<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Klaus Darilion wrote:
<blockquote cite="mid:46B02786.3030804@pernau.at" type="cite">
<pre wrap="">
Zoa wrote:
</pre>
<blockquote type="cite">
<pre wrap="">Because the jitter buffer will never be size zero if its a fixed jitter
buffer and will fluctuate if its a variable length jitter buffer.
So whatever you choose, the timing will be altered.
</pre>
</blockquote>
<pre wrap=""><!---->
I would think that a fixed jitter buffer is a fixed jitter buffer -
thus, only buffering voice frames adding a fixed delay (thus reducing
jitter but increasing delay) and do not change the audio data at all.
A dynamic jitter buffer of course is bad. Because of the dynamic
behavior the audio samples have to be played back sometimes
faster/sometimes slower.
</pre>
</blockquote>
Hi, <br>
<br>
Finally as i wrote yesterday, the problem with faxes was in noise
filter in sangoma HW echocan (fixed now).
<blockquote cite="mid:46B02786.3030804@pernau.at" type="cite">
<pre wrap="">Thus, why does static jitter buffer manipulate the audio?
</pre>
</blockquote>
Static nor adaptive buffer does not manipulate frames. I've compare it
frame by frame when the frame come to buffer and when it goes out. <br>
<blockquote cite="mid:46B02786.3030804@pernau.at" type="cite">
<pre wrap="">
@Martin: What happens if you call the Milliwatt() application via a
jitter-buffer link? Do you hear frequency change?
</pre>
</blockquote>
Asterisk jitterbuffer is not capable of frequency change. It only
buffers frames and if needed it sends info to translater to do PLC (if
it is capable). <br>
<br>
<br>
<br>
</body>
</html>