CS6422
CS6422
speaker, the reflections reaching the microphone
the level of signal without the echo canceller com-
should present no more than 0.9 V to the Acous- pared to the level of signal with the echo canceller.
rms
tic ADC. In fact, it is advisable to allow 6 dB or
When measuring ERLE, it is important that any po-
even 12 dB of margin, such that in the above exam-
tential signal loops be broken; so to measure the
ple, the signal present at the Acoustic ADC is
ERLE of the Acoustic Canceller, the NO output
should be disconnected from the rest of the net-
250 mV
.
rms
After this coupling level is established, the desired work. This will prevent feedback which could oc-
signal gain must be established. To continue from cur when all of the CS6422’s failsafes are disabled.
the previous example, the transmit gain must be ad-
The following example outlines the steps necessary
justed to make sure the near-end talker is easy to
to measure the ERLE of the acoustic echo cancel-
hear at the far-end. If the signal from the near-end
ler.
talker clips at the ADC, it is not significant to the
It is important to choose a good test signal for the
echo path because the AEC should not be updating
tests to be valid. As mentioned in Section 4.1.1.2,
anyway.
“Adaptive Filter”, the CS6422 does not work opti-
In general, to minimize noise system gain should
mally with white noise. The best signal to use
be concentrated before the ADC. However, this is
would be a repeatable speech signal, like a record-
ing of someone counting or saying “ah.”
not practical in all cases, mostly because of the cou-
pling constraint. The CS6422 offers the AGC’d
Use the Microcontroller Interface to disable trans-
gains provided by TVol and RVol to help provide
mit and receive suppression, half-duplex, and the
the desired transmit and receive gains.
Network Echo Canceller. The gains should be set
The CS6422 offers two different programmable
appropriate for good system performance.
gain sources: TGain/RGain and TVol/RVol. TGain
The first measurement is a baseline figure of per-
and RGain provide analog gain at the input to the
formance with no echo canceller. Use the Micro-
ADC of 0 dB, 6 dB, 9.5 dB, or 12 dB. TVol and
controller Interface to clear the acoustic canceller
RVol introduce digital gain and attenuation in 3 dB
coefficients. Inject the test signal at NI and measure
steps. The difference is significant in that the digital
the rms voltage at NO. This measurement gives the
baseline coupling level (denominator).
gain will gain up the noise of the ADC as well as the
desired signal, whereas the analog gain will not.
Furthermore, gains introduced by TVol and RVol Use the Microcontroller Interface to set the acous-
will not result in clipping, since both gains are tic canceller coefficients to normal which will al-
AGC’ed, unlike the gains at TGain and RGain
which are not.
low the adaptive filter to adapt. Inject the test signal
at NI and allow a few seconds for the filter to adapt.
Again, measure the rms voltage at NO. This mea-
surement gives the cancelled echo level (numera-
tor).
4.3.2 Testing Issues
The following tests are suggestions for measuring
echo canceller and half-duplex performance.
Convert both voltages to decibels and subtract the
echo cancelled level from the baseline level to cal-
culate the ERLE. At the factory, with known good
components, we typically see 30 dB of ERLE with
speech.
4.3.2.1 ERLE
Echo Return-Loss Enhancement (ERLE) is a mea-
sure of the attenuation that an echo canceller pro-
vides. The number is an expression of the ratio of
40
40
DS295F1