AEC | Double Talk | ERLE | G.168 | LEC | NLMS

Hi Dhar and to other group members,

the form of the exponential filter is not shown in your post. can you pleas=
e re-post it again?

and by the way, what should be the gain factor of the rectangular sliding w=
indow? I doubt that it is 1.

and lastly, what measurement method must the detector do? is it the RMS met=
hod specified in the recommendation?

thanks to you all and have a good day!:)

-keone-

--- In e...@yahoogroups.com, "Dharmendra Lingaiah"
<dharmendra@...> wrote:
>
> Hello Vinod Prakash,
> 
> I hope you will get some better insight after reading my
> mail..................
> 
> An exponential filter is used to sharpen spectral features; the constant 
is
> varied to change the filter shape.
> A exponential filter is of the form  where  is the filter constant and X
is
> the array (i.e., data file) whose X range is normalized between 0 and 1.
> As you may note.., In Section  6.4.1.2.2,  the exponential filter is
> replaced by a rectangular sliding window of the same length.
> 
> The output of the Exponential Filter is then fed to a DETECTOR,( which the
> G.168 Standard specifies as DETECTOR-INDICATOR ).
> 
> I believe the implementation strategy for the DETECTOR should be based on
> the
> THRESHOLD of detection of the SPECTRAL PEAKS obtained from the EXPONENTIAL
> FILTER......;-)
> 
> Dharmendra Lingaiah
> 
> -----Original Message-----
> From: vinprakash366 [mailto:vinod.prakash@...]
> Sent: Friday, February 01, 2002 11:57 AM
> To: e...@yahoogroups.com
> Subject: [echocancel] Power measurement
> Hi all,
> 
>       In the G.168(2000) standard, under section 6.4.1.2.1, level
> measurement device, i am a little confused abt the detector-indicator
> block. How should this be implemented in order to decect the power in
> dBm0??. Also, is the 1-pole exponential filter mentioned in the block
> a unity gain filter??
> 
> thanks in advance,
> Vinod
>

______________________________
New Code Sharing Section now Live on DSPRelated.com. Learn about the Reward Program for Contributors here.



(You need to be a member of echocancel -- send a blank email to echocancel-subscribe@yahoogroups.com )