[Asterisk-Users] Why echo occurs

[Robert Hajime Lanning]

<quote who="Steve Underwood">
> Maybe http://www.grandstream.com.cn/BT100_Spec_cn.pdf is a bit
> confusing if you can't read Chinese, but I think G.168 should
> be easy to identify :-)

ok, I did miss that.  Then again, the grandstream does have a
speaker phone.  I guess the problem is that I don't know of a
SIP hardphone that doesn't have a speaker phone.

> Acoustic spill gives basically the same effects as hybrid echo, except
> acoustic spill tends to be more variable over time. Hybrid echo also
> bounces back and forth when both ends are causing echo, but the first
> echo is so much stronger than the subsequent ones that you tend not to
> notice them.
>
> I have worked on echo cancellation, and I know the acoustic spill
> issue is serious. In early GSM phones it was often easy to fool the
> canceler, and GSM to GSM calls would suffer really awful echo. They
> seem to have improved the cancelers a lot in the last few years,
> and its rare to get this problem today. This is a broad issue. Echo
> cancelers have generally improved a lot. The latest version of
> G.168 is a very different document from the early versions, and
> incorporates tests for a lot of the problem issues found in earlier
> canceler designs.

I would expect it to be a problem in the GSM (cell) phones.  They
are too small to get proper acoustic separation.

I am talking about the phones that are physically designed the same
as analog phones.  Why do we not hear this echo in the analog device?
But, we do when it is digital.  This type of echo would always be
far end, as the near end would always be seen as sidetone.

> How do you control acoustic spill within a phone through the use of
> directional microphones? Adjusting gains mitigates the issue a bit,
> but is hardly a solution. These are just bodges, not solutions.

You can say the same about echo cancelers.  They patch the symptom,
not the cause.

--
END OF LINE
       -MCP