On 05/17/2013 10:54 AM, Fernando Hernandez wrote:
Thanks a lot to both for your advice, it has been really useful.
Stephen, yes, I am referring to BBB. And I have a question, as far as BBB is not limited by software and we need to run conference larger than 25 we need to scale the hardware. But, I would like to know about the possibility to distribute BBB across multiple servers, how can we do it? do you know if there is any documentation about it?
And about Misheshu comments, I think the licence issue will be a problem. The implementation of G.729 I am talking about is proprietary, not open at all, so we need to find out if it is possible to use it with FreeSWITCH (I am not well versed on the licensing issues, so I will read the licenses and try to understand them). You can check this comparison table to have an idea about this codec performance,
Imágenes integradas 1
And finally, about audio transcoding. We use BBB and speex for the encoding, but as Stephen said before, maybe the audio is not the only problem we have. We need to run more tests...
Thank you very much and best regards,
Fernando Hernandez. What is this table supposed to indicate? u-law, A-law and most forms of ADPCM (the table doesn't indicate which one) are only designed for use at 8k samples/second, and give poor results when used at higher rates. G.729 can *only* be used at 8k samples/second. Its just useless at higher sample rates. If "capacity" is supposed to be the number of channels of the codec which can be supported on an i7-2720QM why is the number for G.729 so low? Even a poor implementation should be faster than 50 channels on a quad core processor like that. The G.729 for Freeswitch is certainly a lot faster than 148 channels on such a processor.
Steve