Polycom OMNIVS500U 用户手册

下载
页码 13
AEC
Nonlinear
Processing
Mic
Speaker
Echoes
Near speech
Far speech
Far speech
Near speech +
echoes + distortion
Far End
Near End
XMT
RCV
Distortion
 
Figure 1:  Illustration of the effects of AEC operation and room acoustics on the transmitted speech.  The far speech that travels 
through the receive path is not modified as it passes through the AEC.  In an echo canceller that is poorly designed, there may be 
residual echoes as well as distortion added to the near speech signal (these effects are described in detail later).  This degrades 
the speech that is transmitted, so that the poor audio quality is noticed on the far end.  
 
 
Why not just use a speakerphone? 
 
Speakerphones are half-duplex, which means that only one person can talk at a time.  The speakerphone 
determines which side is active (or louder) by comparing the signal levels on both sides.  It turns off the 
other side until the louder side is finished.  Once one side has “captured” the circuit, most speakerphones 
do not permit any sort of interruption.  This inhibits the natural flow of conversation, especially if one party 
is long-winded.   
 
Acoustic Echo Cancellation vs. Line Echo Cancellation 
 
Acoustic echo cancellation and line echo cancellation both address similar problems, and are often based 
on the same technology.  However, a line echo canceller generally cannot replace an acoustic echo 
canceller, because acoustic echo cancellation is a more difficult problem.  With line echo cancellation there 
are generally only one or two reflections from 
telephone hybrids or impedance mismatches in the 
telephone line.  These echoes are usually delayed 
by less than 32 ms, and do not change very 
With acoustic echo cancellation, the echo path 
is complex and can vary continuously as people 
move around the room.