Transcript Acoustic interface performance parameters

Using subsystem performance parameters to
optimize end-to-end performance
Scott Pennock
Senior Hands-Free Standards Specialist
QNX Software Systems
Outline
Introduction
Subsystems
Subsystem performance parameters
Optimizing end-to-end (e2e) performance
Conclusions
2
All content copyright QNX Software Systems
Introduction
ITU-T G.799.2 defines a mechanism that can be used to improve the
end-to-end (e2e) speech communications performance of telephone
connections
Each Signal Processing Equipment (SPE) on the telephone
connection communicates its Signal Processing Functions (SPFs)
capabilities to all other SPEs
Based on coordination rules, each SPE will either “enable” or
“disable” its own SPFs to optimize e2e performance
MSC-O
BSC-O
MGW-O
MSC-T
SC-T
Server
Server
MGW-T
SPNE-T
Circuit-switched
network
IP packet
network
MS-O
Land
User
Media Path
Signalling Path
ITU-T G.799.2/Figure I.6  Mobile-to-land call through an IP packet network and a circuit-switched network
3
All content copyright QNX Software Systems
Introduction (continued)
This presentation will describe how e2e performance can be further
optimized by:
Expanding the list of what is considered SPE to include:
Acoustic interfaces
Signal transport equipment (e.g., speech codec, de-jitter buffer, etc.)
Measurement devices (e.g., ITU-T P.561, etc.)
Adding subsystem performance parameters to the information
exchanged between SPEs
4
All content copyright QNX Software Systems
Terminals are becoming dynamic subsystems
5
All content copyright QNX Software Systems
Subsystem performance parameter considerations
When is performance measured?
Classification into performance levels apriori
Real-time measurements of parameters
How is performance measured?
Signal-based analysis
Parameter-based analysis
6
All content copyright QNX Software Systems
Acoustic interface performance parameters
7
All content copyright QNX Software Systems
Example of QoS levels in the Send direction
for the Acoustic interface
QoS level
Measurement parameter
1
2
3
4
5
Speech-to Noise Ratio (SpNR)
SpNR > 25 dB
SpNR > 20 dB
SpNR > 15 dB
SpNR > 10 dB
SpNR > 5 dB
Send Frequency Response (SFR)
SFR mask 1
SFR mask 2
SFR mask 3
SFR mask 4
SFR mask 5
Send ReVerberation (SRV)
SRV > 15 dB
SRV > 10 dB
SRV > 5 dB
SRV > 0 dB
SRV > -5 dB
Send Distortion + Noise (SDN)
SDN < 1%
SDN < 2%
SDN < 3%
SDN < 4%
SDN < 5%
Speech-to-Echo Ratio (SpER)
SpER > 5
SpER > 0
SpER > -5
SpER > -10
SpER > -15
Echo Tail Delay (ETD)
ETD < 5 ms
ETD < 10 ms
ETD < 15 ms
ETD < 20 ms
ETD < 25 ms
Echo Tail coherence (ETco)
ETco > 95%
ETco > 90%
ETco > 85%
ETco > 80%
ETco > 75%
8
All content copyright QNX Software Systems
Speech enhancement performance parameters
9
All content copyright QNX Software Systems
Local transport performance parameters
10
All content copyright QNX Software Systems
Network transport performance parameters
11
All content copyright QNX Software Systems
Optimizing e2e performance
There are many ways subsystem performance parameters can be
used to optimize performance:
Determining if acoustic interface can support requested service (e.g.,
wideband, high SNR, low echo path distortion, stereo telephony, etc.)
Determining if speech enhancement can handle acoustic interface
(e.g., hands-free)
Selection of best acoustic interface (could be due to hardware,
usage, or environment)
Setting speech enhancement parameters based on acoustic interface
or speech codec
Setting radio link parameters based on speech codec/SNR
Selecting speech codec based on local transport error pattern
Selecting speech codec based on upstream (or downstream) speech
codec to minimize transcoding distortion
Turning “off” poorly performing SPFs
G.799.2 allows for defining “attributes” for a SPF, but need
standardized measurement parameters to be effective
12
All content copyright QNX Software Systems
Conclusions
G.799.2 has limited ability to optimize e2e performance
e2e performance can be further optimized by:
Expanding the list of devices considered SPE to include:
Acoustic interfaces
Transport equipment
Measurement devices
Adding subsystem performance parameters, or QoS levels, to the
information exchanged between SPE
Terminal equipment is no longer “static”; it is dynamic with different
subsystems interoperating
There is still much work to be done to identify the validate
subsystem performance parameters
However, this should not prevent designing the information
exchange between SPEs so that when performance parameters
become available, there is a mechanism for utilizing them
13
All content copyright QNX Software Systems
Thank you
Scott Pennock
Senior Hands-Free Standards Specialist
[email protected]
14
All content copyright QNX Software Systems