[VoiceOps] G.729 A/B Experiences
peeip989 at gmail.com
Fri Mar 11 19:13:32 EST 2016
It's not as clear but it's still quite acceptable and in low bandwidth scenarios, one could argue, the quality can be better because you don't drop packets where you might with higher bandwidth codecs.
On Mar 11, 2016, at 19:03, Robert Johnson <robert.j at bendtel.com> wrote:
> On 03/11/2016 03:50 PM, Alex Balashov wrote:
> As far as I can tell, G.729 is still the best intersection of low bandwidth and call quality, although the OPUS fans have their own opinion. It certainly leads to intelligible speech, though it can make for some amusing gibberish when applied to hold music, given the extreme code word contractions it uses to achieve its vicious compression ratio.
> However, it's relatively CPU intensive and frequently requires transcoding from G.711 PSTN table stakes. Moreover, in general things are going in the other direction, e.g. higher bandwidth HD codecs.
> This leads me to ask: why, as a North American operator, would you want to do this today, in light of the capacity and price of available bandwidth today? Generally speaking, G.729 is something like a niche interest for international haulers and folk operating in developing world markets where bandwidth remains stubbornly expensive.
> Alex Balashov | Principal | Evariste Systems LLC
> 1447 Peachtree Street NE, Suite 700
> Atlanta, GA 30309
> United States
One of our strategies in combating QoS issues when a customer is
"off-network" is to order a dedicated 1.5/1 ADSL connection and bring it
back to our network on the ILEC's ATM network. But we quickly run out of
call capacity using G.711. Alternatively, we may order a T1, depending
on a number of items (cost, distance, others).
I'm also looking to deploy G.722, but that's another conversation.
Central Oregon's Own Telephone and Internet Service Provider
VoiceOps mailing list
VoiceOps at voiceops.org
More information about the VoiceOps