Want to implement Skype’s SILK super wideband audio codec? You know, the software that makes Skype’s audio calls sound so rich and like you are “right there”?
If so, just check out Appendix A of this massive Internet-Draft:
http://tools.ietf.org/html/draft-vos-silk
Coming in at 566 pages, this I-D is what the word “ginormous” was meant for… but if you look closer, you’ll see that the first 26 pages describe the SILK codec and provide the usual IETF language. The remaining 540 pages are the C source code for a reference implementation of SILK!
This is part of the work of the IETF’s “CODEC Working Group” (see “MOVING INTO THE IETF” below) to standardize on a new wideband audio codec and essentially upgrade the audio quality of the Internet. Let me lay out some background…
LAUNCHING SILK
Last year at eComm 2009, Skype’s Jonathan Christensen announced the SILK codec and that it was being released for free licensing to third-party software and hard developers. Skype made available a developer site focused on the codec:
https://developer.skype.com/silk
Christensen trumpeted these items as the strengths of SILK:
- improving audio bandwidth going from 8 kHz to 12 kHz, meaning that a SILK conversation sounds like you are in the same room as the person you are speaking with
- providing real-time bandwidth scalability to deal with degraded network conditions
- balancing codec optimization between voice, music and background noise, each of which can have an impact on the overall user experience
- delivering a robust solution that delivers a more consistent audio experience, regardless of network conditions and an individual user’s voice signature
You can watch his eComm announcement here:
or view his slides here:
As a heavy user of Skype, I can personally attest to the fact that the SILK codec, available in all the latest versions of Skype, really does provide an outstanding audio experience.
MOVING INTO THE IETF
With the SILK codec available, folks at Skype started to look at how the codec could be made an open standard so that more folks would implement it. In May 2009, Skype’s Jason Fischl started a discussion around forming a new IETF working group to produce a standardized wideband audio codec. Much discussion and several meetings followed that did lead to the creation of the CODEC Working Group within the IETF. The group, created in January 2010 and co-chaired by Peter Saint-Andre and Michael Knappe, is open to the public – anyone can join the mailing list – and the problem statement is very clear:
According to reports from developers of Internet audio applications and
operators of Internet audio services, there are no standardized,
high-quality audio codecs that meet all of the following three
conditions:
1. Are optimized for use in interactive Internet applications.
2. Are published by a recognized standards development organization
(SDO) and therefore subject to clear change control.
3. Can be widely implemented and easily distributed among application
developers, service operators, and end users.
There exist codecs that provide high quality encoding of audio
information, but that are not optimized for the actual conditions of the
Internet; according to reports, this mismatch between design and
deployment has hindered adoption of such codecs in interactive Internet
applications.
There exist codecs that can be widely implemented and easily
distributed, but that are not standardized through any SDO; according to
reports, this lack of standardization and clear change control has
hindered adoption of such codecs in interactive Internet applications.
There exist codecs that are standardized, but that cannot be widely
implemented and easily distributed; according to reports, the presence
of various usage restrictions (e.g., in the form of requirements to pay
royalty fees, obtain a license, enter into a business agreement, or meet
other special conditions imposed by a patent holder) has hindered
adoptions of such codecs in interactive Internet applications.
According to application developers and service operators, an audio
codec that meets all three of these would: (1) enable protocol
designers to more easily specify a mandatory-to-implement codec in
their protocols and thus improve interoperability; (2) enable
developers to more easily easily build innovative, interactive
applications for the Internet; (3) enable service operators to more
easily deploy affordable, high-quality audio services on the Internet;
and (4) enable end users of Internet applications and services to enjoy
an improved user experience.
Now, here is a critical point:
The CODEC Working Group is *NOT* just about Skype’s SILK codec!
The CODEC WG is working to identify and standardize a wideband audio codec that could be widely used over the Internet. Skype’s SILK codec is one of the potential codec candidates, as shown on this matrix of codecs from the CODEC WG. Another strong contender is the CELT codec from the Xiph.org team. There is a Internet-Draft out now for discussion that specifies the requirements for a standardized codec:
http://tools.ietf.org/html/draft-ietf-codec-requirements
The goal of the CODEC WG is to either identify an existing codec that meets those requirements or create a new codec. Either way, the goal is to have a standardized high quality audio codec that we can all use.
For those interested, the CODEC working group will be holding its first official face-to-face meeting at IETF 77 in Ananheim, California, on March 22nd. A draft agenda has been posted.
SKYPE’S SILK SOURCE CODE
So back on Monday, March 8, a new version of the SILK Internet-Draft was published and announced on the codec mailing list. As the draft notes:
The source code for the reference implementation of the SILK codec is
provided in Appendix A. This fixed-point source code is the
normative specification of the codec. The corresponding text
description in this document is provided for informative purposes.
For more information about SILK, the patent licensing terms, and to
download the source code in a convenient format, please visit
[silk-website]. For any additional questions, please send an email
to SILKSupport@skype.net.
This particular phrase is interesting and amusing to me:
This fixed-point source code is the
normative specification of the codec.
I translate that as “the source code is the specification”.
If you do read the source code, you see Skype’s license on the code is rather open:
Copyright (c) 2006-2010, Skype Limited. All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, (subject to the limitations in the disclaimer below)
are permitted provided that the following conditions are met:
- Redistributions of source code must retain the above copyright notice,
this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
- Neither the name of Skype Limited, nor the names of specific
contributors, may be used to endorse or promote products derived from
this software without specific prior written permission.
NO EXPRESS OR IMPLIED LICENSES TO ANY PARTY'S PATENT RIGHTS ARE GRANTED
BY THIS LICENSE.THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND
CONTRIBUTORS ''AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING,
BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF
USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
Essentially it is a Skype version of the permissive “Simplified BSD License“, which I would summarize as a non-lawyer as essentially “keep the copyrights intact, don’t use Skype’s name for endorsement, and Skype is in no way responsible if bad things happen from using this code”.
Still to be seen is if there are any additional licensing terms from Skype, but overall this is a great step forward for those of us who want to see better overall audio quality on the Internet. Obviously it’s in Skype’s interest to do this so that the CODEC WG might consider SILK as the codec to standardize on versus the others or creating a new one… but still… it’s great to see. What do you all think?
Last week at VoiceCon Orlando 2010 I had a great time giving a 3-hour tutorial on the Session Initiation Protocol (SIP). Over 200 people attended and made it a very interactive session with lots of great questions and comments. The abstract for this session was:
The W3C Voice Browser Working Group recently released the fifth working draft of the VoiceXML 3.0 specification at:
Want to learn about the Session Initiation Protocol (SIP)? Would you like to understand how the SIP protocol works and why it is the dominant open standard for communication today? Want to understand the challenges SIP faces and what’s being done to overcome them?
RSS Feed
