Hello, <br><br>I was looking at the old emails on this mailing list concerning problems with AEC when having playback and capture at different sampling clock. Several propositions came along like: <br><br>* Cross correlation<br>
* clock skew estimation<br>* sample interpolation <br>* resampling the signal<br>* ...<br><br>I would like to know whether a working solution has been found as today or are we always bound to use a sound card having both the playback and capture synchronized ? <br>
My application does calls between two computers that are using a usb webcam with a built-in microphone and their speakers are connected to the sound card. Both computers are running under Ubuntu GNU/Linux and I can't get the AEC to work properly. <br>
<br>Like some have asked on this list, I wonder how programs like the gmail voice/video plugin or skype manage to overcome this situation, knowing that a lot of today's users have a webcam with a built-in microphone whose sampling clock may differ from their speakers. <br>
<br>I will appreciate any help. <br>Thanks in advance,<br clear="all"><br>Lorenzo Fundaró García<br>Université de Technologie de Compiègne<br>Génie Informatique <br>Tel. +33 6 21 13 65 48<br><br>