Version: 1.18
PRELIMINARY/CONFIDENTIAL
TrueSpeech® Co-Processor
9.21.6
Get Electrical Echo Canceller Quality Factor Command
This command is not available in DSVD Speakerphone mode.
1110
101
0 0000 000
(8-1)
HIWORD
(0)
(15-12)
(11-9)
Status:
Electrical Echo Canceller Quality Factor
(15-0)
This command returns either the low-order (HIWORD = 0) or high-order (HIWORD = 1) 16 bits of the electrical
(line-side) echo-cancelled 32-bit quality factor. The quality factor is a measure of how well the training pulse
succeeded in configuring the electrical echo canceller coefficients.
During training, the echo canceller sends the training pulse and receives an echo of the training pulse back from the
line. The echo canceller coefficients are adjusted to achieve maximum cancellation of the training pulse echo.
If no echo occurs (for example, if a 4-wire telephone connection is used), all the echo canceller coefficients are set to
zero, no cancellation is required (perfect cancellation is achieved; there is no reflected signal), and the quality factor
is zero in all 32 bits.
If the training pulse is sent while there is a high-level incoming signal (such as dial tone), the echo canceller is unable
to cancel the incoming signal (training pulse echo + dial tone) and the quality factor becomes very large
(> 100000H). This indicates that training failed and that full duplex operation of the speakerphone can not be
activated (loop attenuation must be greater than or equal to 12 dB).
If the training pulse echo is successfully cancelled, the quality factor should be less than 10000H (all high-order bits
zero). Under these circumstances, full duplex speakerphone operation may be selected (loop attenuation less than 12
dB).
If the training pulse has not been sent, the quality factor will be set to FFFFFFFFH.
The numbers given here are approximate, and depend largely on the design of the external hardware. The user is
expected to experiment and discover suitable threshold values appropriate to his hardware environment.
Note: This command reports the current value of the Electrical Echo Canceller (EEC) Quality Factor.
Since training of the EEC takes several milliseconds, it is possible to read this value before training
has completed. In this case the Quality Factor from the state prior to training will be read. The Host
should allow about 50ms to elapse between sending the command to train the EEC and attempting to
read the new Quality Factor. During the training period, the Host should not issue any additional
commands to the CT8022, as this will delay or disrupt the training process.
The additional AEC control features described in the DSVD AEC section may also be used in standalone
speakerphone mode unless otherwise noted.
CT8022A11AQC FW Revision 0118 DSP GROUP, INC., 3120 SCOTT BOULEVARD
SANTA CLARA, CA 95054 PH: 408 986 – 4300 FAX: 408 986 – 4490
145
All specifications are subject to change without prior notice.
ETC [ ETC ]
