaling is to setup and tear down a call and providing supervisory functions. Signaling Classification
Off-hook Dial-tone Ringing Busy Tone Hookflash ISDN Q.931
Subscriber Signaling
Trunk or Inter-switch SS1-6 SS7 Signaling Router-Router R2 (Analog / PCM H.323 / SIP MGCP
29
Types of Signaling Method of communicating telephony events: Off-hook, busy, on-hook…
Analog • 2-wire • Loop start • Ground start • • • •
E&M 2-wire, 4-wire Five types I-V (Cisco I,II,III,V)
Digital • Digital subscriber lines: 2-wire, 4-wire • Digital trunks: 4-wire • Channel associated signaling (CAS) • In-band signaling • Common channel signaling (CCS) • Out-of-band signaling 30
Basic Local Call Flow
31
Subscriber signaling for local calls
32
Basic Call Progress: On-Hook Telephone Switch
Local Loop
Local Loop
-48 DC Voltage DC Open Circuit No Current Flow
33
Basic Call Progress: Off-Hook Off-Hook Closed Circuit
Telephone Switch DC Current Dial Tone Local Loop
Local Loop
34
Basic Call Progress: Dialing Off-Hook Closed Circuit Dialed Digits Pulses or Tones
Telephone Switch
DC Current Local Loop
35
Basic Call Progress: Switching Off-Hook Closed Circuit
Telephone Switch
DC Current Local Loop
Address to Port Translation
Local Loop
36
Basic Call Progress: Ringing Off-Hook Closed Circuit Ring Back Tone DC Current Local Loop
Telephone Switch DC Open Cct. Ringing Tone Local Loop
37
Basic Call Progress: Talking Off-Hook Closed Circuit
Telephone Switch
Voice Energy DC Current Local Loop
Voice Energy DC Current Local Loop
38
Common Terms
• Local Loop • Switches • Trunks
39
Switch Types • Local Exchange / CO • PBX • Tandem • Transit Switches solve the N² problem
40
Trunk Types
• Private Trunks • CO Trunks • FXO Trunks • FXS Trunks • DID/DOD Trunks • Inter-office trunks 41
2-to-4 wire conversion
• Done in Telephone Set • Done on Switch side as well Result: ????
42
Speech-Coding Techniques
43
Introduction
• Codecs / Speech coding schemes • Subjective impairment analysis: MOS • Digitizing voice • Voice compression ADPCM CELP Silence Removal Techniques (DSI using VAD)
• Processing Power A balance between quality and cost
44
Voice Quality Measure • Bandwidth is easily quantified Voice quality is subjective • MOS, Mean Opinion Score ITU-T Recommendation P.800 Excellent – 5 Good – 4 Fair – 3 Poor – 2 Bad – 1 A minimum of 30 people Listen to voice samples or in conversations 45
ITU-T Voice Quality Standards P.800 recommendations
The selection of participants The test environment Explanations to listeners Analysis of results Toll quality
A MOS of 4.0 or higher
46
ITU-T Voice Quality Standards
• Subjective and objective quality-testing techniques • PSQM – Perceptual Speech Quality Measurement ITU-T P.861 algorithmic comparison between the output signal and a known input type of speaker, loudness, delay, active/silence frames, clipping, environmental noise 47
Voice Compression Technologies Unacceptable 64
Business Quality
Toll Quality *
PCM (G.711)
(Cellular)
Bandwidth (Kbps)
*
32
ADPCM 32 (G.726)
*
24 16 8 0
ADPCM 24 (G.726)
*
ADPCM 16 (G.726)
*
LPC 4.8
*
LDCELP 16 (G.728)
*
CS-ACELP 8 (G.729)
Quality 48
Speech Waveforms & PSD
• Voiced speech
• Power spectrum density 49
Speech Waveforms & PSD (contd..)
• Unvoiced speech
• Power spectrum density 50
Type of Speech Coders • Waveform codecs Sample and code High-quality and not complex Large amount of bandwidth
• Source codecs (Vocoders) Match the incoming signal to a mathematical model Linear-predictive filter model of the vocal tract The information is sent rather than the signal Low bit rates, but sounds synthetic Higher bit rates do not improve much
51
Types of codecs • Hybrid codecs Attempt to provide the best of both Perform a degree of waveform matching Utilize the sound production model Quite good quality at low bit rate
52
Encoding
Quantizing
Sampling
Filtering
Waveform Coders
Waveform ENCODER
1110010010010110
Waveform DECODER
53
PCM Encoder
PCM Decoder
111001001001011
10110010
Parameters
Model Parameters
Sample Frames
Model Parameters
Encoding
Quantizing
Sampling
VocalCords Throat Nose Mouth
Filtering
Vocoders
Human Speech Model
Analysis
Synthesis 54
Voice Digitization • Analog-to-Digital Conversion discrete samples of the waveform and represent each sample by some number of bits A signal can be reconstructed if it is sampled at a minimum of twice the maximum freq.
• Human speech 0-4KHz (300-3400 Hz used in telephony) 8000 samples per second 55
Digitizing Voice: PCM Waveform Encoding • Nyquist Theorem: sample at twice the highest frequency Voice frequency range: 300-3400 Hz Sampling frequency = 8000/sec (every 125us) Bit rate: (2 x 4 Khz) x 8 bits per sample = 64,000 bits per second (DS-0)
• By far the most commonly used method CODEC PCM = DS-0 64 Kbps
56
G.711 •
The most common codec Used in circuit-switched telephone network PCM, Pulse-Code Modulation
•
Uniform quantization (not done) 12 bits * 8 k/sec = 96 kbps
•
Non-uniform quantization 64 kbps DS0 rate mu-law North America & Japan A-law Other countries, including Pakistan A MOS (Mean Opinion Score) of about 4.3
57
DPCM
•DPCM, Differential PCM Only transmit the difference between the predicated value and the actual value Voice changes relatively slowly It is possible to predict the value of a sample based on the values of previous samples The receiver performs the same prediction The simplest form • No prediction 58
ADPCM
•
ADPCM, Adaptive DPCM Predicts sample values based on Past samples Factoring in some knowledge of how speech varies over time The error is quantized and transmitted Fewer bits required G.721 32 kbps G.726 A-law/mu-law PCM -> 16, 24, 32, 40 kbps An MOS of about 4.0 at 32 kbps
59
CELP
• Code excited linear predictive Hybrid coding scheme
• Very high voice quality at low bit rates, processor intensive, use of DSPs • G.728: LD CELP—16 Kbps Smaller Codebook
• G.729: CS ACELP—8 Kbps G.729a variant— “stripped down” 8 kbps (with a noticeable quality difference) to reduce processing load, allows two voice channels encoded per DSP
60
G.729 an Advanced CODEC Cake
Code Excited Linear Prediction (CELP) Consumes ~ 8 Kbps A/D
Code 16-Bit Linear PCM
• DSP = Digital Signal Processing
DSP
Cake Recipe $0.32 10.1.1.1
Code Look-Up
Ingredients: A-sound K-sound
Packet
Directions: Play K, A, and K
Recipe or Code Book 61
G.729x • G.729.B VAD, Voice Activity Detection Based on analysis of several parameters of the input The current frames plus two preceding frames DTX, Discontinuous Transmission Send nothing or send an SID frame SID frame contains information to generate comfort noise CNG, Comfort Noise Generation
• G.729, an MOS of about 4.0 • G.729A an MOS of about 3.7 62
Digital Speech Interpolation (DSI)
• Voice Activity Detection (VAD) • Removal of voice silence • Examines voice for power, change of power • Automatically disabled for fax/modem
63
Bandwidth Requirements
Voice Band Traffic Encoding/ Compression G.711 PCM A-Law/u-Law G.726 ADPCM
Result Bit Rate 64 kbps (DS0)
16, 24, 32, 40 kbps
G.729 CS-ACELP
8 kbps
G.728 LD-CELP
16 kbps
G.723.1 CELP
6.3/5.3 kbps Variable
64
Voice Quality Comparison Anything Above an MOS of 4.0 Is “Toll” Quality Compression Method
MOS Score
Delay (msec)
64K PCM (G.711)
4.4
0.75
32K ADPCM (G.726)
4.2
1
16K LD-CELP (G.728)
4.2
3–5
8K CS-ACELP (G.729)
4.2
15
8K CS-ACELP (G.729a)
3.6
15
65