[Asterisk-Dev] Echo canceller testing

Steve Underwood steveu at coppice.org
Wed Aug 10 06:49:49 MST 2005


Iain Barker wrote:

>Hi Devs,
>
>I am intersted in knowing the status and quality of the echo canceller software in the current Asterisk code. I looked through the sources and the wiki pages http://www.voip-info.org/tiki-index.php?page=Asterisk+echo+cancellation
>
>Looks like there are a number of different s/w ecan implementations; I also note that newer Digium cards have ecans in hardware DSP rather than using the host s/w.
>
>For each of the ecan versions:
>
>Has anyone done objective G.165 or G.168 certification testing?
>Has anyone done MoS subjective (PAMS, PSQM) listener effort testing?
>What is the CPU overhead for each s/w ecan channel (MIPS usage)?
>  
>
None of the cancellers in * will pass any of the tests in G.168. G.165 
isn't really that relevant. Cancellers don't really impact things like 
PESQ much, unless they go horribly wrong. EC takes a lot of CPU power, 
hence the new cards with hardware EC. How good that hardware EC is, I 
have no idea.

Regards,
Steve




More information about the asterisk-dev mailing list