[asterisk-dev] [Code Review] Fax Gateway Implementation T30<->T38

[Steve Underwood]

On 04/15/2011 01:47 AM, Pavel Selivanov wrote:
> 05.04.2011 23:33, Klaus Darilion wrote:
>>> disable echo cancellation. Echo cancellers that are G.168 compliant
>>> detect CED tones generated by *called* FAX terminals and take action
>>> themselves to reconfigure properly so as to not interfere with a FAX
>>> transaction.
>> One more note: the faxdetect code (triggered by CNG) does *NOT*
>> What about the software echo cancellers in DAHDI, are they G.168 
>> compliant?
> dahdi (dahdi-base.c) is searching for CNG tone, unless you disable it 
> (in dahdi_config.h).
> Once CNG is detected - dahdi will switch echo canceller of.
>
> There are a number of echocancellers.
> None of the free echo cancellers g.168 compliant (can not pass all of 
> the g.168 tests).
> The best of the free echo cancellers (my opinion), is OSLEC.
> It will not pass all g.168 tests, but it works very well (converge 
> speed, attenuation, DTD).
> http://www.rowetel.com/blog/oslec.html
> Try it, I think it will fit your needs.
Actually, practically any echo canceller can pass most of the G.168 
tests, because the poor wording in the test spec allows you to cheat. 
OSLEC was tested in a totally honest way, because there was no point in 
trying to cheat with a free open source project. This explains why it 
fails more tests than many commercial cancellers. Some commercial 
cancellers are so poor they can't even pass the tests with the usual 
cheats. They say "Passes G.168 *", and * then says something like 
"except for tests 1, 2, 3, 4, 5".

In practice OSLEC is as solid as most commercial cancellers. None of the 
other free echo cancellers are really worth bothering with.

G.168 has some real value, but it is a widely ridiculed document in the 
industry.

Steve