Having read & listened this far into this series you should now have some grasp of how narrowband (G.711) compares to wideband (G.722/G.722.1) and even super-wideband (G.722.1C) audio for telephony applications. The differences in many cases are quite pronounced, even startling. What you hear in the examples are just the most obvious properties of the encoding, sampling rate and by implication, the available audio bandwidth. It’s worth understanding a bit more about the evolution of the role of the codec over time. This will help you frame up how the Siren codecs fit into the Asterisk realm.
In this third installment I’ll try to broaden your experience with wideband and super-wideband telephony by exposing you to a selection of recorded audio samples using various encoding techniques.
Until now the examples used were strictly in English. This next set of six samples recordings are in six different languages; Norwegian, Chinese, French, German, Russian & Spanish. Each is presented in a comparative form, with three codecs intercut into one example recording. Then again in each of the following; uncompressed, super-wideband (G.722.1C), wideband (G.722/G.722.1) and finally narrowband (G.711) a la PSTN.
In order to truly appreciate the difference between the various recordings you will need to be making use of high-quality audio playback hardware. Good quality computer speakers or, better yet, a high-quality headset will be the most revealing. But then, as someone who’s genuinely concerned about the quality of audio over IP telephony…you knew that, right? I thought so.
In part 1 I gave you an introduction to Polycom’s Siren7 & 14 codecs, as well as a brief overview of their implementation in Asterisk v1.6. Now it makes some sense to try and understand their advantages in use. This is really a more generalized exploration of narrowband (G.711 ala PSTN) vs wideband (G.722/G.722.1) vs Super-Wideband (G.722.1C)
I set about creating a series of audio recordings to illustrate the difference between the three codecs. If Asterisk had been capable of handling all three codecs then recording samples encoded in each fashion would have been relatively simple. The trouble is that in the period leading up to Astricon I didn’t yet have a version of Asterisk capable of handling Siren streams beyond pass through.
Preface: This post is a rework of the HDVoice session I presented in cooperation with Polycom at Astricon 2009. The Powerpoint slides in support of that session as well as a videotape recording of the session are anticipated in a few weeks on the Astricon web site.
In considering this subject I developed more demo material than was possible to use in the 40 minute session at Astricon. This post begins a series that is a kind of superset of the Astricon session, intended to go into more depth with a larger variety of HDVoice examples.
The introduction to the session was given by Tim Yankee, Director of Product Marketing, Voice Communications at Polycom. Tim’s intro gave an overview of the state of HDVoice in the industry. Hopefully his slide set will be included in the presentation materials to be put online at Astricon.net
This coming week I’ll be doing a short presentation at Astricon 2009 in Phoenix. For some reason the description of the session has never made it onto the event web site so I thought I’d describe it here in case anyone was interested.
The topic is “HDVoice & Asterisk: Hearing The Siren’s Song.” The session is essentially an overview of the very recent implementation of the Polycom Siren7 & Siren 14 codecs in Asterisk v1.6. The session is part of the “Tech Track” and the conference and happens Wednesday, October14th at 11:40am.
I’ll be sharing the stage with Tim Yankee, Director of Product Marketing, Voice Communications for Polycom. Tim will start the session, presenting on the state of HDVoice as envisioned by Polycom. When Tim passes the mic to me I’ll offer a demo of the Siren codecs.
We hope to make it both informative and entertaining. And, oh yes….there will be a test…of sorts!
Michael Stanford of Wirevolution has an article called Better Sounding Calls in the March issue of Internet Telephony that was today published on TMCs HDVoice Community site. While very general it’s nevertheless a nice article. He cites Speex developer Jean-Marc Valin referencing the fact that wideband is the principle means of VoIP surpassing the PSTN in terms of end-user call quality.
He notes that transcoding between wideband codecs, or worse wideband and narrowband, is generally a bad idea. He further makes an assertion based upon Polycom’s release of the Siren7 and Siren14 codecs under a royalty free licensing scheme;
There are now three high quality wideband voice codecs that phone vendors can use without paying royalties: Speex and two from Polycom. There is no reason why any phone or soft phone should ship without all three of them.
I whole-heartedly agree, and further assert that Skype’s SILK should be thrown into that mix. Of course, G.722 is royalty free as well, although not nearly as sophisticated as the others mentioned.
It’s also interesting to note that Speex adoption in hardware remains extremely limited. I wonder if that might change as wideband continues to gain momentum? Or does it get left behind in the face of newer royalty free, if not open source, codec offerings? The open source community has also moved on to offer CELT, which is a very new but extremely low-latency, very flexible wideband codec.