Asterisk & HDVoice: Hearing The Siren’s Song Part 2

Audio Sample #1: The Female Voice

The first of my audio samples is Mrs Evelyne Resnick, wife of VUC founder Randy Resnick.  That Mrs Resnick is a native of France is also noteworthy. In many cases the advantages of wideband audio are very evident when dealing with cross-cultural conference calls, where heavy accents and foreign languages can obscure what’s being said.

I had Randy ask Evelyne the following question, “What’s it like to be married to a VoIP geek?” Here’s her response in English.

The video clip is a screencast of the Cool Edit Pro timeline in spectral display mode. It clearly highlights the fact that I’ve intercut three clips to create this comparative version. The Y axis of the display reads from 0 to 22,000 Hz, consistent with the fact that the clip is sampled at 44.1 KHz like a commercial audio CD.

The first portion of the clip is G.711 encoded. This is reflected by the fact that there is no energy above about 3 KHz. The middle section, being G.722 encoded, contains energy to just over 7 KHz. The final section has energy present all the way up to 14 KHz, as we’d expect from Siren14/G.722.1C encoding.

I’m not completely certain that YouTube doesn’t in some way process the audio when I upload such clips. To overcome this potential here is an MP3 encoded version of each sample, along with a screen shot to illustrate the energy distribution in each case. Click on the text label below each screen shot to play the example file.

astricon-2009-evelyne-n-comparative-500

Mrs Evelyne Resnick, English comparative version

astricon-2009-evelyne-n-wave-500

Mrs Evelyne Resnick, English original uncompressed wave file

astricon-2009-evelyne-n-g722_1C-500
Mrs Evelyne Resnick, English G.722.1C encoded

astricon-2009-evelyne-n-g722-500
Mrs Evelyne Resnick, English G.722 encoded

astricon-2009-evelyne-n-g711-500
Mrs Evelyne Resnick, English G.711 encoded

6 thoughts on “Asterisk & HDVoice: Hearing The Siren’s Song Part 2”

  1. Thanks for doing this, I’ve been thinking about something similar for months, but kept putting it on the back burner as it was obviously a low priority. The visuals and video are great for showing customers and potential customers the real advantages of VoIP’s codec flexibility.

  2. Michael, you’ve just set up and instant sales demo for Asterisk/Freeswitch system resellers and for IT folks who are pitching Asterisk or Freeswitch to their CTO’s. Thanks for doing that work and posting it online!

  3. Thanks for highlighting HD voice. I have been using HD voice and FreeSWITCH for the better part of a year and it’s been wonderful! FreeSWITCH comes ready-to-run with both Polycom Siren codecs (7 and 14) and FreeSWITCH runs them in 16kHz and 32kHz beautifully. A polycom phone connected to FS is like magic – nothing to do but enjoy REALLY high quality audio with other HD-supported polycom phones.

    FYI, we do a weekly FS conf call each Friday and we’ve got people calling in from all over the world with various codecs and signaling protocols. SIP, Skype, PSTN, etc. are presented and people using Siren, CELT, Speex, G.711a/u, etc. It’s way cool. BTW, we have an audio clip of crickets chirping that we play when there’s silence. Interestingly, the narrowband codecs don’t carry the audio. You have to be on 16k, 32k, or 48k to hear it. (Yes, we have people calling in on 48kHz CELT. It sounds incredible, assuming that the audio equipment being used is up to snuff.)

    Keep up the good work!
    -MC

    1. It would be interesting to know of any other software that supports CELT. Better yet, any hardware that might support it? I still fear that lack of hardware support dooms open source codecs like SPEEX & CELT to a minor role in industry.

  4. Good point. The only software that I personally know of that says it supports CELT is Ekiga. However, our users report lots of headaches with Ekiga. YMMV.

    We use CELT for FS-to-FS setups mostly. Anthony Minessale and Brian K West connected their Macs together with FreeSWITCH and CELT. Tony fired up his guitar and did a solo and Brian listened to it over the Internet. (Tony was in WI, Brian in OK.) Brian said it sounded awesome. It probably helps that Tony has awesome audio equipment. 😉

    CELT probably won’t be widely used because Polycom his getting Siren out there at a price point that is difficult to ignore. Also, our experience with Speex is that it sucks up a lot of CPU relative to the bandwidth it saves. We have lots of Polycoms out in the wild with the “HDVoice” stamp on them, so I’m guessing that’s the way things are going to go.

    -MC

    1. >>our experience with Speex is that it sucks up a lot of CPU relative to the bandwidth it saves.

      Now FS supports bv32 codec 8)
      X-lite 3.0 /win supports bv32 , speech, dv* too.

Comments are closed.