MediaLib has EchoSuppression in the flowgraph, and I could find no
references to AEC. This project has brought me back into C++ after
many years, so I certainly am no expert and I may have missed it. I am
usuing Visual Studio and it may be in a file not referenced in the
solution. EchoSuppression gave me no sound, and I found that this
matches the behavior described in the issue tracker on XMR-43. I
didn't care for suppression, and moved on. The functions for
controlling AEC from sipXtapi were not connected to anything, but my
patch fixed this. MpMisc does have an echoQ that MprToSpkr populates,
but I didn't find a consumer, until I used it.
Do you have any quick reference information on the new MpBuff
interface, or do I need to just install Doxygen? In moving back to
sipXtapi branch I dumped all my changes that weren't critical for the
project. They should merge in well, but I expect these changes are
better suited in the media branch.
As for inconsitant timers, has anyone looked at the code contributed by
Kenichi Aramaki? I must admit that I haven't had the chance yet.
Charlie
Alexander Chemeris wrote:
Hello,
On 8/21/06, Charlie Hedlin <[EMAIL PROTECTED]> wrote:
I did on my local machine, but I am now going to move my aplication to
the sipXtapi branch so I can get major bug fixes without loosing
stability. I was using the media branch so I could implement the echo
canceler, but we don't need it any more (our headset adapters implement
it). I hope to work on these projects on my own time, but as we get
closer to a useragent that meets our needs I am expected to move on with
company time.
We really appreciate time you're spending on this project. You're working
with really important and complicated places.
BTW, I have a talk with Dan Petrie about echo cancelation and he said that
AEC, included in mediaLib is a giood one. Did you test it?
BTW, did you look into dedicated libraries for adaptive Jitter Buffer
implementation?
I noticed that the speex library has one, and have considered using it,
but I haven't had the time to research it. If it weren't for the
garbled audio I wouldn't worry about it, we are using this on a LAN
where there shouldn't be much jitter. While I believe mpJitterBuffer
was causing problems on the boundry of either an empty or full audio
buffer (probably full, do to missed timer events and the Media thread
falling behind. It wasn't the spkr queue, as the distorted audio was
recorded in the 'ToSpkr' recording object.) I don't feel like I have
much of a problem with jitter itself.
I think that inconsistancy of timer events is the root of many audio related
problems here. Under Windows frame begin signals are generated by
WOM_DONE events in speaker thread. Due to debug output this events
often get delayed and then bursted out.
Regards,
Alexander Chemeris.
_______________________________________________
sipxtapi-dev mailing list
[email protected]
List Archive: http://list.sipfoundry.org/archive/sipxtapi-dev/
|
_______________________________________________
sipxtapi-dev mailing list
[email protected]
List Archive: http://list.sipfoundry.org/archive/sipxtapi-dev/
