Talk:Echo cancellation

Incorrect statement regarding echo suppression
The following paragraph, modified by Darmstro at 02:04, 10 June 2006 has an incorrect statement:

Echo suppression may have the unwanted side-effect of removing valid voice signals from the transmission. This can cause audible signal loss that is called "clipping." For that reason, echo suppression has been obsoleted by echo cancellation.it is done with due respect

In modern software applications echo cancellation and suppression work in conjunction to remove echo. The variability of computer hardware creates all kinds of difficulties that generally make it impossible for echo cancellation alone to perform sufficiently well. I have changed the paragraph to:

Echo suppression may have the unwanted side-effect of removing valid signals from the transmission. This can cause audible signal loss that is called "clipping." In an ideal situation then, echo cancellation alone will be used. However this is insufficient in many applications, notably software phones. Here, echo cancellation and suppression can work in conjunction to achieve acceptable performance. --Andrew J. MacDonald 03:13, 18 August 2006 (UTC)

Merge echo suppression with echo cancellation
I agree with the merging of the two entries. I understand "echo suppression" to be synonymous with "echo cancellation"... the only difference is the term used. BOTH attempt to achieve complete removal of echo in telecommunication circuits. Does anyone have a clearer definition of the difference between suppression and cancellation than I'm assuming? --220.233.25.235


 * Neither is ever complete. Suppression comes closer, but is less often used than it once was.  Echo suppression was developed in the 1950s and 60s for intercontinental telephony.  It turns off the receive leg when the talk leg is talking, thus sacrificing full duplex in order to suppress echo.  Echo cancellation was developed in the 1970s.  It figures out the time/amplitude characteristic of the echo, listens to the talk leg, produces an equal and opposite echo signal, and feeds that down the listen leg, thus cancelling the echo.  It was greatly expanded in the 1980s for use in modems, and nowadays the principal users of echo cancellation are DSL modems.  Jim.henderson 01:40, 29 March 2007 (UTC)

Basic Article Layout
I think the problem with this article is the lack of any description of how talker A signal gets mixed in with talker B signal and heard. In Telephony this is the two wire to four wire converters in the network. In acoustical, the end "room" is the two wire converter. So we have 3 cases. A) Telephones (far echo), the signal from person A goes to person B but some is reflected back at the far 2/4 wire converter. B) Acoustic, the reflection is the room and C) the reflection is in the local handset. Hmmm. Any suggestions on how to first introduct how signals may be reflected to cause an echo?. That would be a good introductory piece --User: jlpayton —The preceding  signed but undated.

Don't merge echo suppression with echo cancellation
The distinction between the two concepts is important. It is important to understand the difference and to know which one you are talking about. Having two separate articles means that when you search on one, that is what you get. It is good, to explain the difference between them in each article, of course and to refer to the other with SEE ALSO. —Preceding unsigned comment added by Braithwa (talk • contribs) 06:03, 18 March 2008 (UTC)


 * So, Braithwa, Jim.henderson and I agree that supression and cancellation are separate processes and probably deserve separate articles. As of today, you'll notice that today there is no Echo suppression article so there's no merge to do. --Kvng (talk) 16:18, 2 September 2008 (UTC)
 * Well, I didn't mean that the distinction between the two methods calls for separate articles. These two methods tackle the same problem differently, is all.  However, it seems unfair or inefficient that a lot of work done in good faith in creating the other article, even if it was a misguided or inadvertent Content forking, should be made to disappear without trace, discussion, or even prior notice.  Jim.henderson (talk) 14:21, 8 September 2008 (UTC)
 * I didn't realize that the Echo suppression article ever existed. Is there no way to retrieve it? --Kvng (talk) 20:38, 8 September 2008 (UTC)
 * Oops! Kindly disregard.  It's Echo suppressor and still alive.  Anyway I'm pretty sure that's how I went wrong.  As for a merger, alas I have overfilled my plate of Wikipedia things to do, and won't spare any of my limited brain power on this question at least until lovely summery weather fades away and allows me to work into my backlog of unprocessed outdoor Wikiphotography.  Jim.henderson (talk) 12:35, 14 September 2008 (UTC)

July 2013
You should have redirected Echo suppression to Echo suppressor back in 2008. Well I finally did it today. As for merging, both articles are totally dead. There are no references in either of them and no visible progress in the last 5 years. Time to merge and add references. -- Petri Krohn (talk) 01:02, 25 July 2013 (UTC)

NLP (Non-Linear Processing)
I've removed the following text as it appears to be original research.


 * NLP (Non-Linear Processing) algorithms attempt to adjust preemptively to changes in the acoustic environment. They can be applied in varying degrees of aggressiveness to avoid its overuse. More aggressive NLP increases the chance of over-cancellation, which will actually add echo. By keeping in mind how AEC cancels a signal by adding one that is 180° out of phase but of the same amplitude, one sees that over-cancellation can result  in a small, out of phase echo. The NLP algorithm is distinct from the AEC algorithim in that inaccuracies with NLP are more audible than those with the primary AEC, which is also subject to over/under cancellation.

Non linear processing is generally used to suppress any residual echo that is not perfectly cancelled by the AES. This non-linear processing is essentially a noise gate. Other forms of NLP are used to assist with double talk detection. --Kvng (talk) 00:09, 15 January 2010 (UTC)

Suspected COI
An edit made to this article on 5 September 2008 was performed from an Israeli IP address (217.132.143.17), whose sole history on Wikipedia is adding two external links to the website of Solicall Ltd., an Israeli company. I suspect that Conflict of Interest may be at play. 69.180.205.28 (talk) 19:30, 8 December 2010 (UTC)

Full duplex echo cancellation
This article does not discuss full duplex echo cancellation and echo canceler used in video conferencing system and associated techniques such as background vs foreground echo cancellation. Is this the right place to do it? Could not find much anywhere. PCleve (talk) 21:22, 2 January 2011 (UTC)


 * This is also known as acoustic echo cancellation (AEC). I agree, the article is thin on this. I've made some minor improvements. The topic is substantial enough to merit its own article. Perhaps someone will create one someday. --Kvng (talk) 14:42, 3 January 2011 (UTC)

Broken external link
The "Echo cancellation technology (IBM)" link goes to a 404 page after you get through the SSL certificate warnings.

144.32.48.87 (talk) 09:12, 5 September 2011 (UTC)


 * It's working for me at the moment. --Kvng (talk) 12:30, 7 September 2011 (UTC)

July 2013 merge proposal
User:Petri Krohn has added merge banners proposing to merge Echo suppressor into Echo cancellation. As discussed earlier on this talk page, echo cancellation and echo suppression are different techniques for avoiding echoes in voice communications. You may argue that there should be one article that covers more than one technique to achieve the same end. I think it would be potentially confusing to call the combined article either Echo suppressor or Echo cancellation. ~KvnG 14:37, 27 July 2013 (UTC)
 * There is no need for separate articles on "Echo suppressor" and "Echo suppression" Echo suppressors can be covered in the History section of Echo cancellation. -- Petri Krohn (talk) 06:43, 28 July 2013 (UTC)
 * There seems to be confusion again. We are now talking about three different articles: Echo suppressor, Echo suppression, and Echo cancellation. There is no difference between the -or and the -ion version, and there should be a redirection in one of them, linking to the other. Echo cancellation is not echo suppression. Neither is echo suppression an "old version" of echo cancellation. In modern systems, both of them are used, even simultaneously. There is no reason to merge the two subjects of suppress(or/ion) and cancellation. 82.113.121.132 (talk) 22:58, 3 November 2013 (UTC)
 * There are only two articles; Echo suppression has always been a redirect. ~KvnG 13:58, 7 November 2013 (UTC)
 * I (82.113.121.132) was referring to Petri Krohn, and I agree that there "is no need for separate articles on 'Echo suppressor' and 'Echo suppression'", but I do not agree that they should "be covered in the History section of Echo cancellation", since suppression and cancellation are different approaches to the same problem. In particular, echo cancellation does not make echo suppression obsolete. --195.81.5.154 (talk) 14:26, 13 November 2013 (UTC)

Is Echo suppression part of Echo cancellation or the other way around? Or are they separate topics? If so, there must be a general topic – by whatever name – that covers both. -- Petri Krohn (talk) 20:12, 13 November 2013 (UTC)


 * They are two different techniques for eliminating echoes. I am personally not aware of a single overarching term for this topic. ~KvnG 21:31, 17 November 2013 (UTC)
 * After a bit of Googling, I came across the term "echo avoidance". For lack of a better catch-all term, I think this is a pretty good term for the general topic. --Bigpeteb (talk) 18:18, 12 December 2013 (UTC)
 * Echo avoidance seems to come up more frequently in association with the Precedence effect than it does with the topics being discussed here. In the current cotex, do you see it used by anyone else besides Asterisk? ~KvnG 17:54, 15 December 2013 (UTC)

I support this merge for several reasons. One, both articles are fairly small and have almost no references. Two, both articles reference each other and give at least a partial explanation of the other; this is needless duplication. Three, taken together, the articles are generally describing methods of echo avoidance, and spend comparatively little time discussion their respective methods (suppression or cancellation). Therefore, I suggest that the pages be merged, and moved to echo avoidance, where it can be treated as a general topic and the specific methods described with proper context given. --Bigpeteb (talk) 18:19, 12 December 2013 (UTC)
 * I appreciate these arguments but I am concerned that with the Echo avoidance proposal we would be creating new terminology. ~KvnG 17:54, 15 December 2013 (UTC)
 * Well, how about something as simple as Echo suppression and cancellation? That makes it clear that they're two distinct but related topics, and it would be natural for the article to start with an overview of echo, and then discuss and compare the two different methods of dealing with it. --Bigpeteb (talk) 14:03, 17 December 2013 (UTC)
 * Good idea. I can see how this could work. I don't see any other open objections. ~KvnG 19:59, 21 December 2013 (UTC)