A couple of handset manufacturers I recently talked to mentioned that they were looking at implementing AMR-WB into their IP phones. Everyone currently looking at AMR-WB, be it for a mobile device, softphone implementation, or network core, needs to skip it. EVS is the future and it has four key advantages over AMR-WB that make it worth the effort to embrace sooner rather than later.
EVS, short for "Enhanced Voice Services," comes out of a 3GPP working group for Voice over LTE (News - Alert) (VoLTE). EVS is designed to be a drop-in for VoLTE and existing AMR-WB use, with backwards compatibility with both the AMR-WB wideband mode and the narrowband AMR mode.
Advantage one here is self-evident: EVS provides AMR-WB and AMR modes, plus a bunch of upside for doing the full range of the human voice and high-quality music. AMR-WB delivers a frequency (sound) range from around 300 Hz to 7000 Hz, while EVS goes from 50 Hz to over 14,000 Hz. EVS is the clear winner when it comes to a wider frequency range.
EVS also delivers AMR-WB quality (baseline cellular HD voice) in less bandwidth than the current AMR-WB codec, the second advantage. Depending on the carrier configuration, AMR-WB needs anywhere from 12 kbps to nearly 24 Kbps, while EVS will deliver the same sound quality in roughly 5.95 kbps to 7.36 kbps.
Carriers will end up doing system trades between existing HD voice sound – packing more phone calls in existing bandwidth – or delivering high quality calls at the same or better bandwidth. EVS's superwideband mode starts at 13.2 kbps. T-Mobile (News - Alert) US says it uses a full 24 kbps for HD voice, so it could crank up superwideband mode AND save bandwidth, depending on how it decided to configure the codec on network and handsets; knowing T-Mobile, it will probably go for as high quality of an experience as possible.
EVS, so I've been told by Fraunhofer (News - Alert) representatives, can also be used as a straight-up drop in replacement for AMR-WB on existing 2G and 3G HD voice deployments, so you don't even need VoLTE to put it in operation.
Advantage three is improved error resilience for both legacy circuit switched and (more importantly, in my view) VoIP applications. If there's delay, dropped packets, and/or other issues with bandwidth, EVS can compensate for some problem without taking a quality hit.
Finally, there may be fewer firms to deal with for licensing use of EVS. Maybe. Securing the licensing rights to use AMR-WB currently requires at least four (4) different arrangements with different patent holders. EVS may require only two or three parties involved, with the best case being licensing agreements with Qualcomm (News - Alert) and a separate one through an EVS patent pool arrangement. Qualcomm apparently doesn't want to pool its licenses under any circumstances, even though it would be the most convenient for third-party developers.
It would be in everyone's best interest to have "one stop shopping" for EVS licensing, but I'm not optimistic. Hopefully it will be better than the headaches that AMR-WB licensees have gone through. And people wonder why Opus is so popular.