ETSI TR 103 590 V1.1.1 (2018-09)

TECHNICAL REPORT
Digital Enhanced Cordless Telecommunications (DECT);
Study of Super Wideband Codec in DECT for narrowband,
wideband and super-wideband audio communication
including options of low delay
audio connections (≤ 10 ms framing)

2 ETSI TR 103 590 V1.1.1 (2018-09)

Reference
DTR/DECT-00316
Keywords
codec, DECT, superwideband
ETSI
650 Route des Lucioles
F-06921 Sophia Antipolis Cedex - FRANCE

Tel.: +33 4 92 94 42 00  Fax: +33 4 93 65 47 16

Siret N° 348 623 562 00017 - NAF 742 C
Association à but non lucratif enregistrée à la
Sous-Préfecture de Grasse (06) N° 7803/88

Important notice
The present document can be downloaded from:
http://www.etsi.org/standards-search
The present document may be made available in electronic versions and/or in print. The content of any electronic and/or
print versions of the present document shall not be modified without the prior written authorization of ETSI. In case of any
existing or perceived difference in contents between such versions and/or in print, the only prevailing document is the
print of the Portable Document Format (PDF) version kept on a specific network drive within ETSI Secretariat.
Users of the present document should be aware that the document may be subject to revision or change of status.
Information on the current status of this and other ETSI documents is available at
https://portal.etsi.org/TB/ETSIDeliverableStatus.aspx
If you find errors in the present document, please send your comment to one of the following services:
https://portal.etsi.org/People/CommiteeSupportStaff.aspx
Copyright Notification
No part may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying
and microfilm except as authorized by written permission of ETSI.
The content of the PDF version shall not be modified without the written authorization of ETSI.
The copyright and the foregoing restriction extend to reproduction in all media.

© ETSI 2018.
All rights reserved.
TM TM TM
DECT , PLUGTESTS , UMTS and the ETSI logo are trademarks of ETSI registered for the benefit of its Members.
TM TM
3GPP and LTE are trademarks of ETSI registered for the benefit of its Members and
of the 3GPP Organizational Partners.
oneM2M logo is protected for the benefit of its Members. ®
GSM and the GSM logo are trademarks registered and owned by the GSM Association.
ETSI
3 ETSI TR 103 590 V1.1.1 (2018-09)
Contents
Intellectual Property Rights . 5
Foreword . 5
Modal verbs terminology . 5
Introduction . 5
1 Scope . 6
2 References . 6
2.1 Normative references . 6
2.2 Informative references . 6
3 Definitions and abbreviations . 7
3.1 Definitions . 7
3.2 Abbreviations . 7
4 Investigations on an enhanced DECT codec . 8
4.1 Overview . 8
4.2 Design constraints/features to be supported . 9
4.2.1 Improved exploitation on DECT slots . 9
4.2.2 Transmission latency . 9
4.2.3 Supported sampling rates, audio bandwidths and sample depths . 9
4.2.4 Support for music streaming . 9
4.2.5 Packet loss concealment . 9
4.2.6 Low codec complexity . 9
4.3 Investigations on impact of block-based codec . 10
4.3.1 Probability and distribution of bit errors . 10
4.3.1.1 Normal slots - transmission error profile I . 10
4.3.1.2 Long slots - transmission error profile II . 13
4.3.2 Study of CRC/FEC protection schemes . 16
4.3.2.1 Static rates . 16
4.3.2.2 Dynamic rate switching of source and channel coder . 19
4.3.2.2.1 General . 19
4.3.2.2.2 Graceful degradation at DECT range limit . 19
4.3.2.2.3 Audio bandwidth switching . 19
4.3.2.2.4 Potential channel coder configuration . 20
4.3.3 Comparison to current DECT codecs . 20
4.3.4 Usability of I _minimum_delay service . 20
N
4.4 Evaluation results of LC3 . 21
4.4.1 Audio quality evaluation for clean channels . 21
4.4.1.1 NB/WB experiment tandeming I . 21
4.4.1.1.1 Setup . 21
4.4.1.1.2 Observations . 23
4.4.1.2 SSWB/SWB experiment tandeming I . 23
4.4.1.2.1 Setup . 23
4.4.1.2.2 Observations . 25
4.4.1.3 WB/SWB experiment on short frame size . 25
4.4.1.3.1 Setup . 25
4.4.1.3.2 Observations . 26
4.4.2 Audio quality evaluation for error prone channels . 26
4.4.2.1 Reference conditions . 26
4.4.2.2 Codec characterization depending on PLR . 27
4.4.2.2.1 Setup . 27
4.4.2.2.2 Observations . 28
4.4.2.3 Codec characterization depending on PLR and BER . 28
4.4.2.3.1 Setup . 28
4.4.2.3.2 Observations . 31
4.5 Integration of LC3 in existing DECT infrastr ucture . 31
ETSI
4 ETSI TR 103 590 V1.1.1 (2018-09)
4.5.1 General . 31
4.5.2 High-level codec description . 31
4.5.2.1 Codec overvie w . 31
4.5.2.2 Audio channels . 31
4.5.2.3 Complexity . 31
4.5.2.4 Audio bandwidth detection . 31
4.5.3 High-level options for LC3 integration into the DECT specifications. 32
4.5.3.1 Fixed bitrate, multi-mode or adaptive multi-mode operation . 32
4.5.3.2 Codec negotiation at session setup . 33
4.5.3.3 Codec mode set negotiation . 33
4.5.3.4 Adaptation signalling . 34
4.5.3.5 Audio bandwidth selection/adaptation . 35
4.5.3.6 DTX . 35
4.5.3.7 Acoustic front-end . 36
4.5.3.8 VoIP, RTP payload format . 36
4.5.3.9 Backwards compatibility . 36
4.5.3.10 Interworking with external networks and devices . 36
4.5.3.11 Repeater operation, relays . 37
5 Conclusions . 37
Annex A: Change History . 39
History . 40

ETSI
5 ETSI TR 103 590 V1.1.1 (2018-09)
Intellectual Property Rights
Essential patents
IPRs essential or potentially essential to normative deliverables may have been declared to ETSI. The information
pertaining to these essential IPRs, if any, is publicly available for ETSI members and non-members, and can be found
in ETSI SR 000 314: "Intellectual Property Rights (IPRs); Essential, or potentially Essential, IPRs notified to ETSI in
respect of ETSI standards", which is available from the ETSI Secretariat. Latest updates are available on the ETSI Web
server (https://ipr.etsi.org/).
Pursuant to the ETSI IPR Policy, no investigation, including IPR searches, has been carried out by ETSI. No guarantee
can be given as to the existence of other IPRs not referenced in ETSI SR 000 314 (or the updates on the ETSI Web
server) which are, or may be, or may become, essential to the present document.
Trademarks
The present document may include trademarks and/or tradenames which are asserted and/or registered by their owners.
ETSI claims no ownership of these except for any which are indicated as being the property of ETSI, and conveys no
right to use or reproduce any trademark and/or tradename. Mention of those trademarks in the present document does
not constitute an endorsement by ETSI of products, services or organizations associated with those trademarks.
Foreword
This Technical Report (TR) has been produced by ETSI Technical Committee Digital Enhanced Cordless
Telecommunications (DECT).
Modal verbs terminology
In the present document "should", "should not", "may", "need not", "will", "will not", "can" and "cannot" are to be
interpreted as described in clause 3.2 of the ETSI Drafting Rules (Verbal forms for the expression of provisions).
"must" and "must not" are NOT allowed in ETSI deliverables except when used in direct citation.
Introduction
Since the introduction of additional codecs in New Generation DECT [i.5] in 2007, wideband services have been widely
established for fixed line, mobile and OTT communications networks. This trend is gaining even more momentum by
services using cutting edge codecs like 3GPP EVS and the upcoming new Bluetooth™ codec by offering super-
wideband audio bandwidth.
NOTE: Bluetooth™ is the trade name of a wireless technology standard for exchanging data over short distances
(using short-wavelength UHF radio waves in the ISM band from 2,4 - 2,485 GHz) from fixed and mobile
devices, and building personal area networks (PANs), owned by the Bluetooth Special Interest Group.
This information is given for the convenience of users of the present document and does not constitute an
endorsement by ETSI of the technology named. Equivalent technologies may be used if they can be
shown to lead to the same results.
Recent market research from several relevant DECT infrastructure providers indicates a demand for upgrading DECT
services and standard with additional features enabled by evolved speech and audio codecs.
The present document collects performance requirements to add a real benefit to current and upcoming DECT
applications and evaluates the Low Complexity Communication Codec (LC3) on suitability for this as well as discusses
possible adaptations for DECT environments in terms of error protection and signalling.

ETSI
6 ETSI TR 103 590 V1.1.1 (2018-09)
1 Scope
The present document provides a study of technical updates to the DECT standard to enable super wideband (SWB)
audio calls in existing DECT slot formats as well as technical improvements to narrowband (NB) and wideband (WB)
calls. All required change requests are listed and defined for the different DECT layers to enable high quality audio
communication between DECT FP and PP including DECT repeaters (relays). The study includes an investigation on
FEC for block-based codecs. Information is provided on the audio quality in some DECT use cases for NB, WB and
SWB and potential improvements by a new audio codec are studied.
2 References
2.1 Normative references
Normative references are not applicable in the present document.
2.2 Informative references
References are either specific (identified by date of publication and/or edition number or version number) or
non-specific. For specific references, only the cited version applies. For non-specific references, the latest version of the
referenced document (including any amendments) applies.
NOTE: While any hyperlinks included in this clause were valid at the time of publication ETSI cannot guarantee
their long-term validity.
The following referenced documents are not necessary for the application of the present document but they assist the
user with regard to a particular subject area.
[i.1] ETSI EN 300 175-1: "Digital Enhanced Cordless Telecommunications (DECT); Common
Interface (CI); Part 1: Overview".
[i.2] ETSI EN 300 175-8: "Digital Enhanced Cordless Telecommunications (DECT); Common
Interface (CI); Part 8: Speech and audio coding and transmission".
[i.3] EP2901594B1: "Error Detection for sub-band ADPCM encoded sound signal".
[i.4] ETSI EN 300 175-3: "Digital Enhanced Cordless Telecommunications (DECT); Common
Interface (CI); Part 3: Medium Access Control (MAC) layer".
[i.5] ETSI TS 102 527-3: "Digital Enhanced Cordless Telecommunications (DECT); New Generation
DECT; Part 3: Extended wideband speech services".
[i.6] ETSI EN 300 700: "Digital Enhanced Cordless Telecommunications (DECT); Wireless Relay
Station (WRS)".
[i.7] ETSI EN 300 175-5: "Digital Enhanced Cordless Telecommunications (DECT); Common
Interface (CI); Part 5: Network (NWK) layer".
[i.8] ETSI EN 300 176-2: "Digital Enhanced Cordless Telecommunications (DECT); Test
specification; Part 2: Audio and speech".
[i.9] Recommendation ITU-T P.863 (09-2014): "Perceptual objective listening quality assessment".
[i.10] Recommendation ITU-T P.800 (08-1996): "Methods for subjective determination of transmission
quality".
[i.11] Recommendation ITU-T G.191 (03-1996): "Software tools for speech and audio coding
standardization" .
[i.12] Recommendation ITU-T G.722 (09-2012): "7 kHz audio-coding within 64 kbit/s".
ETSI
7 ETSI TR 103 590 V1.1.1 (2018-09)
[i.13] Recommendation ITU-T G.726 (12-1990): "40, 32, 24, 16 kbit/s Adaptive Differential Pulse Code
Modulation (ADPCM)".
[i.14] ETSI TS 126 071: "Digital cellular telecommunications system (Phase 2+) (GSM); Universal
Mobile Telecommunications System (UMTS); LTE; Mandatory speech CODEC speech
processing functions; AMR speech Codec; General description (3GPP TS 26.071)".
[i.15] ETSI TS 126 171: "Digital cellular telecommunications system (Phase 2+) (GSM); Universal
Mobile Telecommunications System (UMTS); LTE; Speech codec speech processing functions;
Adaptive Multi-Rate - Wideband (AMR-WB) speech codec; General description (3GPP
TS 26.171)".
[i.16] ETSI TS 126 441: "Universal Mobile Telecommunications System (UMTS); LTE; Codec for
Enhanced Voice Services (EVS); General overview (3GPP TS 26.441)".
[i.17] SIG Bluetooth™ Hands-Free Profile 1.7.1.
3 Definitions and abbreviations
3.1 Definitions
For the purposes of the present document, the terms and definitions given in ETSI EN 300 175-1 [i.1] and the following
apply:
Fullband (FB): speech or audio sampled at 48 kHz
Fullband, compact disc (FBCD): speech or audio sampled at 44,1 kHz
Narrowband (NB): speech or audio sampled at 8 kHz
Semi-Super Wideband (SSWB): speech or audio sampled at 24 kHz
Super Wideband (SWB): speech or audio sampled at 32 kHz
Wideband (WB): speech or audio sampled at 16 kHz
3.2 Abbreviations
For the purposes of the present document, the following abbreviations apply:
rd
3GPP 3 Generation Partnership Project
ACR Absolute Category Rating
NOTE: See Recommendation ITU-T P.800 [i.10].
AMR Adaptive Multi-Rate,
NOTE: See ETSI TS 126 071 [i.14].
AMR-WB Adaptive Multi-Rate Wideband
NOTE: See ETSI TS 126 171 [i.15].
BCH Bose-Chaudhuri-Hocquenghem
BER Bit Error Rate
CELT Constrained Energy Lapped Transform
CMR Codec Mode Request
CNG Comfort Noise Generation
CRC Cyclic Redundancy Check
CS Circuit Switched
DECT Digital Enhanced Cordless Telecommunications
ETSI
8 ETSI TR 103 590 V1.1.1 (2018-09)
DSP Digital Signal Processor
DTX Discontinuous Transmission
EP Error Protection
ETSI European Telecommunications Standards Institute
EVS Enhanced Voice Service
NOTE: See ETSI TS 126 441 [i.16].
FB Fullband
FBCD Fullband Compact Disc
FEC Forward Error Correction
FER Frame Erasure Rate
FP Fixed Part (DECT bas station)
FT Frame Type
GFSK Gaussian Frequency-Shift Keying
GSM Global System for Mobile Communications
IETF Internet Engineering Task Force
I higher layer Information channel (unprotected), minimum delay operation
NA
NOTE: See ETSI TS 102 527-3 [i.5].
I higher layer Information channel (unprotected), normal delay operation
NB
NOTE: See ETSI TS 102 527-3 [i.5].
IP Internet Protocol
LAN Local Area Network
LC3 Low Complexity Communication Codec
MIPS Million Instructions Per Second
MNRU Modulated Noise Reference Unit
MOS-LQO Mean Opinion Score - Listening Quality Objective
mSBC modified Subband Coding
NOTE: See [i.17].
NB Narrowband
OTT Over-The-Top content
PLC Packet Loss Concealment
PLR Packet Loss Rate
PP Portable Part (DECT handset)
RAM Random-Access Memory
RSSI Received Signal Strength Indicator
RTP Real-Time Protocol
SSWB Semi-Super Wideband
STL Software Tool Library
SWB Super Wideband
VoIP Voice over Internet Protocol
WB Wideband
WMOPS Weighted Millions of Operations Per Second
WRS Wireless Relay Station
4 Investigations on an enhanced DECT codec
4.1 Overview
The investigations are organized as follows:
1) Definition of general required codec features (clause 4.2).
2) Study on error profiles and protection schemes for DECT systems (clause 4.3).
ETSI
9 ETSI TR 103 590 V1.1.1 (2018-09)
3) Characterization of the LC3 as potential candidate (clause 4.4).
4) Definition of required update to DECT specifications (clause 4.5).
4.2 Design constraints/features to be supported
4.2.1 Improved exploitation on DECT slots
Table 1 compares the slot related requirements of the legacy DECT codecs with the proposed new DECT codec.
Table 1: Overview of DECT slot related requirements for new codec
Slot usage Legacy DECT codecs Enhanced DECT codec
Normal slots NB calls (G.726) NB and WB calls
Long slots WB calls (G.722) WB and SWB calls

NB and WB audio quality should be comparable to or better relative to the legacy DECT codecs.
The user experience for error prone channels is expected to be comparable to or better relative to the legacy DECT
codecs.
4.2.2 Transmission latency
The codec should operate on 10 ms frame sizes as provided by the DECT transmission slots [i.1]. On top of the framing
delay of 10 ms, the additional algorithmic delay should be less than or equal to 2,5 ms.
Additionally, the codec should support frame sizes of 5 ms as well to enable new low delay application besides
telephony. For instance, in-room conferencing/amplification or parliament systems require a microphone to loudspeaker
delay of less than 20 ms. This guarantees lip-synchronism of the speaker to the amplified signal.
4.2.3 Supported sampling rates, audio bandwidths and sample depths
The codec should support NB, WB, SWB and FB audio bandwidths at the native sample rates of 8 kHz, 16 kHz, 32 kHz
and 48 kHz. Additionally, 24 kHz (SSWB) should be supported.
The codec should support the coding of lower audio bandwidths for a given sample rate, e.g. coding of NB signals at
32 kHz. The codec should support the coding of audio samples with 16 bits per samples and may support coding of
audio samples with 24 bits per sample.
4.2.4 Support for music streaming
The codec should provide decent audio quality for music streaming services and may provide additional coding features
to support stereo music channels.
4.2.5 Packet loss concealment
The codec should support packet loss concealment without adding further algorithmic delay. As the main application is
voice, the packet loss concealment should perform well for speech signals.
4.2.6 Low codec complexity
The codec should run with a low computational complexity and low memory footprint to be implementable on typical
DECT handheld devices. The complexity should be measured and reported using the latest ITU-T STL complexity
measurement toolbox [i.11].
ETSI
10 ETSI TR 103 590 V1.1.1 (2018-09)
4.3 Investigations on impact of block-based codec
4.3.1 Probability and distribution of bit errors
4.3.1.1 Normal slots - transmission error profile I
An error profile was measured using a real DECT system simulating that a DECT caller is moving through an office
building. The caller starts close to the base station and walks away through the office. The measurement recorded the
number of bit errors, the position of the bit error, the signal strength (RSSI, in steps of eight) and complete frame losses.
Figure 1 outlines the error profile where complete frame losses are indicated by 384 bit errors (A+B field) in
combination with signal strength zero.

Figure 1: Characterization of Error Profile I
Regarding the position of the bit errors inside the frame, no specific dependency can be found. In order to structure the
analysis of the pattern, the bit error profile is further grouped into certain signal strength classes as outlined in Figure 2.
ETSI
11 ETSI TR 103 590 V1.1.1 (2018-09)

Figure 2: Plot of absolute number (blue) and average number (red) of bit errors for each signal
strength (top); Histogram of number of bit errors inside frame for
signal strength levels 32 dB, 40 dB, 48 dB and 56 dB
The plots show that:
• For this profile, bit errors only occur for a signal strength level ≤ 56 dB.
• For the signal strength level 48 dB and 56 dB, most packets show less than three bit errors in one packet.
• For signal strength level 40 dB, the bit error rate per packet is significantly higher compared to level 48 dB.
• For signal strength level 32 dB, almost all bits are affected; this level is thus not considered for any recovery
activity.
• Four different error protection classes may be appropriate to address the different error characteristics
depending on the signal strength, i.e. clean channel, 56 dB, 48 dB, 40 dB.
The packet loss rate (PLR) per signal strength can be estimated by averaging the signal strength over one second.
Figure 3 shows packet losses in relation to the averaged signal strength.
ETSI
12 ETSI TR 103 590 V1.1.1 (2018-09)

Figure 3: Packet loss rate estimation
According to the given data, a specific PLR can be assigned to a certain signal level as outlined in Table 2.
Table 2: Packet loss and bit error rates
Normalized Averaged signal # Frames # Packet PLR PLR rounded BER rounded
strength (RSSI) losses [%] [%] [%]
1 (136 dB) 430 0 0 0
0,94 (128 dB) 728 0 0 0
0,88 (120 dB) 41 0 0 0
0,82 (112 dB) 78 0 0 0
0,76 (104 dB) 20 0 0 0
0,71 (96 dB) 28 0 0 0
0,65 (88 dB) 75 0 0 0
0,59 (80 dB) 2 380 0 0 0
0,53 (72 dB) 320 0 0 0
0,47 (64 dB) 436 5 1,15 1
0,41 (56 dB) 203 2 0,99 1 0,01
0,35 (48 dB) 3 195 28 0,88 1 0,31
0,29 (40 dB) 1 190 88 7,39 7 2,92
0,24 (32 dB) 543 133 24,49 24 17,13

Please note that for realistic simulations packet losses come in addition to any bit errors. Figure 4 shows the cumulative
bit error probability as number of bit errors per frame for the relevant RSSIs. This plot provides a compact view of the
expected number of bit errors per packet, e.g. at RSSI 48, about 62 % of the packets show no bit error, 70 % show less
than 2 errors and 80 % of the packets show less than 4 bit errors.
ETSI
13 ETSI TR 103 590 V1.1.1 (2018-09)

Number of bit errors
Figure 4: Cumulative probability of bit errors per frame (frame length 40 bytes)
Given the fact that the bit error rate correlates to the signal strength, an adaptively controlled forward error correction
code may be beneficial for a DECT system using a block-based audio codec. In order to find the best compromise
between voice quality and channel robustness at any signal condition, the codec needs to provide a seamless rate
switching scheme to allow the adaptation of codec rate and error protection rate.
NOTE: The parameter signal strength/RSSI might not be usable in a real DECT system and might be replaced by
other metrics.
4.3.1.2 Long slots - transmission error profile II
Since a long slot is established based on two normal slots, it is assumed that probability and distribution of bit errors
behave similar to those of normal slots.
To prove this assumption, error profiles were derived by Fraunhofer IIS using the following setup:
• An FP (AVM FRITZ!Box 7490) with updated firmware for improved testing and logging capabilities provided
by the manufacturer (see note)
• A PP (AVM FRITZ!Fon C4) with updated firmware for improved testing and logging capabilities provided by
the manufacturer (see note)
• A computer connected to the PP via LAN, where the following software was used on this computer
• A web browser to configure and monitor the FP
• A soft phone to call the PP
• Wireshark to record the RTP packed G.722 DECT streams (outgoing and incoming)
NOTE: AVM FRITZ!Box 7490 and AVM FRITZ!Fon C4 are examples of suitable products available
commercially. This information is given for the convenience of users of the present document and does
not constitute an endorsement by ETSI of these products.
In preparation for the measurements, the PP was put into a special mode, in which it responds automatically to an
incoming call after 15 s, loops back the incoming signal and terminates the connection after 60 s.
ETSI
14 ETSI TR 103 590 V1.1.1 (2018-09)
The measurements itself were performed as follows:
• The PP was called from the soft phone.
• A test signal was sent from the soft phone. First, a signal was used where a harmonic part (pitch pipe) and a
non-harmonic part (clicks) were sent alternant (which helped for the visual inspection), later a noise signal was
used (which allows better cross-correlation).
• The data transfer of the computer was recorded by a network protocol analyser.
• Afterwards the RTP streams containing the G.722 DECT signals (outgoing and incoming), were dumped.
• Dumped streams were subsequently aligned (by correlation measures).
• Statistics were derived with regard to the differences between the outgoing and the incoming stream, which
can be treated as transmission errors.
• Successful recording was proved by decoding the streams.
A total of 41 measurements were performed:
• Some test measurements to verify the set-up.
• Some measurements outdoor at different distances.
• Some measurements indoor on a long corridor at different distances.
• Some further measurements indoor on a long corridor but moving some meters into side corridors or rooms.
• Some further measurements indoor, where the PP was moved around.
Figure 5 gives an overview of the cumulative probability of bit errors per frame for the 41 measurements. Note that the
derived error statistics reflect two-way transmission.
Cumulative bit error probability
1,1
0,9
0,7
0,5
0,3
0,1
0 100 200 300 400 500 600 700
-0,1
Number of bit errors
Figure 5: Cumulative probability of bit errors per frame (frame length 80 bytes) - overall figure
ETSI
Probability
15 ETSI TR 103 590 V1.1.1 (2018-09)
Assuming a symmetric channel, one could halve the number of errors for a one-way transmission. The setup provides
no option to distinguish between bit errors occurring from PP to FP or from FP to PP.
A legend for the different measurements is left out by intention since it provides no additional information. In fact, there
is no strong correlation between the distance and the error rate. Instead, a lot of soft factors (people on the floor, other
electromagnetic sources, walls, etc.) seem to have an impact on the channel quality. In contrast to the experiment
undertaken for normal slots, no RSSI was available on a frame-by-frame base, thus no RSSI based evaluation can be
performed. Having a closer look at the data, a post-screening seems to be appropriate. On one hand there are nine
measurements without any error and seven further measurements with an extremely low amount of bit errors. On the
other hand, there are two measurements showing a large amount of bit errors per frames, that no meaningful connection
is possible anymore. Figure 6 shows just the remaining 23 measurements in a close-up. For comparison reasons, the
error profiles described in clause 4.3.1.1 for normal slots are extrapolated to long slots and plotted as well.
Cumulative bit error probability
1,2
0,8
0,6
0,4
0,2
0 102030 405060 7080
Number of bit errors
movement_4_indoor_noise 080m_indoor 080m+01r_indoor_noise
028m+05r_indoor 110m_indoor_noise_2 movement_1_indoor_noise
110m_indoor_2 030m+01l_indoor_noise 110m_indoor_1
035m+05l_indoor 028m+cinema_indoor_noise 080m+04r_indoor_noise
065m+03l_indoor movement_3_indoor_noise 100m_outdoor
110m_indoor_behindGlass 090m+02l_indoor 060m+10r_indoor
065m+03l_indoor_noise 028m+05r_indoor_noise 035m+04l_indoor_noise
065m+06l_indoor 035m+10l_indoor RSSI32
RSSI40 RSSI48 RSSI56
Figure 6: Cumulative probability of bit errors per frame (frame length 80 bytes) - post-screened:
Showing only the profiles where FEC may be beneficial
As can be seen from Figure 6, the number of correctly decodable frames can in many situations already be increased
significantly by correcting a small number of bits. Afterwards a saturation is visible in almost all measurements,
meaning that an increased FEC capability will have almost no positive effect on the number of correctly decodable
frames. In some rare cases there is a second area (between 30 and 70 bit errors), in which the number of correctly
decodable frames could be increased, but this would require an extremely strong FEC.
ETSI
Probability
16 ETSI TR 103 590 V1.1.1 (2018-09)
In general, the extrapolated error profiles for normal slots show a very similar slope compared to the real measurements
for long slots.
4.3.2 Study of CRC/FEC protection schemes
4.3.2.1 Static rates
Assessments based on Recommendation ITU-T P.863 [i.9] were carried out for 18 speech items (clean speech, duration
58 s each). The following encoder configurations were assessed:
• narrowband (8 kHz sampling rate) at 32 kbps
• wideband (16 kHz sampling rate) at 32 kbps
• wideband (16 kHz sampling rate) at 64 kbps
• super wideband (32 kHz sampling rate) at 64 kbps
The following forward error correction schemes were assessed:
• no protection;
• Bose-Chaudhuri-Hocquenghem (BCH) codes with a max correction capability starting from 1 bit and going up
to the maximum possible correction rate, which depends on the minimum supported audio payload for the
tested configuration; note that the used correction capability is 1 bit less than the max correction capability in
order to have a sufficient error detection capability.
Figure 7 shows the LC3 net bitrate relative to the applied forward error protection scheme.
Forward error protection vs. net bitrate
normal slots - 32kbps gross bitrate long slots - 64kbps gross bitrate

Figure 7: Forward error protection vs. LC3 net bitrate
The following error profiles ware applied (for PLR and BER see Table 2):
• clean
• RSSI 56 dB
• RSSI 48 dB
• RSSI 30 dB
The averaged MOS-LQO scores for the mentioned configurations (as given by the Recommendation ITU-T P.863 [i.9]
assessment) are shown in Figure 8, Figure 9, Figure 10 and Figure 11. The bars with different colours refer to the
different signal strength classes, indicated by their RSSI. It turns out that a BCH code can improve the perceived audio
quality substantially if it is adjusted to the signal strength.
ETSI
17 ETSI TR 103 590 V1.1.1 (2018-09)
5,00
4,00
O
3,00
Q
L
-
clean
S
O
2,00
56dB
M
1,00
48dB
0,00 40dB
Forward error protection scheme

Figure 8: Averaged MOS-LQO scores for NB at 32 kbps gross bitrate
5,00
4,00
O
3,00
Q
L
-
clean
S
O
2,00
56dB
M
1,00
48dB
0,00 40dB
Forward error protection scheme

Figure 9: Averaged MOS-LQO scores for WB at 32 kbps gross bitrate
5,00
4,00
O
Q3,00
L
-
clean
S
O2,00
56dB
M
1,00
48dB
0,00
_ 2 4 6 8 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 0 2 40dB
e 0 0 0 0 1 1 1 1 1 2 2 2 2 2 3 3 3 3 3 4 4 4 4 4 5 5
n h h h h h h h h h h h h h h h h h h h h h h h h h h
c c c c c c c c c c c c c c c c c c c c c c c c c c
o
b b
n b b b b b b b b b b b b b b b b b b b b b b b b

Forward error protection scheme

Figure 10: Averaged MOS-LQO scores for WB at 64 kbps gross bitrate
ETSI
18 ETSI TR 103 590 V1.1.1 (2018-09)
5,00
4,00
O
Q
3,00
L
-
clean
S
2,00
O
56dB
M
1,00
48dB
0,00
_
2 4 6 8 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 0 2
40dB
e
0 0 0 0 1 1 1 1 1 2 2 2 2 2 3 3 3 3 3 4 4 4 4 4 5 5
n h h h h h h h h h h h h h h h h h h h h h h h h h h
c c c c c c c c c c c c c c c c c c c c c c c c c c
o
b b b b b b b b b b b b b b b b b b b b b b b
n b b b
Forward error protection scheme

Figure 11: Averaged MOS-LQO scores for SWB at 64 kbps gross bitrate
The sweet spots, i.e. the best objective quality for the different channel conditions relative to the FEC schemes are given
in Table 3, Table 4, Table 5 and Table 6.
Table 3: Best objective quality for different channel conditions relative to
FEC scheme for NB @ 32 kbps
NB @ 32 kbps Clean 56 dB 48 dB 40 dB
max MOS-LQO score 3,43 3,28 2,96 1,75
FEC none bch03 bch08 bch18
net bitrate 32,0 kbps 28,8 kbps 24,8 kbps 16,0 kbps

Table 4: Best objective quality for different channel conditions relative to
FEC scheme for WB @ 32 kbps
WB @ 32 kbps Clean 56 dB 48 dB 40 dB
max MOS-LQO score 4,11 3,88 3,53 1,95
FEC none bch03 bch08 bch17
net bitrate 32,0 kbps 28,8 kbps 24,8 kbps 17,6 kbps

Table 5: Best objective quality for different channel conditions relative to
FEC scheme for WB @ 64 kbps
WB @ 64 kbps Clean 56 dB 48 dB 40 dB
max MOS-LQO score 4,32 4,09 4,03 2,64
FEC bch01 bch09 bch21 bch44
net bitrate 62,4 kbps 54,4 kbps 43,2 kbps 23,2 kbps

Table 6: Best objective quality for different channel conditions relative to
FEC scheme for SWB @ 64 kbps
SWB @ 64 kbps Clean 56 dB 48 dB 40 dB
max MOS-LQO score 4,66 4,42 4,30 2,73
FEC none bch04 bch21 bch44
net bitrate 64,0 kbps 60,0 kbps 43,2 kbps 23,2 kbps

Figure 12 outlines the achievable quality (by means of the MOS-LQO scores) for the investigated codec configurations
and signal strengths in a graphical manner.
ETSI
19 ETSI TR 103 590 V1.1.1 (2018-09)
Achivable quality for different configurations and different
channel conditions
4,5
3,5
O 3
Q
L
-
2,5
S
O
M
1,5
0,5
clean 56dB 48dB 40dB
signal strength / RSSI
NB@32kbps WB@32kbps WB@64kbps SWB@64kbps

Figure 12: Achievable quality for different configurations at various channel conditions
According to those results, the following recommendations can be made:
• 64 kbps allows for all tested conditions a better quality than 32 kbps.
• For 64 kbps, the SWB mode provides better quality than the WB mode for all channel conditions.
• For 32 kbps, the WB mode provides better quality than the NB mode for all channel conditions.
4.3.2.2 Dynamic rate switching of source and channel coder
4.3.2.2.1 General
Based on the given results, a bit allocation of source and channel coding dependent on the present channel condition is
advisable. As a consequence, the source coder needs to support seamless rate switching, as e.g. LC3 does.
4.3.2.2.2 Graceful degradation at DECT range limit
For the tested channel conditions (clean channel, 56 dB, 48 dB, 40 dB), a graceful degradation toward the range limit
can be confirmed. For 32 dB, no meaningful output can be achieved anymore.
This behaviour is also confirmed by Recommendation ITU-T P.800 [i.10] experiment documented in clause 4.4.2.3.
4.3.2.2.3 Audio bandwidth switching
According to Figure 12, no bandwidth switching seems advisable for 32 kbps. This is also confirmed by the P.800 [i.10]
experiment documented in clause 4.4.2.3.
For 64 kbps a bandwidth switching from SWB to WB may lead to a slight improvement for the 40 dB channel
condition.
ETSI
20 ETSI TR 103 590 V1.1.1 (2018-09)
4.3.2.2.4 Potential channel coder configuration
The DECT channel analysis provided in clause 4.3.2.1 indicates that four error protection (EP) classes could be used for
a real DECT system for normal slots. Table 7 outlines a potential configuration of the EP classes for LC3, operating for
WB signals at 32 kbps.
Table 7: EP configuration for DECT normal slots
EP class Codec rate EP rate Correction Detection Complexity
capability capability (note 2)
(note 1)
ep_class_1 28 800 3 200 0 bits 99,999 995 % 4,31
ep_class_2 27 200 4 800 2 bits 99,999 995 % 7,77
ep_class_3 24 800 7 200 7 bits 99,999 995 % 7,76
ep_class_4 17 600 14 400 16 bits 99,999 995 % 9,12
NOTE 1: Detection capability is only relevant if bit errors occur. It includes the cases where a
correctable frame is marked as un-correctable as well as an un-correctable frame is decoded
as a valid frame.
® ®
NOTE 2: Complexity in million cycles per second measured with an Arm Cortex -A9 simulator
(simulating the A9MPx1) (see note 3). Code only moderately optimized, especially for low
protection classes.
® ®
NOTE 3: Arm Cortex -A9 simulator is an example of a suitable product available commercially. This
information is given for the convenience of users of the present document and does not
constitute an endorsement by ETSI of this product."
...