Cirrus Logic AN168 - Manual (Page 19)

AN168
AN168REV2 19
Similar to harmonic distortion, these ‘buzzing’
artifacts cause elevated levels of residual echo
because they result from a non-linear phenomenon
that the echo canceler cannot model.
The test for buzzing is similar to the frequency
sweep distortion test. A full-scale sine-wave is
injected into the speaker-driver input (or the NI
input on the CS6422), and swept slowly between
100 Hz and 4 kHz. This sweep can be performed
manually. As the frequency is swept, listen for the
induced buzzing vibrations. Try placing your hand
on the speaker housing to test for mechanical
vibrations as well. When a buzz is detected, try to
isolate the mechanical cause of the induced
vibration and fix it. Typical solutions involve
inserting spongy foam gaskets between connecting
pieces of plastic or inserting spongy foam between
the speaker’s mounting rim and the housing.
1.5.3 Acoustic ERLE
In order to tell if the system is properly designed, it
is useful to measure the ERLE (Echo Return Loss
Enhancement) that the echo canceler is able to
provide. We present two methods to test ERLE, the
White Noise RMS method and the Loop Gain
method.
The White Noise RMS method measures the
difference in the spectral power level of the signal
with the echo canceler present in the acoustic path
and removed. The Loop Gain method measures the
worst-case ERLE that the echo canceler is able to
achieve by measuring the difference in loop gain
with the echo canceler present in the loop versus
absent.
Both tests are useful. The White Noise RMS
method requires a white noise source and a
band-limited RMS voltage measurement
instrument. The Loop Gain method requires no
additional hardware beyond the ability to configure
the CS6422.
1.5.3.1 White Noise RMS Method
In this test, the CS6422 in configured such that the
transmit and receive suppressors and half-duplex
are disabled. A 2.8 Vpp white noise signal is
injected at NI, and the band-limited (preferably
C-message) RMS voltage is measured at NO with
the echo canceler enabled (ACC = ‘normal’) and
disabled (ACC = ‘cleared).
The test procedure is as follows:
1) Set up the speaker and microphone and adjust
the acoustic coupling to -9 dB by the methods
Coupling-Weighted THD (%)
0
0.2
0.4
0.6
0.8
1
1.2
0 500 1000 1500 2000 2500 3000 3500 4000
Frequency (Hz)
Figure 14. Coupling-Weighted THD+N