If you can't get 1 call to sound good you clearly have an environmental problem to solve.
I am not sure where you will ever be advised from us to use ubuntu but I guess anything's possible,

We currently endorse centos5 for older hardware, debian 6 for anything recent from the last 3-4 years and are still vetting centos6


There are not really enough details from this post to diagnose your problem.
I suggest you answer every one of these questions.......

What hardware are you using?  (based on the willingness to try so many distros I fear virtualized which is a whole separate can of worms).
What device is the other end of the call?  Is it a provider or a IP phone or what?
What format are you recording to?
Do the calls sound unsatisfactory in real time or only the recordings?
Does this same problem only happen in your special conference setup or even on ordinary bridged calls?


In general, digital audio records verbatim to the stream that is being received.  If you are using g711and recording pcm wav, there is little reason for quality to be changed from the encoding.


You may want to setup another box on the same network and test calling the milliwatt extension or some other tone and work your way back from there.




On Fri, Nov 8, 2013 at 5:28 PM, Scott Ross <slross16@gmail.com> wrote:
I've developed a tool which leverages freeswitch event_socket to originate 2 calls (sofia/external/number@ip) whom then navigate an IVR, connect to the same conference bridge then take turns playing back a .wav while the partner call records channel input. The goal is to run the two recorded files against a PESQ library to monitor whether our conference bridges are performing as expected. 

Problem: I'm hearing poor channel quality (static, degradation) from the recordings and when I listen in on the calls while the test is running. My PESQ scores are coming out much lower than expected. Out of curiosity I tried asterisk and scores were about a point higher (2.9 vs 3.9) so I should be able to rule out the hardware/environment/destination lines I'm running it on. 

I need help improving freeswitch audio quality for my use-case. I know this is asked alot - I've read all the threads/wiki articles I can find and tried: loading only necessary modules, running a single codec (ulaw), I've tried ensuring my playback reference is also ulaw format without headers, I've tried running on several different OSs: Ubuntu 11.04, RHEL 6.8, RHEL 5.6. Even originating on two separate sofia profiles (although I don't think this should have mattered regardless). 

I prefer freeswitch over asterisk in all other aspects but need to solve this (important) quality issue.  
Let me know what outputs I can provide.

 

_________________________________________________________________________
Professional FreeSWITCH Consulting Services:
consulting@freeswitch.org
http://www.freeswitchsolutions.com

FreeSWITCH-powered IP PBX: The CudaTel Communication Server
http://www.cudatel.com

Official FreeSWITCH Sites
http://www.freeswitch.org
http://wiki.freeswitch.org
http://www.cluecon.com

FreeSWITCH-users mailing list
FreeSWITCH-users@lists.freeswitch.org
http://lists.freeswitch.org/mailman/listinfo/freeswitch-users
UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users
http://www.freeswitch.org




--
Anthony Minessale II

FreeSWITCH http://www.freeswitch.org/
ClueCon http://www.cluecon.com/
Twitter: http://twitter.com/FreeSWITCH_wire

AIM: anthm
MSN:anthony_minessale@hotmail.com
GTALK/JABBER/PAYPAL:anthony.minessale@gmail.com
IRC: irc.freenode.net #freeswitch

FreeSWITCH Developer Conference
sip:888@conference.freeswitch.org
googletalk:conf+888@conference.freeswitch.org
pstn:+19193869900