<div dir="ltr">Ok.<br><br>Thank you for your help.<br><br><div class="gmail_quote">On Sat, May 8, 2010 at 1:55 PM, Martin Vit <span dir="ltr"><<a href="mailto:vit@lam.cz">vit@lam.cz</a>></span> wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
On Sat, May 8, 2010 at 2:34 PM, mosbah abdelkader<br>
<div class="im"><<a href="mailto:mosbah.abdelkader@gmail.com">mosbah.abdelkader@gmail.com</a>> wrote:<br>
</div><div class="im">> Thank you Martin,<br>
><br>
> So the MOS-LQE does not inform bout payload itself but predicts the MOS<br>
> based on networks metrics<br>
<br>
</div>yes exactly. LQE is Listen Quality Emodel (E-model is parametric model<br>
which takes into account some more parameters. I've used static<br>
parameters except for loss and burstiness. So if your network is<br>
stable and you want to measure MOS, there is no way how to do that on<br>
unknown samples. You can do only automated tests.<br>
<div class="im"><br>
<br>
> and P862 and P863 uses also payload (voice) to<br>
> calculate the MOS. Is it true what I have understood.<br>
><br>
<br>
<br>
</div>yes, P.862 (PESQ) compare two samples. Original and degraded (and<br>
about 20 seconds). P.563 does not need original sample and can predict<br>
only degraded sample (only about 20 seconds). It cannot analyze whole<br>
conversation. Both methods is suited for automated tests with specific<br>
samples. These objective methods compare new codecs, transmittion path<br>
etc. etc.. It will never work as real live passive monitoring. I've<br>
used P.862 to calibrate MOS-LQE.<br>
<font color="#888888"><br>
MV<br>
</font><div><div></div><div class="h5"><br>
<br>
> Best regards.<br>
><br>
> On Sat, May 8, 2010 at 1:17 PM, Martin Vit <<a href="mailto:vit@lam.cz">vit@lam.cz</a>> wrote:<br>
>><br>
>> Hello,<br>
>><br>
>> I've choosen only MOS-LQE because it is calculated only on network<br>
>> parameters, which is loss, burstinnes and delay (which is converted to<br>
>> loss by jitterbuffer simulator). It does not takes into account voice<br>
>> (payload). There is no effective objective methods (today) which<br>
>> predicts MOS. Only ITU-T P.862 and P.563 which is patented and it can<br>
>> analyze only 20 seconds samples. I've tried implementing P.563 and it<br>
>> is not usable for real live use, only for automated tests which is not<br>
>> in my interest now (and because of patents). I've calibrated MOS-LQE<br>
>> with polynomial functions using P.862 PESQ. I will write more on<br>
>> <a href="http://voipmonitor.org" target="_blank">voipmonitor.org</a> documentation once I've found more time.<br>
>><br>
>> I'm using voipmonitor on central gateway and succesfully monitoring<br>
>> all SIP traffic and filtering calls by the worst MOS. So yes, you can<br>
>> use that tool for measuring quality of IP network in realtime. If you<br>
>> save PCAP files, you can analyze it with wireshark in more depth.<br>
>><br>
>><br>
>><br>
>><br>
>><br>
>> On Sat, May 8, 2010 at 1:42 PM, mosbah abdelkader<br>
>> <<a href="mailto:mosbah.abdelkader@gmail.com">mosbah.abdelkader@gmail.com</a>> wrote:<br>
>> > Hello,<br>
>> ><br>
>> ><br>
>> > First, thank you for your great job.<br>
>> ><br>
>> ><br>
>> > I want to know why you have choosed to calculate only MOS-LQE. Why you<br>
>> > have<br>
>> > only used G107. Is that model suitable for VoIP operators to have a<br>
>> > calculated QoS value so they can confirm their quality.<br>
>> ><br>
>> ><br>
>> > Thanks again and best regards.<br>
>> ><br>
><br>
><br>
</div></div></blockquote></div><br></div>