PSB 4860
Functional Description
Two pairs of coefficients, GDS/PDS when speech is detected, and GDN/PDN in case of
noise, offer a different echo handling for speech and non-speech.
With speech, even if very strong resonances are present, the performance will not be
worsened by the high GDS needed. Only when speech is detected, a high reserve
prevents clipping. A time period ET [ms] after speech end, the parameters of the
comparator are switched to the “noise” values. If both sets of the parameters are equal,
ET has no function.
Table 5 Speech Comparator Parameters
Parameter # of bytes Range
Comment
G
1
– 48 to + 48 dB
Base Gain
GDS
1
0 to 48 dB
Gain Reserve (Speech)
PDS
1
0.025 to 6 dB/ms Peak Decrement (Speech)
GDN
1
0 to 48 dB
Gain Reserve (Noise)
PDN
1
0.025 to 6 dB/ms Peak Decrement (Noise)
ET
1
0 to 992 ms
Time to Switch from speech to noise
parameters
2.1.3.3 Attenuation Control
The attenuation control unit controls the attenuation stages GHX and GHR and performs
state switching. The programmable attenuation ATT is completely switched to GHX
(GHR) in receive state (transmit state). In idle state both GHX and GHR attenuate by
ATT/2.
In addition, attenuation is also influenced by the automatic gain control stages (AGCX,
AGCR).
State switching depends on the signals of one speech comparator and the
corresponding speech detector. While each state is associated with the programmed
attenuation, the time is takes to reach the steady-state attenuation after a state switch
can be programmed (TSW).
If the current state is either transmit or receive and no speech on either side has been
detected for time TW then idle state is entered. To smoothen the transition, the
attenuation is incremented (decremented) by DS until the evenly distribution ATT/2 for
both GHX and GHR is reached.
Table 6 shows the parameters for the attenuation unit. Note that TSW is dependent on
the current attenuation by the formula Tsw = SW × ATT .
Semiconductor Group
37
10.97