[Asterisk-Users] BAD/GOOD Echo Cancel

[Steve Underwood]

Kevin P. Fleming wrote:

> Ken D'Ambrosio wrote:
>
>> They use the G.168-2002 algorithm; I, personally, had never heard of it
>> before, but I bounced it off a friend of mine (he's a hardware architect
>> for a major VoIP switch manufacturer -- they sell to places like Time
>> Warner), and he was of the opinion that G.168 is the _ONLY_ algorithm.
>
>
> You might want to do some research before making statements like 
> that... G.168 is a set of recommendations of characteristics that an 
> echo canceler implementation should provide, and a set of tests 
> designed to determine whether an implementation is compliant with the 
> recommendations. The current recommendations were released in 2004.
>
> G.168 is _not_ an algorithm, so your friend must have a slight 
> misunderstanding of the situation. There are many (dozens, maybe 
> hundreds) of implementations out there claiming to be G.168 compliant, 
> but most of them have not been tested by a third party to verify those 
> claims.

More than that, in their fine print some only claim to pass maybe two or 
three of the tests. There is nothing that defines what you must achieve 
before you can claim G.168-2002 compliance.

Regards,
Steve