On 12/6/06, John Lange <[EMAIL PROTECTED]> wrote:
On Wed, 2006-12-06 at 10:12 +0100, [EMAIL PROTECTED] wrote:
> with tcpdump you have a pcap-format that you can compare bit for bit
with
> the origingal audio, but that is hardly productive.
> rtptools also seems to keep this on a packet level, so we need to find
> a tool to transform pcap to audio or a tool to store audio directly,
ethereal can save the payload but I don't believe it can be done from
the command line.
Ethereal can even replay it (latest version) if it is g711a/u law.
and then to find a nice simple scriptable tool to compare the audio
quality.
I'm pretty sure this tool does not exist, at least not in the open
source world. I looked for something like this a few years back and the
only things available were commercial packages for very high prices.
Let me know if you find anything.
I could think of several options (but none of them readily available I
believe):
- Run automated tests like ITU PESQ (P.862). The principle is that you play
a reference audio file on one end, record it on the other end, and apply the
algorithm to get a score. This score maps to a MOS score, from which you can
judge quality perceived from the end user. PESQ is quite complex and account
for volume difference, temporal synchronization and other things
-> The algorithm is licensed and, if you implement it, you must pay
license fees for psytechnics.
- Play a sequence of in-band DTMF one one side, record the RTP stream on the
other side, decode the audio and decode the DTMF digits, to see if you
recognize the same string that has been sent. This is basic, but helps a
little bit. There is a GPL DTMF recognition algorithm used in Asterisk.
<!-- Play a pre-recorded PCAP file (RTP stream)
-->
> <nop>
> <action>
> <exec play_pcap_audio="pcap/g711a.pcap"/>
> <exec command="tcpdump-vent 9 -s 0 -w
uac_echo_[remote_ip]_6000_`date +%F_%T`.pcap src host [remote_ip] and dst
port 6000"/>
> </action>
> </nop>
I'm confused; why are you playing back and capturing the file in the
same script? I assume this is just proof of concept?
In any case I do see the application of your idea and it looks very
promising.
I only used SIPp for the first time yesterday and already it has helped
me track down a very elusive problem (can you believe we had a client
with one PRI channel that was bad!? I didn't even think that was
possible. I never would have found that without an automated testing
tool like SIPp).
Great :)
--
HP OpenCall Software
http://www.hp.com/go/opencall/
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Sipp-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/sipp-users
