Cisco Voice over IP (CVOICE)

Commonly POTS or VoIP, required for inbound and outbound call legs.

Outbound Core Elements:
- Destination pattern (digit matching)
- Port (POTS DP) e.g. 0/1/0
- or session target ipv4/6:x.x.x.x (VoIP destination)

Example POTS dial-peer for FXS port 0/2/0 with ext. 4444

!
dial-peer voice 4444 pots
description DP FXS Ext 4444 Reception
destination-pattern 4444
port 0/2/0 

Example VoIP dial-peer for extensions. 5555 and 5556 over WAN to router with loopback 192.168.2.1

!
dial-peer voice 100 voip
description DP VoIP SiteA Ext. 555[5-6]
destination-pattern 555[5-6]
session target ipv4:192.168.2.1 

Example FXS voice port configuration for UK phone handset:

voice-port 0/2/0
description VP-FXS | Mail Room | +441224987654
signal loop-start
cptone GB (sets call progress tone specific to country code)
ring-cadence pattern# (not required if cptone is configured)
busy-out (effectively shut port down, continuous busy tone)
station-id name Mail Room 7654 (name send to called party)
station-id number 7654 (number sent to called party)

Example configuration of FXO port for BT PSTN line

voice-port 0/1/0
description VP-FXO | BT PSTN | +441224567890
signal loop-start 
ring number 3 (postpones answer until 3rd ring)
dial-type dtmf (determines dialing type from calling party)

Codecs used to compress and encode voice transmissions are not effective when the stream is a fax transmission.  As the stream is digitised data the codec can cause adverse affects.

• Cisco Fax Relay (pre-standards), detects a fax call and downgrades the stream to an appropriate size for transmission for the IP network.  DSPs are required for conversion from analog to digital.
• T.38 fax relay - standards based transmission over IP.  DSPs convertor the stream into an appropriate format, e.g. 9600bps group 3.  T.38 is non-propriatory, devices such as servers can support sending/recieving if they meet the T.38 standard.
• T.39 store and forward - supports the above, but with the ability to digitally scan/email.  Devices are configured as on/off ramps depending on the direction of the transmission.
• Fax Pass-through - treated as a normal call with G.711 codec, 64Kb/sec, without VAD.  Fax transmission is suceptable to packet loss in the IP network, which could break transmission.

CCS uses a common channel for all signalling. 

The signalling protocols that can be used with CCS are:

ISDN - Q931  - BRI (2B+D) and PRI 23/30+D varients,
Q Signalling (QSIG) - based on Q931, often used for propriatory PBX/PBX connections
Digital Networks Private Signalling System (DNPSS) - created by BT prior to ISDN standards, typically used in UK
Signalling System 7 (SS7) - typically used on PSTN, signalling can be undertaken before seizure of circuit, high speed and packet based signalling system.
SIGTRAN - transportation of SS7 signals end-to-end over an VoIP network

Waveform algorithms:
- Sample at 8000 samples/sec
- Use prediction techniques
- G.711/G.726

Source algorithms:
- Vocoders/voice coders
- Utilise speech characteristics
- Predict human speech waveforms using speech codebook
- G.728/G.729

Pulse Code Modulation (PCM), takes 8000 samples/sec, either a/u-law

Adaptive Differential PCM (ADPCM), looks at different between current and predicted sample

Code Excited Linear Prediction (CELP) predicts the next wave form and synthesises and generates the waveform with the speakers characteristics.

G.729 uses CS-ACELP, G279A is a variation that allows for less CPU intensive processing, G279B includes VAD to elimiate transmission of 'silence' within a voice transmission.

Analogue to digital conversion uses the following steps:

1. Sample (snapshot) the analog signal at high frequency intervals
2. Match the signal to a specific step, identifying a chord
3. Encode the chords determined above, into an 8bit value (binary form)


Most spoken voice falls within the 300-3400Hz frequency, Nyquist theorem determines the number of samples required to quantise voice, the number of samples required is twice the highest frequency.  Nyquest determined the highest frequency for spoken voice is 4000Hz, therefore 8000 samples/sec would be required.  Encoding 8000 samples into 8bit values generates 64Kb/sec.


The G711 codec takes 8000 samples/sec and encodes them into 8bit values, streaming voice at 64kb/sec.


There are two variants of the G711 codec, a-law and u-law (u-law is used in the US, a-law is used elsewhere).    The variation is specific to the quantisation levels used.

Two types of echo exist, acoustic and electrical:

• Acoustic when the received voice is caught by the microphone and sent back to the reciever

• Electrical is when there is an inconsistancy in the telephone circuits, called impedence mismatch

Echo suppression - transmitting speak in the forward direction, whilst preventing it in the return direction, the return path is broken to prevent feedback.

Echo cancellation is much more sophisticated, it analyses the transmitted voice and subtracts it in the return path. Rather than stopping all return transmission, it removes specific parts.