Confidential and Proprietary - Qualcomm Technologies, Incorporated. All Rights Reserved. Enhanced Voice Services – EVS IWAENC 2014 Presentation September 2014
Confidential and Proprietary - Qualcomm Technologies, Incorporated. All Rights Reserved.
Enhanced Voice Services – EVS IWAENC 2014 Presentation
September 2014
2
Topics of this Presentation
Benefits of EVS
Standardization framework
Specifications
Algorithmic overview and special aspects
Delay and complexity
Embedding EVS in 3GPP system
Quality characterization
Deployment / commercialization aspects
3
kbps4.75 12.2
6.6 23.85kbps
5.9 kbps 128
AMR
AMR-WB
EVS
EVS – Next Gen 3GPP Speech Coding for Improved User Experience in Telephony
AMR
AMR-WB
EVS
Qu
ality
1995 2002 2014
4
Standards
Commercial
AMR AMR-WB EVS
3GPP Voice Service Evolution
WCDMA+
2002 2014
2015
2000
20102001
5
EVRC Family AMR Family
Different Voice Solutions
Carrier Grade Voice Best Effort Voice
OTT-VOIPIMS/RCS VoIPCS Voice
VoLTEVoMBB
Ensures Consistent & Seamless
Rich Voice Experience
EVS FamilyCustomized Family Internet Family G.7xx Family
CDMA UMTS
OPERATOR
PROVIDED
SERVICE
TAPI
Dialer
OPERATOR
PROVIDED
SERVICE
Operator
Client
3RD PARTY
PROVIDED
SERVICE
3rd party
Client
PORTAL
PROVIDED
SERVICE
HLOS
Client
VoWANVoWiFi
Very reliable, fully interoperable, but
lacks personalization
Interoperable but Limited to
VoLTE Coverage
Lacks IP Mobility, Connectivity & Interop
with other mVoIP
6
What is EVS?
3GPP Speech Conversation / Telephony Coder
EVS – Enhanced Voice Services− Next generation 3GPP speech coding
− Following the successful FR, HR, EFR, AMR, AMR-WB codecs
− Designed for packet-switched networks / mobile VoIP
− VoLTE is a key target application
− Application in other networks
− AMR-WB interoperable mode
− Rel-12 Work Item in 3GPP preceded by a Study Item TR 22.813
Key features− Super-wideband speech (32 kHz sampling) – improved speech quality
− Source-controlled variable bit-rate operation – improved capacity
− Designed for VoIP – improved robustness
− Improved music performance
− Wide bit-rate range and all bandwidths for maximum flexibility
− Backward interoperable mode to AMR-WB
Standardization process− Qualification phase
− Selection phase
− Characterization phase
7
3GPP EVS is the Next Generation Speech CoderSpeech quality determines user experience
− Ensuring voice quality on new VoLTE deployments
− EVS addresses all networks – mobile VoIP with QoS, best effort VoIP, CS
3GPP goals of Enhanced Voice Services (EVS) standardization
− Feature-rich coder
− Designed for VoIP applications such as MTSI in TS 26.114
− It is further desirable that the benefits of EVS are available for users of other networks such as CS
− NB, WB, SWB bandwidths, FB optional, high robustness mode
− Bit rates: 7.2, 8, 9.6, 13.2, 16.4, and 24.4 kb/s gross rates that comply with LTE TBSs; 32, 48, 64, 96, 128 kb/s
− Quality improvements – improving user experience
− Better quality in VoLTE and UMTS (with no new RAB)
− Evolution path: EVS provides SWB at around 13 kbps – lower rate and lower delay SWB than other industry
coders without sacrificing quality
− Better quality for music and mixed content in conversational applications
− Capacity improvements – increasing system efficiency
− VBR at 5.9 kbps provides high capacity mode
− Robustness improvements – optimized behavior in VoIP applications
− More robust NB/WB through significantly better error resilience
− High robustness mode
8
EVS- Enhanced Voice Services
EVS @13.2kbps provides Super Wideband Voice Quality at comparable bit-rate to AMR & AMR-WB
AMR
AMR-WB
EVS
Be
tter V
oic
e Q
ua
lity
NB
WB
SW
B
Improves voice clarity and
intelligibility @12.65 kbps
More natural sounding speech
and improved music quality
@13.2 kbps
Toll quality narrowband voice
@ 12.2 kbps
The Ultimate Codec of Choice for Mobile Telephony
9
Better Capacity
Superior Quality
Enhanced Error ResiliencySuper Wideband: 13.2 – 128 kbps
Wideband: 5.9 – 128 kbps
Narrowband: 5.9 – 13.2 kbps
Support of Source-Controlled
Variable Bit Rate operation
Extended audio bandwidth: 50 Hz to 16 kHz
Better quality NB and WB Voice compared to
AMR & AMR-WB
Entertainment quality music coding
Optimized for VoLTE and Circuit
Switched Networks
Improved Robustness to packet
loss compared to AMR-WB
EVS Benefits
Wideband(Voice)
Narrowband(Voice)
Super Wideband(Voice and Music)
Low frequencies increases naturalness, presence
and comfort
High frequencies improves voice clarity and
intelligibility
16kHz~7kHz
Reproduces better
audio and music
10
Enhanced In-Call Music Quality
Enhanced Error Resiliency
2
2.5
3
3.5
23.85 13.2
AMR-WB EVS
Error Resilience Improvement for 3GPP
delay loss profile (6% FER)
2
2.5
3
3.5
19.85 23.85 13.2
AMR-WB EVS
Music coding performance (3% FER)
2
3
4
12.65 23.85 13.2 12 16
AMR-WB EVS Opus
Speech coding performance
(with background car noise)
Superior Voice Quality
EVS Performance
11
Wh
y D
ep
loy
EV
S?
EVS – Solution for Each Situation
Jitter Buffer Management
NB WB SWB FB Stereo
New EVS
Modes
(CBR)
New EVS
Modes
(VBR)
AMR-WB
Interop
Modes
New EVS Modes (CBR)
7.2-128 kb/s
New EVS
Modes
(VBR)
New EVS Modes
7.2-128 kb/s
New EVS Modes
(optional)
New EVS Modes
(optional)
Speech Speech Speech Speech Music Speech Speech Music Speech Music Speech Music
7.2-13.2 kb/s 5.9kb/s
(avg)
6.6-23.85
kb/s
7.2-128 kb/s 7.2-128
kb/s
5.9 kb/s
(avg)
13.2-128
kb/s
13.2-128
kb/s
7.2-128 kb/s 7.2-128 kb/s 7.2-128 kb/s 7.2-128 kb/s
Better Capacity
Same NB/WB quality as legacy
Better Music
Near AAC Quality at much
lower delay
Better Quality
Same capacity as legacy
NB/WB
Improved Error Resilience
Much better than AMR-WB, VoIP
Optimizations
12
3GPP EVS Standardization Process in Rel-12
Requirements phase – design constraints and performance requirements
Candidate coders− 13 companies submitted a candidate by 16 November 2012
− Ericsson, Fraunhofer, Huawei, Motorola, Nokia, NTT, NTTDoCoMo, Orange, Panasonic, Qualcomm, Samsung, VoiceAge, ZTE
− Standardization by competition
Qualification phase− Aim is to keep the most promising 5 candidates for selection
− Extensive testing− 12 experiments, each candidate is tested in-house and in another listening lab− Global Analysis Lab performs collection and analysis of test results
− Qualification meeting in March 2013 agreed in 5 candidates
All proponents announced a collaborative development of a joint candidate
Selection phase – single joint candidate− Codec selection is based on extensive testing in neutral listening labs
− Selection meeting in August 2014 agreed to adopt the joint candidate as EVS standard
− Agreement on most EVS specifications
Characterization phase− Aim is to test the coder performance for all conditions and special signals / conditions
Approval of remaining EVS Specifications and Technical Report
13
2014Jun Jul Aug Sep Oct Nov Dec
2015
Approval of EVS Technical Report and floating-point spec
Dec 10
SA approval of EVS standard
Sep 15
3GPP SA4#80bis: codec selection and approval of specifications
Aug 30
3GPP SA4#80
Aug 4
Submission of EVS executable for testing
June 27
EVS Prototypes Available for Preliminary Lab/Field Testing EVS Engineering Build Available For IOT and Field Trials
EVS Rel-12 Standardization Timeline
Selection Testing Characterization Testing
Nov 6
3GPP SA4#81
EVS over 3G UTRAN CS Work Item (Rel-13)
2015
14
EVS
Qualcomm
Samsung
Nokia
Panasonic
NTT
NTT DoCoMo
Orange
ZTE
Huawei
Ericsson
VoiceAge
Fraunhofer
EVS Is A Global Collaboration
Broad Industry Support Across the Ecosystem
Keys For Successful Deployment
• Codec hw/sw support (i.e., chipset, IMS/RCS client,
voice pre/post-proc, etc.)
• Super Wideband terminal acoustic designs
• Infra support (IMS, gateways, etc)
• Test Equipment Support (call box, IMS, SWB
acoustics, voice quality)
• EVS support in voice services outside of mobile
ecosystem (e.g., wireline VoIP, Enterprise VoIP &
Video Telephony, etc.)
12 Party
Collaboration
15
EVS Design Requirements
Narrowband
(0-4 KHz) Coding of
Speech better than
AMR
Wideband
(0-8 kHz) Coding of
Speech better than
AMR-WB; inclusion of
AMR-WB IO
Superwideband (0-
16 kHz) Coding of
Speech better than
AMR-WB
Improved Error
Resilience
for both Circuit
Switched and
Packet Switched
Communication
and
VoIP Capability
Source
Controlled
Variable Rate
Coding
Improved
Coding of
Music
for In-call Music
(Music on hold
and Ringback)
Constraints on
Frame Length,
Max.
Algorithmic
Delay,
Complexity,
JBM, Rate
Switching,
PLC, RTP
Payload
Format,
VAD/DTX/CNG
16
EVS Requirements in SWB at Low RatesCategory Bitrate (kbit/s) FER DTX Requirements
Clean speech
-26,-16,-36dBov
13.2 0% On†/Off NWT G.722.1C @ 32
16.4 NWT G.722.1C @ 48
24.4 NWT G.718B @ 36
Clean speech
-26 dBov
13.2 x=3%,
6%
Off
On† for 13.2
NWT G.722.1C @ 48, x% FER
16.4 NWT G.719 @ 48, x% FER
24.4 NWT G.719 @ 56, x% FER
Noisy Speech (Car, Office,
Street)
-26 dBov
13.2 0% On‡/Off NWT G.722.1C @ 24 when EVS DTX off
NWT AMR-WB @19.85 DTX on when EVS DTX on
16.4 NWT G.722.1C @ 32 when EVS DTX off
NWT AMR-WB @23.05 DTX on when EVS DTX on
24.4 NWT G.722.1C @ 48 when EVS DTX off
NWT AMR-WB @23.85 DTX on when EVS DTX on
Noisy Speech (Car, Office,
Street)
-26 dBov
13.2 x=3%,
6%
Off
On‡ for 13.2
NWT G.722.1C @ 24, x% FER and DTX off
NWT AMR-WB @19.85, x% FER and DTX on when EVS DTX on
16.4 NWT G.722.1C @ 32, x% FER
24.4 NWT G.722.1C @ 48, x% FER
17
3GPP EVS Specifications
Spec No. Title Status: agreed
TS 26.441 EVS Codec General Overview For approval
TS 26.442 EVS Codec ANSI C code (fixed-point) For approval
TS 26.443 EVS Codec ANSI C code (floating point) Draft
TS 26.444 EVS Codec Test Sequences For approval
TS 26.445 EVS Codec Detailed Algorithmic Description For approval
TS 26.446 EVS Codec AMR-WB Backward Compatible Functions For approval
TS 26.447 EVS Codec Error Concealment of Lost Packets For approval
TS 26.448 EVS Codec Jitter Buffer Management For approval
TS 26.449 EVS Codec Comfort Noise Generation (CNG) Aspects For approval
TS 26.450 EVS Codec Discontinuous Transmission (DTX) For approval
TS 26.451 EVS Codec Voice Activity Detection (VAD) For approval
TS 26.114 MMTel CR Part for approval;
part draft
TR 26.952 EVS Codec Performance Characterization Not existing yet
18
Transmit Side
19
Receive Side
20
EVS Bit Rates and Supported Bandwidths
Source codec bit-
rate (kbit/s)
Supported audio
bandwidthsDTX availability
5,9 (SC-VBR) NB, WB
Yes (Always On; Bit
Rates are 2.8, 7.2, 8.0
kbit/s)
7,2 NB, WB Yes
8.0 NB, WB Yes
9,6 NB, WB, SWB Yes
13,2 NB, WB, SWB Yes
13,2 (channel
aware)WB, SWB Yes
16,4 NB, WB, SWB, FB Yes
24,4 NB, WB, SWB, FB Yes
32 WB, SWB, FB Yes
48 WB, SWB, FB Yes
64 WB, SWB, FB Yes
96 WB, SWB, FB Yes
128 WB, SWB, FB Yes
Source codec bit-rates for the EVS codec
The SID bit rate for EVS primary modes is 2.4 kbit/s.
Source codec bit-rate (kbit/s)
6,6
8,85
12,65
14,25
15,85
18,25
19,85
23,05
23,85
Source codec bit-rates for the AMR-WB Interoperable
Modes of the EVS codec
The SID bit rate in AMR-WB IO corresponds to AMR-WB (2.0 kbit/s).
21
Encoding Modes
Bitstream
Multiplex
LP-based
Coding
Frequency
Domain
Coding
Inactive Signal
Coding/CNG
Common Processing
Signal
Resampling
Signal
Analysis
Command Line
Parameters
Classifier Information
Classifier
Decision
Input
Signal
Hybrid coding scheme combining linear predictive (LP) coding techniques based upon
ACELP (Algebraic Code Excited Linear Prediction), predominantly for speech signals
Transform coding method, for generic content, as well as inactive signal coding in
conjunction with VAD/DTX/CNG (Voice Activity Detection/Discontinuous Transmission/
Comfort Noise Generation) operation
EVS codec is capable of switching between these coding modes without artefacts
22
ACELP
Bitstream
Multiplex
High Band
Parameterization
Residual
Quantization
LP Filter
Analysis
Bandwidth
Splitter
Low Band
Signal
Input
Signal
LP Coefficient
Estimation &
Interpolation
High Band
Signal
The input signal is split into high frequency band and low frequency band paths
The high-frequency portion of the signal is represented with several different parametric
representations. The parameters vary as a function of the bit-rate and the residual
quantization strategy. The transmitted parameters include some or all of spectral
envelope, energy information and temporal evolution information.
In the LP based core, the configuration of the LP coefficient estimation, parametric HF
representation and the residual quantization is similar to those of AMR-WB
23
Frequency Domain Operation
Separation into a control layer and a signal processing layer
The control layer performs signal analysis to derive several control and configuration
parameters for the signal processing layer. The time-to-frequency transformation is
based on the Modified Discrete Cosine Transform (MDCT) and provides adaptive time-
frequency resolution. The control layer derives measures of the time distribution of the
signal energy in a frame and controls the transform.
The MDCT coefficients are quantized using a variety of direct and parametric
representations depending upon bit rate signal type and operating mode.
24
Inactive Signal Coding
When the codec is operated in DTX on mode the signal classifier depicted in Figure 1
selects the discontinuous transmission (DTX) mode for frames that are determined to
consist of background noise. For these frames a low-rate parametric representation of the
signal is transmitted no more frequently than every 8 frames (SID frame).
The low-rate parametric representation is used in the decoder for comfort noise
generation (CNG) and includes parameters describing the frequency envelope of the
background signal, energy parameters describing the overall energy and its time
evolution.
25
Source Controlled Variable Bit-Rate Coding
VBR coding describes a method that assigns different number of bits to a speech frame
in the coded domain depending on the characteristics of the input speech signal
This method is often called source-controlled coding
− Typically, a source-controlled coder encodes speech at different bit rates depending on how the current
frame is classified, e.g., voiced, unvoiced, transient, or silence. Note that DTX operation can be combined
with VBR coders in the same way as with Fixed Rate (FR) coders; the VBR operation is related to active
speech segments.
The VBR solution provides narrowband and wideband coding using the bit rates 2.8, 7.2
and 8.0 kbps and produces an average bit rate at 5.9 kbps.
Due to the finer bit allocation, in comparison to Fixed Rate (FR) coding, VBR offers the
advantage of a better speech quality at the same average active bit rate than FR coding
The benefits of VBR can be exploited if the transmission network supports the
transmission of speech frames (packets) of variable size, such as in LTE and UMTS
networks.
26
EVS – Source Controlled Variable (SC-VBR) Bit-Rate Encoder
Pre-processing
Pre-emphasis -> Voice-Activity-
Detection -> LPC analysis ->
Open-Loop Pitch Estimation
EVS speech type
classification
EVS-SID (Silence Descriptor)
NELP mode
Inactive
speech
Unvoiced-
speech
Voiced-
speechHas there
been 2
consecutive
EVS-Voiced
frames ?
PPP WI mode
EVS Voiced mode
EVS Transient mode
First Voiced-
speech after
voiced onset
NO
YES
EVS Generic mode
All other
speech
27
SC-VBR Salient FeaturesEnabled use of low-rate modes for certain speech types
PPP WI (Prototype Pitch Period Waveform Interpolation) mode for coding voiced speech (2.5 kbps)
Noise-Excited linear prediction (NELP) mode for unvoiced speech (2.3 kbps for narrow-band and 2.4 kbps
for wide-band)
PPPWI and NELP are low-rate LP residual (excitation) signal coding schemes
Illustration – improved system efficiency through lower average bit-rate
7.36
12.65 12.81
7.36
12.5913.04
1
2
3
4
5
EVRC-WB AM R-WB at
12.65k bps
VM R-WB,
M ode 0
EVRC-WB AM R-WB at
12.65k bps
VM R-WB,
M ode 0
CT1 - Input Level and FER Conditions CT2 - Noise Conditions
Qu
ali
ty
0
2
4
6
8
10
12
14
Avera
ge D
ata
Rate
(kb
ps)
Speech Quality Active ADR
28
Advanced Error Resilience for VoLTE – Partial RedundancyAvailable in a special mode of the EVS coder (channel aware mode) offers improved performance under packet loss conditions in a VoIP system
EVS offers partial redundancy based error robust channel aware mode at 13.2 kbps for both wideband and super-wideband audio bandwidths.
Packets arrive at the decoder with random jitters in their arrival time; packets may also arrive out of order at the decoder.
Since the decoder expects to be fed a speech packet every 20 ms to output speech samples in periodic blocks, a de-jitter buffer is required.
Partial copies of the current speech frame are piggybacked on future speech frames, without increase in the total bit rate. Partial copy of a lost frame can be retrieved by polling the de-jitter buffer.
Frame nFrame n+3 Frame n+2 Frame n+1n
• Adding partial copy of previous critical frame for better error resilience
0
0.5
1
1.5
2
2.5
3
3.5
4
AMR-WB 12.65 3% FER AMR-WB 23.85 3% FER EVS-SWB-AdvErrRes-on -13.2 kbps10% FER
EVS-SWB-AdvErrRes-off -13.2 kbps10% FER
EVS-SWB-13.2 kbps with Advanced Error Resilience at 10% packet loss is equivalent in quality to AMR-WB-23.85 kbps @ 3% packet loss
29
Advanced Error Resilience in EVS fits well for VoLTE
VoLTE network
AMR-WB
Packets
Received
Packets
DeJitter
Buffer
AMR WB
Encoder
AMR WB
Decoder
VoLTE network
EVS Packets Received
Packets
DeJitter
Buffer
EVS
Encoder
EVS
Decoder
30
Frame Loss Concealment at Decoder Side
The EVS codec includes frame loss concealment algorithms
An extrapolation algorithm estimates the signal in a lost frame
− For the LP based core this estimation operates on the last received residual and LP
coefficients.
− For the frequency domain core in some cases the last received MDCT coefficients are
extrapolated and in addition the resulting time domain signal is guaranteed to give a smooth
time evolution from the last received frame into the missing frames.
Once the frame loss is recovered, i.e., the first good frame is received the codec
memory is updated and frame boundary mismatches towards the last lost frame
are minimized.
For situations of sustained frame loss the signal is either faded to background
noise or its energy is reduced and finally muted when no reasonable extrapolation
can be assumed.
31
EVS Delay and Complexity
Sampling frequencies− 48, 32, 16, 8 kHz at input and output
EVS algorithmic delay− The coder operates on 20 msec frames
− The algorithmic delay is less or equal to 32 msec
− For 48, 32, 16 kHz sampled output, the delay consists of one 20 msec frame, 0.9375 msec delay of input resampling filters on the encoder-side, 8.75 msec for the encoder look-ahead, and 2.3125 msecdelay of time-domain bandwidth extension on the decoder-side resulting in 32 msec
− For 8 kHz sampled output, the decoder delay is reduced to 1.25 msec needed for resampling using a complex low-delay filterbank, resulting in 30.9375 msec overall algorithmic delay.
EVS complexity− With all features supported and measured according to EVS-8b, the worst case complexity of
the coder is 85.9 WMOPS which splits up to 56.7 WMOPS for encoder (24.4 kbit/s SWB with DTX on) and 29.2 WMOPS for decoder (48 kbit/s SWB with DTX off, FER=6%).
− The coder uses 175KW of RAM (with no JBM included), 157 KW of ROM, and 116 KW of Program ROM.
− The JBM solution was measured to consume 18 WMOPS and 49 KW RAM.
32
Embedding EVS in 3GPP System
Multimedia Telephony (MMTel) Servíce over IMS uses EVS
− Support of AMR and AMR-WB for VoIP in 3GPP networks prior to Rel-12
− Inclusion of EVS support for VoLTE in Rel-12
EVS for 3G UTRAN Circuit-Switched networks
− Agreed Work Item
− Rel-13 time frame
− Enable users of UMTS to make benefit of the enhanced quality due to EVS
33
EVS Quality – Clean Speech Inputs (Proprietary Tests)
2
2.5
3
3.5
4
4.5
5
0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 64
Sub
ject
ive
DM
OS
Test
Bit Rate (kbps)
AMR-WB
EVS-SWB
EVS-WB
Opus
G718B
G719
G722.1c
Note: OPUS bit rates are target average bit rates. The true average and standard deviations of the frame bit rates vary across bit rates
34
Official EVS Selection Tests in 3GPP – Experiments Exp. Content Methodology # of Exp.N1 NB clean speech under clean channel condition including input level dependency ACR 1
N2 NB clean speech under impaired channel conditions including delay/jitter profiles ACR 1
N3 NB noisy speech under clean channel condition and impaired channel conditions DCR 1
N4 NB mixed content and music under clean channel condition and impaired channel conditions including delay/jitter profiles
ACR 1
W1 WB clean speech under clean channel condition including input level dependency ACR 1
W2 WB clean speech under impaired channel conditions including delay/jitter profiles ACR 1
W3 WB noisy speech under clean channel condition DCR 1W4 WB noisy speech under impaired channel conditions including delay/jitter profiles DCR 1
W5 WB mixed contents and music under clean channel condition DCR 1W6 WB mixed contents and music under impaired channel conditions DCR 1
W7 WB mixed contents and music under impaired channel conditions including delay/jitter profiles
DCR 1
I1 AMR-WB IO clean speech under clean channel condition including input level dependency
ACR 1
I2 AMR-WB IO clean speech under impaired channel conditions ACR 1I3 AMR-WB IO noisy speech under clean channel condition DCR 1I4 AMR-WB IO noisy speech under impaired channel conditions DCR 1I5 AMR-WB IO mixed contents and music under clean channel condition DCR 1
I6 AMR-WB IO mixed contents and music under impaired channel conditions DCR 1
S1 SWB clean speech under clean channel condition including input level dependency DCR 1
S2 SWB clean speech under impaired channel conditions including delay/jitter profiles DCR 1
S3 SWB noisy speech under clean channel condition DCR 1S4 SWB noisy speech under clean channel condition DCR 1S5 SWB noisy speech under impaired channel conditions DCR 1S6 SWB mixed contents and music under clean channel condition DCR 1
S7 SWB mixed contents and music under impaired channel conditions including delay/jitter profiles
DCR 1
Total 24
35
4.3
8
4.0
4
3.7
1
2.9
0
1.9
4
1.2
9
3.7
6
3.7
9 3.9
2
3.7
1
3.5
1
3.3
3
3.6
1
3.4
5
4.0
7
3.9
8 4.1
5
3.9
6
3.9
0
4.2
8
4.2
4 4.3
7
4.3
4
4.1
2 4.2
5
4.2
1 4.3
8
4.3
8
4.2
3
4.2
6
3.9
8 4.1
0 4.1
9
4.0
3
4.2
4
4.2
1
4.4
2
4.1
8 4.3
1
4.2
7 4.3
8
4.4
3
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
4.50
5.00
Experiment N1 -- NB Clean Speech, Clean Channel
36
4.3
1
4.0
4
2.8
7
2.0
2
1.3
7
1.14
3.5
1
3.9
8
4.0
5
4.0
3
4.0
6
3.9
9
2.7
9
3.2
5
3.3
1
3.1
7 3.2
8
3.2
9
3.7
2
4.1
1
4.0
8
4.1
7
4.1
9
4.2
2
4.3
0 4.4
5
4.4
8
4.0
7
3.9
3 4.0
3
4.2
4
4.3
2 4.4
4
4.4
5
3.8
4
4.0
2
4.1
1 4.2
4 4.3
5
4.3
1 4.4
6
4.0
9
4.0
2
4.0
7 4.2
2
4.2
1 4.3
1 4.4
1
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
4.50
5.00
Experiment W1 -- WB Clean Speech, Clean Channel
37
4.8
5
4.0
6
3.4
5
2.4
0
1.2
7
1.0
3
2.5
7
3.4
9 3.6
8
1.9
2
2.9
6
2.9
2
3.7
8
4.1
1
4.1
3
4.3
9
3.7
1
4.2
2
3.2
4
3.5
9
3.6
4
3.9
1
3.1
9
3.6
6
0.00
1.00
2.00
3.00
4.00
5.00
6.00
Experiment W7 -- WB Music & Mixed Content at FER
38
4.7
9
4.6
9
4.6
3
3.5
9
2.0
7
1.2
8
3.9
8
4.2
4
4.6
0
4.6
0 4.7
3
4.6
8
3.6
2 3.7
8
4.4
0 4.5
3
3.9
4 4.1
6
4.6
5
4.6
1
4.4
8
4.5
3 4.7
3
4.6
6
4.7
6
4.5
8
4.6
2 4.7
5
4.5
5
4.5
7
4.5
9
4.6
2
4.6
2
4.5
0
4.5
6 4.7
3
0.00
1.00
2.00
3.00
4.00
5.00
6.00
Experiment S1 -- SWB Clean Speech, Clean Channel
39
4.4
2
4.3
6
4.2
8
3.9
7
2.3
8
1.3
8
3.7
7
4.0
5 4.1
8
4.3
9
4.3
5
4.3
5
3.6
3
3.5
9
3.6
4
4.0
6
4.0
6
4.2
9
4.3
5
4.1
3
4.3
8
3.9
0
3.8
9
4.1
5
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
4.50
5.00
Experiment S3 -- SWB Speech / Background Noise (Street, SNR=20dB)
40
4.7
2
4.7
0
4.7
4
3.9
2
1.7
9
1.10
2.8
2
3.0
9 3.2
8
3.4
9
3.5
6 3.7
2
3.6
5
3.0
5
2.5
5
2.8
2
2.8
8 3.0
8 3.2
4
3.2
3
3.2
1
2.7
9
3.7
6 3.8
5
4.2
3
3.7
2
4.2
8
4.1
4
3.8
1
3.5
3
3.4
6
3.9
5
3.3
3
3.9
5
3.7
8
3.5
2
0.00
1.00
2.00
3.00
4.00
5.00
6.00
Experiment S5 -- SWB Speech / Background Noise, FER (Car, SNR=15dB)
41
4.6
7
4.6
1
4.6
3
4.4
1
3.2
6
1.2
1
2.3
6
2.5
9
3.6
7 3.7
7
3.9
8 4.1
1
1.9
4
2.1
7
3.1
8 3.2
9
3.6
8
3.5
7 3.6
8
4.1
9
3.3
9
3.6
3
4.1
1
4.1
2
4.3
7
4.3
2
3.2
2
3.0
3
3.3
2
3.6
7 3.7
9 3.8
9
3.9
3
3.0
5
4.0
6
4.5
2
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
4.50
5.00
Experiment S7 -- SWB Music & Mixed Content, FER
42
2014Jun Jul Aug Sep Oct Nov Dec
2015
Approval of EVS Technical Report and floating-point spec
Dec 10
SA approval of EVS standard
Sep 15
3GPP SA4#80bis: codec selection and approval of specifications
Aug 30
3GPP SA4#80
Aug 4
Submission of EVS executable for testing
June 27
EVS Prototypes Available for Preliminary Lab/Field Testing EVS Engineering Build Available For IOT and Field Trials
EVS Rel-12 Standardization Timeline
Selection Testing Characterization Testing
Nov 6
3GPP SA4#81
EVS over 3G UTRAN CS Work Item (Rel-13)
2015
43
EVS Deployment
EVS targets – VoLTE and other networks− VoLTE mass deployment is on-going
− EVS in 3G UTRAN CS networks
VoLTE trials− Goal is to make EVS available for VoLTE trials and deployment
− Pre-commercial phase during 2015
Qualcomm is key partner in EVS deployment− Serve ecosystem by making EVS codec available in mobile chipsets
− Pre-standard version available for VoLTE trials and branded voice services
Deployment of standardized version can begin− IOT, field testing is first step
Qualcomm supports best-in-class voice quality for IMS based voice service deployments on LTE with a complete suite of tools and features, including EVS, IMS client, and voice enhancement.
44
All data and information contained in or disclosed by this document is confidential and proprietary information of Qualcomm
Technologies, Incorporated and all rights therein are expressly reserved. By accepting this material the recipient agrees that this
material and the information contained therein is to be held in confidence and in trust and will not be used, copied, reproduced in
whole or in part, nor its contents revealed in any manner to others without the express written permission of Qualcomm Technologies,
Incorporated.
© 2013 Qualcomm Technologies, Inc. All rights reserved. Qualcomm, Snapdragon, and Gobi are trademarks of Qualcomm Incorporated, registered in
the United States and other countries. Trademarks of Qualcomm Incorporated are used with permission. Other products and brand names may be
trademarks or registered trademarks of their respective owners.
QUALCOMM Technologies, Incorporated, 5775 Morehouse Drive, San Diego, CA 92121-1714
Thank you