Sie sind auf Seite 1von 15

User Description, Adaptive Multi Rate

Contents
1 Introduction
1.1 Main changes in Ericsson BSS G11B
1.2 Main changes in Ericsson BSS G10A
1.3 Main changes in Ericsson BSS 09A

2 Capabilities

3 Technical Description
3.1 General
3.2 Inband Signalling
3.3 Codec Mode Adaptation
3.4 Multirate Configuration
3.5 DTX
3.6 Handover
3.7 Tandem Free Operation
3.8 Transcoder Free Operation
3.9 Related Statistics

4 Engineering Guidelines
4.1 Fine Tuning Codec Sets

5 Parameters
5.1 Main Controlling Parameters
5.2 Parameters for Special Adjustments
5.3 Value Ranges and Default Values

6 Concepts

Glossary

Reference List

1 Introduction

Adaptive Multi Rate (AMR) is a speech and channel codec for both half rate and full rate
GERAN channels and for UTRAN. By adapting the codec rate to the radio conditions the
speech quality is enhanced. At low C/I, a large amount of channel coding is applied and
less speech coding. When the C/I increases the speech coding is increased and the
channel coding is decreased.

Both the BTS (uplink ) and the MS (downlink) continuously measures the radio quality
(C/I) and based on these measurements the codec rate is changed. AMR requires support
in all network nodes, that is MSC, BSC, BTS and MS and AMR is only supported in cells
where all radio units are AMR capable.
1.1 Main changes in Ericsson BSS G11B
Impact on legacy counters changed.

1.2 Main changes in Ericsson BSS G10A


Possibility to set Initial Codec Mode introduced.

1.3 Main changes in Ericsson BSS 09A


Impact from A-Interface over IP with TrFO added.

2 Capabilities

The feature AMR makes it possible to offer an enhanced speech quality for AMR mobiles
in a network. The enhanced speech quality also provides better coverage at the edges of
the cell, thus making it possible to increase the coverage area. AMR also tolerates more
interference than the old speech codecs. This enables higher traffic loads which leads to
higher capacity, but at the cost of somewhat lower speech quality. The speech quality
enhancement is divided in two parts and can therefore be of different interest to the
operator. The first is the robust FR channel that provides high speech quality at low C/I.
This makes it possible to tighten the cell planning in a network with AMR mobiles. The
second is an audible better HR codec than HR speech version one. This will increase the
capacity in the networks and reduce transmission costs.

3 Technical Description

3.1 General
AMR makes it possible to change codec rate during a call. There are 8 different codec
rates available in the AMR transcoder:

 12.2 kbps
 10.2 kbps
 7.95 kbps
 7.40 kbps
 6.70 kbps
 5.90 kbps
 5.15 kbps
 4.75 kbps

According to the standard, 12.2 kbps and 10.2 kbps can only be used in FR channels and
the other six can be used in both FR and HR channels, on the air interface.

Note:
Since 7.95 kbps is not specified for 8 kbps sub-multiplexing on the Abis
interface, this codec mode is not supported for HR channels.
At call setup and handover an Active Codec Set (ACS) and a channel rate is
selected. The ACS is a subset of the 8 (5 for AMR HR) codecs above and
may contain up to 4 codec modes, all with the same channel rate.

There are five codec sets for Full Rate (FR) and four codec sets for Half Rate
(HR) in the Ericsson system. It is possible to decide, by command, which
codec set that shall be used for each channel rate. This decision is made per
BSC.

Three of the FR codec sets and two of the HR codec sets are pre-defined.
The remaining two codec sets for each channel rate can be defined by the
operator.

With A-interface over IP the codec sets must be defined in accordance with
Preferred Configuration 1 in TS 28.062.

The use of the RATSCCH protocol to modify the codec configuration is not
supported.

AMR FR is also referred to as Full Rate Speech version 3 and AMR HR is also
referred to as Half Rate Speech version 3.

3.2 Inband Signalling


To be able to perform the codec mode adaptation the receiving side
performs quality measurements on the incoming link. This measurement
results in a Quality Indicator. The Quality Indicator produced on the MS side
is converted to a Mode Request that is sent to the BTS using inband
signalling. There, a codec mode control entity selects the codec mode to be
used on the downlink, based on the Mode Request sent by the MS.

The codec mode control entity also selects the codec mode to be used on
the uplink, based on the Quality Indicator produced in the BTS. The
selected uplink mode is sent inband to the MS. The presently applied codec
mode is also sent in band both uplink and downlink. This mode indication is
sent every second speech frame and the Mode Request (uplink)/Mode
Command (downlink) is sent in the speech frames in between, see
Reference [10].
Figure 1 AMR Functional Diagram

In band messages:

UL: Mode Request (2 bits per speech frame, giving a suggestion of a DL


codec mode)
UL: Mode Indication (2 bits per speech frame, giving the present UL codec
mode)
DL: Mode Command (2 bits per speech frame, telling the MS which UL
codec mode to use)
DL: Mode Indication (2 bits per speech frame, giving the present DL codec
mode)

3.3 Codec Mode Adaptation


At call setup, a channel rate is selected by the network. The codec set for
that channel rate is communicated to the MS and the BTS. The codec set
can consist of up to 4 different codec modes, all using the same channel
rate. It is not possible to have different channel rates within the same codec
set. The selection of codec mode within the codec set is based on
measurements of the uplink and downlink radio channel.

The codec modes in an Active Codec Set is defined as follows:

CODEC_MODE_1 - Represents the codec mode with the lowest bit rate

CODEC_MODE_2 - Represents the second lowest mode (if the ACS includes
more than one mode)

CODEC_MODE_3 - Represents the third lowest mode (if the ACS includes
more than two modes)

CODEC_MODE_4 - Represents the highest mode (if the ACS includes four
modes)

For each of the pre-defined codec set (two for HR and three for FR) there is
an associated set of decision thresholds that determine which codec mode
that should be used for a certain C/I (channel quality). To avoid continuous
rapid changes between 2 codec modes a hysteresis is used for the decision
thresholds.

A threshold (THR) is given as an absolute value between 0 and 63 which


represent normalized C/I values as follows:

THR 0 = 0.0 dB

THR 1 = 0.5 dB

THR 2 = 1.0 dB

::

THR 62 = 31.0 dB
THR 63 = 31.5 dB

1,2 or 3 thresholds needs to be defined depending on the number of codec


modes in ACS.

The hysteresis values (HYST) are given as an absolute value between 0 and
15 and are coded as follows:

HYST 0 = 0.0 dB

HYST 1 = 0.5 dB

HYST 2 = 1.0 dB

::

HYST 14 = 7.0 dB

HYST 15 = 7.5 dB

1,2 or 3 hysteresis values needs to be defined depending on the number of


codec modes in ACS.

Figure 2 Definition of Thresholds and Hysteresis

The figure above illustrates the definition of decision thresholds, hysteresis


and the operational range of codec modes. The lower decision threshold for
switching from a higher codec rate to a lower is given by one of THR_1,
THR_2 or THR_3. As an example, when C/I drops below decision threshold
THR_3, codec mode 3 (CODEC_MODE_3) will be in operation instead of
codec mode 4 (CODEC_MODE_4). The sum of the decision threshold and
hysteresis constitutes the upper threshold between two codec modes, e.g
when C/I exceeds THR_3 + HYST_3 codec mode 4 (CODEC_MODE_4) will
be in operation instead of codec mode 3 (CODEC_MODE_3). It is possible to
change codec mode every second speech frame but only to the closest
higher or lower codec rate in the set. Different codec modes (within the
ACS) can be used on the uplink and downlink. The change of codec mode
within a set is called fast codec mode adaptation.

Note:
The closer to ideal frequency hopping the better will the codec mode
adaptation for AMR work, see Reference [8].

3.4 Multirate Configuration

3.4.1 General

At setup and handover, the codec set to be used is sent from the
BSC to the MS and the BTS together with the decision threshold
and hysteresis values. The codec set (including its thresholds and
hysteresis values) is the same for the uplink and downlink. There
are three pre-defined full rate codec sets (FR1, FR2 and FR5) and
two pre-defined half rate codec sets (HR1 and HR2) in the BSC.
These codec sets with their associated threshold and hysteresis
values (described below) are hard coded and can not be changed.

The Initial Codec Mode (ICM) to be used at call setup and after
handover is sent from the BSC to the BTS and the MS. The initial
codec mode is settable with the MML command RLADC.

TFO is supported for any defined codec set, both for AMR FR and
for AMR HR. This is valid for both the pre-defined codec sets,
described below, as well as codec sets defined by the operator. For
more information TFO, see Reference [7].

3.4.2 A-interface over IP

When A-interface over IP is used in the BSC the codec sets


selected in BSC MUST be FR codec set 5 and HR codec set 2 (also
known as Preferred configuration 1 in TS 28.062 ).

It is also possible to use the User defined codec sets, if customized


settings are preferred, but they MUST include 4.75, 5.90, 7.40,
12.2 for FR then codec set and 4.75, 5.90, 7.40 for HR the codec
set, see also Reference [9].

3.4.3 FR Codec Set 1

Table 1 Full Rate Codec Set 1 (= Preferred Configuration 12 (TS 28.062))


CODEC_MODE_1 4.75 kbps
CODEC_MODE_2 5.9 kbps
CODEC_MODE_3 6.7 kbps
CODEC_MODE_4 10.2 kbps
This codec set has been selected because extensive evaluations
have shown that it provides a good overall speech quality during
all radio conditions (both for clean speech and with different
background noise), that is it provides a very small deviation from
optimum AMR performance (when all 8 codec modes are used). At
the same time it also allows the operator to take full advantage of
the possibility to use AMR as a capacity booster in the network.
This set was defined by the TFO Subgroup as preferred for AMR FR.

This codec set, together with the Half Rate Codec Set 1, see below,
satisfies the AMR FR and AMR HR operation for TFO connections.
With these two codec sets it will be possible to run TFO on the
three lower modes (6.7 - 5.9 - 4.75) since they are common.

Table 2 Decision Thresholds and Hysteresis Values for Full Rate Codec Set 1
Threshold Value Hysteresis Value
THR_1 12 (6 dB) HYST_1 2 (1 dB)
THR_2 16 (8 dB) HYST_2 3 (1.5 dB)
THR_3 21 (10.5 dB) HYST_3 3 (1.5 dB)

3.4.4 FR Codec Set 2

Table 3 Full Rate Codec Set 2 (= Preferred Configuration 14 (TS 28.062))


CODEC_MODE_1 4.75 kbps
CODEC_MODE_2 5.9 kbps
CODEC_MODE_3 7.95 kbps
CODEC_MODE_4 12.2 kbps

This codec set provides basically the same overall speech quality
as full rate codec set 1 but offers a slightly better speech quality
under clean speech conditions.

This codec set does not support optimal TFO to HR AMR.

Table 4 Decision Thresholds and Hysteresis Values for Full Rate Codec Set 2
Threshold Value Hysteresis Value
THR_1 12 (6 dB) HYST_1 2 (1 dB)
THR_2 17 (8.5 dB) HYST_2 3 (1.5 dB)
THR_3 25 (12.5 dB) HYST_3 3 (1.5 dB)

3.4.5 FR Codec Set 5

Table 5 Full Rate Codec Set 5 (= Preferred Configuration 1 (TS 28.062))


CODEC_MODE_1 4.75 kbps
CODEC_MODE_2 5.9 kbps
CODEC_MODE_3 7.4 kbps
CODEC_MODE_4 12.2 kbps

This codec set, together with the Half Rate Codec Set 2, see below,
satisfies the AMR FR and AMR HR operation in an optimal way for
TFO connections. With these two codec sets it will be possible to
run TFO on the three lower modes (7.4 - 5.9 - 4.75).

This codec set, together with the Half Rate Codec Set 2, see below,
is the only codec set that is allowed when A-Interface over IP is
used. See Section 3.4.2 and Reference [9].

Table 6 Decision Thresholds and Hysteresis Values for Full Rate Codec Set 5
Threshold Value Hysteresis Value
THR_1 12 (6 dB) HYST_1 2 (1 dB)
THR_2 17 (8.5 dB) HYST_2 3 (1.5 dB)
THR_3 25 (12.5 dB) HYST_3 3 (1.5 dB)

3.4.6 HR Codec Set 1

Table 7 Half Rate Codec Set 1 (= Preferred Configuration 10 (TS28.062))


CODEC_MODE_1 4.75 kbps
CODEC_MODE_2 5.9 kbps
CODEC_MODE_3 6.7 kbps
CODEC_MODE_4 7.4 kbps

This codec set has been selected since extensive evaluations have
shown that it provides a good overall speech quality during all
radio conditions (both for clean speech and with different
background noise), that is it provides a very small deviation from
optimum AMR performance (when all 6 codec modes are used).

This codec set, together with Full Rate Codec Set 1, see above,
satisfies the AMR FR and AMR HR operation for TFO connections.
With these two codec sets it will be possible to run TFO on the
three lower modes (6.7 - 5.9 - 4.75) since they are common.

Table 8 Decision Thresholds and Hysteresis Values for Half Rrate Codec Set 1
Threshold Value Hysteresis Value
THR_1 23 (11.5 dB) HYST_1 3 (1.5 dB)
THR_2 28 (14 dB) HYST_2 4 (2 dB)
THR_3 32 (16 dB) HYST_3 4 (2 dB)

3.4.7 HR Codec Set 2

Table 9 Half Rate Codec Set 2 (= Preferred Configuration 1 (TS 28.062))


CODEC_MODE_1 4.75 kbps
CODEC_MODE_2 5.9 kbps
CODEC_MODE_4 7.4 kbps

This codec set provides basically the same overall speech quality
as half rate codec set 1 but offers a slightly better speech quality
under noisy speech conditions.
This codec set, together with Full Rate Codec Set 5, see above,
satisfies the AMR FR and AMR HR operation for TFO connections.
With these two codec sets it will be possible to run TFO on the
three common modes (7.4 - 5.9 - 4.75).

This codec set, together with the Full Rate Codec Set 5, see above,
is the only codec set that is allowed when A-Interface over IP is
used. See Section 3.4.2 and Reference [9].

Table 10 Decision Thresholds and Hysteresis Values for Half Rate Codec Set 2
Threshold Value Hysteresis Value
THR_1 23 (11.5 dB) HYST_1 3 (1.5 dB)
THR_2 31 (15.5 dB) HYST_2 4 (2 dB)

3.5 DTX
The radio network feature Discontinuous Transmission (see
Reference [3]) is supported for the AMR codec as well as for the
other codec types.

The DTX operation has been slightly changed for AMR because of
the inband signalling, which needs to be carried out as part of the
codec mode adaptation. The Codec Mode Indication or Codec Mode
Command/Request are transmitted in the SID_FIRST frames and
both Codec Mode Indication and Codec Mode Command/Request
are sent in every SID_UPDATE frame. The new ONSET frame
allows a very reliable detection of the end of a speech pause / start
of new speech. Less front end clipping is the result. The
interleaving scheme for SID_UPDATE is optimized and needs only
half the time slots as for previous codec types. SID_FIRST and
ONSET do not need any extra time slots, but utilize the “left-overs”
from the speech frame.

3.6 Handover
AMR is handled in the same way as all other speech versions at
handover except for the parameter AMRSPEECHVERUSE (see
Reference [1]).

3.7 Tandem Free Operation


For more information about TFO than given in this UD, see
Reference [7].

3.8 Transcoder Free Operation


For more information about A-Interface over IP and TrFO than
given in this UD, see Reference [9].

3.9 Related Statistics


3.9.1 Impact on Legacy Counters

Statistics may change when turning on AMR FR and AMR HR. The
AMR codecs will provide increased protection to the speech, in
particular when codec mode 4.75 kbps is used. The channel coding
for codec mode 4.75 kbps is more robust compared to the SACCH,
this may increase the dropped call rate. Analysis has shown that
users using codec modes higher than 4.75 kbps may adjust their
location when speech quality is low in bad RF spots. Users using
codec mode 4.75 kbps may not notice the poor RF spots due to
good enough speech quality, hence they are more likely to drop.

While speech robustness is increased with AMR, the SACCH


robustness will remain the same with or without AMR. SACCH is
used for measurement reports and needed for Locating. With
frequent loss of SACCHs bursts when holding on to a call, an
automatic drop may eventually be triggered. The result is that the
subscriber may not experience degraded speech quality prior to
the drop. It may therefore be a shift of distribution between drops
(system initiated due to lost signalling on SACCH) and subscriber
initiated disconnections (due to unsatisfying speech quality).

One potential remedy against this is to lower the lowest threshold


a bit, such that the second less robust AMR mode in the fringe area
is used longer, thus displaying more error events to the customer.

Figure 3 Subscribers Experience Better Speech Quality with AMR


and Hold on to the Call until It Drops Due to Bad SACCH
Performance Instead of Initiating a Disconnection Themselves.

When analyzing the above described behavior drop & disconnection


parameter (BADQDL/UL, LOWSSDL/UL), drop& disconnection
counters (for example TFDISQAUL, DISBQA), SACCH parameters
(RLINKT,
RLINKTAFR,RLINKTAHR,RLINKUP,RLINKUPAFR,RLINKUPAH
R) and cellplan (coverage, interference) should be considered. The
parameters
RLINKTAFR,RLINKTAHR,RLINKUPAFR,RLINKUPAHR might be
set to a higher value than RLINKT,RLINKUP to give a more
robust SACCH to the expense of longer holding time of the
equipment at dropped call. For more information about these
parameters, see Reference [2], Reference [4] and Reference [5].

3.9.2 Statistics for Performance Management

There are a number of STS counters related to AMR:

 SQI counters, belong to object types CELLSQI and


CELLSQIDL.
 Codec mode utilization counters (uplink and downlink),
belong to object types CLTCHFV3C and CLTCHHV3C.
 FER counters, belong to object types CELLFERF, CELLFERH,
CELLAFFER and CELLAHFER.
 Dropped call counters, belong to object types CELTCHF,
CELTCHH, CLTCHDRAF and CLTCHDRAH.
 Traffic level counters, belong to object types CELTCHF and
CELTCHH.

Additional information on counters can be found in Reference [6].

Measurement Result Recording (MRR) is a feature that can be used


in troubleshooting, in periodic monitoring of the network
performance and in determining the impact of changes to the
network. If AMR is used in the system an extra 'AMR Connection
Filter' is available at the Recording Definition in MRR. Using this it
is possible to specify:

 'No Filter', that is all connections will be recorded.


 'Record Only Measurement Reports from non AMR full rate
Connections'.
 'Record Only Measurement Reports from AMR full rate
Connections'.

This enables targeted optimization activities when AMR FR mobile


stations are introduced in the network.

4 Engineering Guidelines

4.1 Fine Tuning Codec Sets


Besides the pre-defined codec sets (1, 2 and 5 for AMR FR and 1
and 2 for AMR HR), both AMR FR and AMR HR codec set 3 and 4
are available. These codec sets can be configured by the operator
giving the possibility to fine tune AMR parameters, based on
statistics, and optimize the different codec rates within a codec set.
However, the pre-defined codec sets will work well in most
networks.

Note:
If fine tuning is needed when A-interface over IP is used the codec rates
included in the codec sets must be same as in FR 5 and HR 2 that is 12.2,
7.4, 5.9 and 4.75 for fullrate and 7.4, 5.9 and 4.75 for halfrate.

5 Parameters

5.1 Main Controlling Parameters


AMR FR and AMR HR are activated by changing BSC
exchange properties. Transcoder pools for AMR FR and
AMR HR must also be defined in order for AMR to work.
Note if AMR HR shall be activated the feature Half Rate
Channels must have been purchased by the operator.

The following exchange properties is set by the BSC


command RAEPC:

 AMRFRSUPPORT

Indicates if AMR FR is turned ON or OFF and also


which full rate codec set that shall be used within
the BSC.

 AMRHRSUPPORT

Indicates if AMR HR is turned ON or OFF and also


which half rate codec set that shall be used within
the BSC.

5.2 Parameters for Special Adjustments


Codec sets 3 and 4 for both FR and HR channels can be
defined per BSC using the command RLADC. The following
parameters are set using this command:

 SET

Indicates which of the codec sets for a specific


channel rate that shall be modified.

 MODE

Specifies the codec modes that shall be included in


the codec set. For Full Rate all 8 codec modes can
be used, expressed as a numeral between 1 and 8,
and defined in ascending order.

1 = 4.75 kbps

2 = 5.15 kbps
.

8 = 12.2 kbps

For Half Rate only the 5 lower codec modes can be


used and at least one of the three modes 4.75
kbps, 5.15 kbps or 5.90 kbps must be included in
the set. The reason for this is that one of these
modes must be used as initial mode to secure a
good signalling quality on the Abis interface.

1 = 4.75 kbps

5 = 7.4 kbps

 THR

Specifies the thresholds between the codec modes


in the codec set, expressed as a numeral between
0 and 63, in steps of 0.5 dB.

 HYST

Specifies the hysteresis values for the threshold in


the codec set. It is expressed as a numeral
between 0 and 15, in steps of 0.5 dB.

 ICMODE

Specifies the initial codec mode to be used when


the channel is activated (at call set up or at
handover). This parameter is specified if a non-
standard setting is prefered. If not given for a
codec set containing 3 codec modes the lowest will
be used as initial codec mode while for codec set
with 4 codec modes the second lowest will be used
as initial codec mode. In case A-interface over IP is
used the standard setting is recommended.

5.3 Value Ranges and Default Values

Table 11 Parameters for AMR


Parameter name Default Recommended Value Unit
value value range
AMRFRSUPPORT 0 — 0 to 5 0=off,
1=Fullra
1,

2=Full ra
set 2,

3=Full ra
set 3,

4=Full ra
set 4

5=Full ra
set 5
AMRHRSUPPORT 0 — 0 to 4 0=off,

1=Half ra
set 1,

2=Half ra
set 2,

3=Half ra
set 3,

4=Half ra
set 4
SET — — FR3, FR4, Codec Se
HR3, HR4
MODE — — 1 to 8 Codec Mo
THR — — 0 to 63 Threshold
HYST — — 0 to 15 Hysteres
ICMODE — — 1 to 8 Codec Mo

6 Concepts

Active Codec A set of up to 4 different codec modes (all using the same
Set rate).

Adaptive Multi Speech and channel codec capable of operating at gross b


Rate 11.4 and 22.8 kbps. The codec can operate at various com
of speech and channel coding bit rates.

Channel Rate Full rate (gross bit rate of 22.8 kbps) or half rate (gross b
11.4 kbps) operation.

Codec Mode A codec rate used within a codec set is called a codec mod

C/I The ratio between signal strength and interference signal


(expressed i dB).
TFO Tandem Free Operation is a feature that removes unneces
speech encoding/decoding in the transcoders for mobile to
calls but needs the transcoder hardware in the path. Spee
encoding/decoding is then only done in the mobiles, see R
[11].

TrFO Transcoder Free Operation, avoids the transcoding functio


networks. Does typically not need the transcoder hardwar
path.

Das könnte Ihnen auch gefallen