To: Craig K who wrote (16467 ) 6/30/1998 11:22:00 PM From: TheLineMan Respond to of 50264
Ok....time to speak....I work in the area of communications...I deal with voice compression and encoding to some degree, although not VO/IP... IP is just the packaging of the data for internet routing...what remains is the voice sampling usually at an 8KHz rate for a good frequency response from 300Hz to 4KHz... so far so good, but you neglected to mention the samples are taken from a A/D converter (usually 16 bits) - giving you a raw data rate of 128kbps The output of this sampling is a high rate data stream about 12bits wide...way to much data to packetize and send over the net... you're beginning to show your ignorance - what do you mean by 12 bits wide???? The key to minimizing bandwidth utilization is the algorithms that are used..this is where the argument exists..what algorithms are being used....Standard or proprietary....???? The problems that sometimes exist with standard algorithms is that they are not optimal meaning that the output would require more bandwidth per call...I don't know that this is the case here, but with the OVERIDING significance of bandwidth if DGIV has a proprietary algorithm that requires less bandwidth than the standard, it may be correct to use it.... They appear to use a proprietary algorithm - I know nothing about - if it is better than the current algorithms for quality and bandwidth - why is it not being adapted as a standard by the ITU? Interoperability has often been the key, what will happen to DGIV's products as ASND, CSCO, ERCY, LU start selling their interoperable system worldwide. DGIV had better start addressing the Qos issues The other part of the story is; Even if it is proprietary,, it can be converted at the gateway to a standard format....if this conversion is done at the gateway, bandwith is not an issue.So you're saying that they can re-encode already compressed data to a standard ITU format - at another gateway. That's clever - not to mention the degradation of voice quality after the voice has been sent through two encoders. You have to realize that we would not have an incompatibility at the IP packaging and routing layer, just the voice encoding format... Are you saying it's fine that the IP packet gets to its destination - even if the voice data is garbage to the decoder? Another question that may be significant is....Does Digitcom' business plan require any interopability with any other vendors products.....My guess would be no....Digitcom plans to have gateways a both ends of the POP...they will encode and decode their on voice and switch it into analog lines...no compatibility with anything isrequired....from what I see, this is Digitcoms path even if they do have a standard algorithm.So how will users place calls to areas with no DGIV gateways? Will they route these calls to a hub of DGIV servers and use voice standard voice circuits? Most companies in the VoIP arena are way ahead on DGIV on this one.