Plot twist: If I play the same files via the dialplan, no popping sound. So, if I call the locally running asterisk on the extension that goes to the app, I hear a pop at the beginning of the even first prompt. It's using STREAM FILE via asterisk-java. However if I add another extension:
exten => 555,1,Answer
exten => 555,n,While(true)
exten => 555,n,Background(/sounds/XXX)
exten => 555,n,EndWhile
exten => 555,n,Hangup
using Background or Playback, it plays the first and subsequent times with no pop.
On Tue, Oct 17, 2017 at 8:48 PM Naftoli Gugenheim <naftoligug@xxxxxxxxx> wrote:
On Tue, Oct 17, 2017 at 1:08 PM Barry Flanagan <barryf-lists@xxxxxxxxxxx> wrote:Hi, thanks for replying.
1. I have opened the files in audacity. When I play it on my computer I can't hear anything like the popping sound. Visually I'm not really sure what to look for. I mean I can't see a some that looks like it would account for such a loud pop.
I have had this in the past. Look at the very end of the recording in audacity - if you see where the main voice ends - starts to flatline - and if there is any small non-flatline after it, delete it. I have found that when someone is making recordings the keyboard click to stop the recording can get picked up. I usually manually delete from the end of the recording I want, and replace it with silence in Audacity just to be sure. Often you will not here it during playback in Audacity, but it is very apparent over a phone line.Is there any measurement that can be said? I'm dealing with a very large number of files. I would prefer if it's something I could check in code.Also, according to this, what if a file doesn't end with silence at all (recording stops pretty simultaneous to the end of speaking... or just background noise) -- would it be the same problem? Is the rule that you need to allow audio files to begin with a ramp up from silence and end with a ramp down to silence? And if so are there any numbers that can be put on it, like ramping duration, or duration of actual silence at the beginning and end? And measured in wall time or samples (if that isn't a dumb question)? Note: I'm not looking for a scientific number that less is bad and more is unnecessary. But it would be nice to have numbers that are known to be safe, without being too perfectionist...Hope this helps_______________________________________________-Barry Flanagan
asterisk-app-dev mailing list
asterisk-app-dev@xxxxxxxxxxxxxxxx
http://lists.digium.com/cgi-bin/mailman/listinfo/asterisk-app-dev
_______________________________________________ asterisk-app-dev mailing list asterisk-app-dev@xxxxxxxxxxxxxxxx http://lists.digium.com/cgi-bin/mailman/listinfo/asterisk-app-dev
