Skype Announces Support for Opus Audio Codec

skype-opus-logoAs discussed in the “Skype’s Core Applications” section of Experience Skype to the Max, one of Skype’s unique technologies has been its royalty-free SILK audio codec that provides crystal clear audio on Skype-to-Skype voice and video calls. However, Skype also submitted SILK to the IETF to request its addition to IETF standards as one element of Internet infrastructure that allows software developers, hardware vendors and service providers to work towards common interoperability standards for voice and video communications. The podcast in this post provides an example of the clarity that comes when using Skype for iPhone over a 3G or WiFi wireless connection compared to the narrowband audio heard on a standard phone call on a wireless carrier’s voice channel.

Earlier this week, the IETF announced the adoption of the royalty-free Opus audio codec as an IETF standard after three to four years of debate amongst major participants in the audio communications space, including Google, Mozilla Corporation (Firefox) and Nokia. Wednesday, in Skype and a New Audio Codec, Karlheinz Wurm, Audio/Video Product Engineering Director for Skype, announced Skype’s intention to adopt this codec in future product releases. Quoting from his post:

This is the kind of goal that drives our engineers at Skype. We hope that the fact that Opus is now a standard will mean that it will be more widely adopted, helping us in our mission to deliver better quality as part of improving communications for the Skype community no matter the platform or device….

… We believe that Opus is the first codec with state-of-the-art performance for any type of audio signal and any application (communications, streaming and storage) under any condition.

Because Opus was designed for the Internet, it can adjust seamlessly on-the-fly between any of its operating modes to adjust to variations in available internet resources, whether moving from 3G to WiFi or competing with the house next door for broadband bandwidth.

Fundamentally Opus is a fusion of Skype’s SILK codec technology for voice with Xiph.Org’s CELT codec for music. It includes the infrastructure to adapt the use of this codec as the voice channel during Skype or other publishers’ voice and video calls to underlying network conditions. It represents a key component to the overall availability of enhanced HD voice as IP-based communications take over the telecom infrastructure.


I had the opportunity to interview Koen Vos, Senior Skype Architect and Distinguished Engineer, and ask some questions about Opus and how it will benefit end users. Amongst the topics we discussed were:

What are the major end user benefits of using Opus?

Not only improved audio quality but also better adoption to the resources available on the web. Skype users will also benefit from stereo audio and “full band” – essentially providing a codec that supports audio across the entire audio frequency range of the human ear. (Note, while voice covers up to about 14 KHz, music produced by instruments actually can be heard out to 22kHz.)

Translated to end user terms is that there is more opportunity, amongst others, for those who use Skype for “remote” music production and training. Of course it also means more robust support of the crystal clear voice conversations currently supported by SILK.

We went on to talk about:

  • adapting to network conditions during a call,
  • adapting to the resources on the CPU of an endpoint device or client that supports Opus (where PC CPU’s have more horsepower than, say, mobile smartphones and tablets),
  • deployment of Opus on end point devices as well as supporting Opus features across various gateways and other telecom interconnections,
  • inclusion of Opus in Skype clients, whether on PC’s or mobile devices,
  • the role of Opus in the evolution of the webRTC specification (which is intended to become a standard for making voice and video calls within a web browser)
  • what benefits can be derived from having stereo available in Skype calls
    • calls from board rooms
    • calls involving Skype for TV where people are sitting across, say, a family rooms
    • reduction of background noise
  • the impact of Opus on network bandwidth demands
  • how the three modes of Opus operation (SILK only, hybrid, CELT only) adapt to the actual call setup and conditions
  • the impact of Opus on any time delay in a voice conversation
  • Opus and working in conjunction with video codecs such as H.264, VP7 and VP8
  • The role of Opus in the Microsoft/Skype CU-RC-Web proposal to the W3C WebRTC working group.

Bottom line: Opus provides an opportunity through an industry recognized standard that addresses several issues related to Internet voice communications: transmitting the entire audio hearing range of the human ear; adapting a conversation to the underlying network conditions and providing the hardware and telecom interconnections required to allow end-to-end crystal clear conversations and audio in any call. Its actual deployment will take a few years to implement but at least there is now a common reference point. Obviously Skype has the opportunity to become one of the first application developers to implement Opus.

What follows are two videos of my interview with Koen Vos; they provide more detailed answers to the topics mentioned. (Warning: due to network conditions and Koen’s connection via an Amsterdam hotel’s Internet service, Koen’s video froze during the recording of the second video; however, he makes some very informative comments in the audio channel.)

Part I covers the discussion down to “inclusion of Opus in Skype clients”; Part II starts with the discussion of the role of Opus in the webRTC specification.

Related articles

Enhanced by Zemanta

About Jim Courtney

Bringing over thirty years' experience in the sales, marketing and management of cutting edge technology businesses.

, , , , , , , , , , , , , , , , , , , , ,

No comments yet.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.