Voice Over IP/Voice-Data Integration

Voice Over IP/Voice-Data Integration

Post by Dennis Hartman » Fri, 15 Jan 1999 04:00:00



    I believe G.711 is the highest quality voice compression you can get
(Cisco's web site has a listing of various codecs, their data rates, and
their MOS scores) .  This represents regular 64k PCM (Pulse Code
Modulation).  G.729 and G.729a, as well as others offer much lower bandwidth
requirements, but at the expense of quality (their quality is acceptable
though).

    What are your goals?  Netmeeting is a very simple implementation of
voice over IP type of stuff.  Are you looking to use your same data
infrastructure for voice (i.e. replace your current expensive, proprietary
PBXs in the future).  While the data networking industry has been growing by
leaps and bounds, the voice industry is still that same closed-off,
proprietary, call Lucent or AT&T for support infrastructure that it was 10
years ago.  Cisco's plans are to dominate this market with voice-data
networking integration.

Cisco's voice primer (a great place to start) document is available at
http://www.veryComputer.com/
_c/vcoview.htm#xtocid82636

-Dennis Hartmann
CCNA/MCSE

P.S.  I have a beta CTI book on my ftp site:
ftp://24.3.131.142/Tech_Books-White_papers/

----- Original Message -----


Sent: Thursday, January 14, 1999 3:13 PM
Subject: Re: Voice Quality


>> we  tried some days ago Microsoft Netmeeting between two ethernet LAN
>>interconnected by an ISDN link ( two B-Channnel in PPP multilink ). We had
>>two Cisco2503. We have tested sharing application and voice, no video. The
>>demo has been successful but the quality of the voice was not as good as I
>>expected.
>> What kind of bandwidth management and queuing do you suggest me for this
>>environments?

>The problem is that NetMeeting uses G.711, which is not as high quality of
>a voice codec as G.729 or G.729a.

>Try replacing the 2503's with 2600's with VoIP cards.  Also set, the IP
>precedence of the voice traffic to something other than 0 so WFQ can
>prioritize it properly.

>Wade
>---------------------------------------------------------------------------
>Wade Williams                      "Put your message in a modem, and throw
>Systems Engineer, CCIE #3373        it in the * sea."
>Cisco Systems, Inc.                       - N. Peart
>Brentwood, TN
>615-221-2918

>---------------------------------------------------------------------------

 
 
 

Voice Over IP/Voice-Data Integration

Post by Paul Ferguso » Fri, 15 Jan 1999 04:00:00



Quote:>    I believe G.711 is the highest quality voice compression you can get
>(Cisco's web site has a listing of various codecs, their data rates, and
>their MOS scores) .  This represents regular 64k PCM (Pulse Code
>Modulation).  G.729 and G.729a, as well as others offer much lower bandwidth
>requirements, but at the expense of quality (their quality is acceptable
>though).

Ah. "Quality" is difficult to evaluate, is pretty much left to human
interpretation. Craig Partridge eloquently calls it the "human-in-the-
loop problem."

For the most part, packetizing voice traffic, as opposed to representing
voice in a TDM environment, is quite a tricky business -- quoting standards
doesn't really buy you anything more than BS assumptions -- delivering
voice traffic in a packetized environment comparable to TDM "quality" is
virtually impossible. Anyone that tells you differently is yanking your chain.

These are diametrically opposed technologies.

Having said that, I assume you had some sort of point? If so, I didn't
get it.

Cheers,

- paul

--
Paul Ferguson                                           ||        ||
Consulting Engineering                                  ||        ||
Internet Architecture                                  ||||      ||||
Herndon, *ia   USA                            ..:||||||:..:||||||:..


 
 
 

Voice Over IP/Voice-Data Integration

Post by Wade William » Sat, 16 Jan 1999 04:00:00



Quote:>    I believe G.711 is the highest quality voice compression you can get
>(Cisco's web site has a listing of various codecs, their data rates, and
>their MOS scores) .  This represents regular 64k PCM (Pulse Code
>Modulation).  G.729 and G.729a, as well as others offer much lower bandwidth
>requirements, but at the expense of quality (their quality is acceptable
>though).

Whoa....a big, embarrasing typo there.  I was thinking G.723 and wrote
G.711 instead.  

From looking on Microsoft's site though, it appears NetMeeting also now
supports G.711.

Wade
---------------------------------------------------------------------------
Wade Williams                      "Put your message in a modem, and throw
Systems Engineer, CCIE #3373        it in the * sea."
Cisco Systems, Inc.                       - N. Peart
Brentwood, TN                        
615-221-2918                            

---------------------------------------------------------------------------

 
 
 

Voice Over IP/Voice-Data Integration

Post by Dennis Hartman » Sat, 16 Jan 1999 04:00:00


  My "point" was merely to correct the statement: (The problem is that
NetMeeting uses G.711, which is not as high quality of
a voice codec as G.729 or G.729a.) that was posted.  Didn't want anybody to
be lead astray by a type-o or a simple mistake.  That was basically.... "My
point"

    It's not anyone's best interest to be "brash" on a list in which the
constituents are trying to collaborate.

Dennis Hartmann
CCNA/MCSE
http://www.veryComputer.com/

----- Original Message -----





Sent: Thursday, January 14, 1999 9:19 PM
Subject: Re: Voice Over IP/Voice-Data Integration


>>    I believe G.711 is the highest quality voice compression you can get
>>(Cisco's web site has a listing of various codecs, their data rates, and
>>their MOS scores) .  This represents regular 64k PCM (Pulse Code
>>Modulation).  G.729 and G.729a, as well as others offer much lower
bandwidth
>>requirements, but at the expense of quality (their quality is acceptable
>>though).

>Ah. "Quality" is difficult to evaluate, is pretty much left to human
>interpretation. Craig Partridge eloquently calls it the "human-in-the-
>loop problem."

>For the most part, packetizing voice traffic, as opposed to representing
>voice in a TDM environment, is quite a tricky business -- quoting standards
>doesn't really buy you anything more than BS assumptions -- delivering
>voice traffic in a packetized environment comparable to TDM "quality" is
>virtually impossible. Anyone that tells you differently is yanking your
chain.

>These are diametrically opposed technologies.

>Having said that, I assume you had some sort of point? If so, I didn't
>get it.

>Cheers,

>- paul

>--
>Paul Ferguson                                           ||        ||
>Consulting Engineering                                  ||        ||
>Internet Architecture                                  ||||      ||||
>Herndon, *ia   USA                            ..:||||||:..:||||||:..


 
 
 

Voice Over IP/Voice-Data Integration

Post by Petri Heleniu » Sat, 16 Jan 1999 04:00:00


Quote:Paul Ferguson writes:

 > For the most part, packetizing voice traffic, as opposed to representing
 > voice in a TDM environment, is quite a tricky business -- quoting standards
 > doesn't really buy you anything more than BS assumptions -- delivering
 > voice traffic in a packetized environment comparable to TDM "quality" is
 > virtually impossible. Anyone that tells you differently is yanking your chain.

Yank!

The above holds true only for toll-bypass applications (and Cisco has
been one-track-minded about that until the late aquisition of
Selsius).

If your packetized voice devices run their voice above the usual 8kHz
8bit sampling, like run at 16kHz sample rates you can do better with
VoIP than you can with the 64kbps TDM channels we've gotten so used to
with. The ear really appreciates the upper frequency tones that are
properly transported.

Pete

 
 
 

Voice Over IP/Voice-Data Integration

Post by Petri Heleniu » Sat, 16 Jan 1999 04:00:00


Quote:Dennis Hartmann writes:

 >     I believe G.711 is the highest quality voice compression you can get
 > (Cisco's web site has a listing of various codecs, their data rates, and
 > their MOS scores) .  This represents regular 64k PCM (Pulse Code
 > Modulation).  G.729 and G.729a, as well as others offer much lower bandwidth
 > requirements, but at the expense of quality (their quality is acceptable
 > though).

G.711 does not compress. It could be described as either
"packetization" or "transport".
 >
 >     What are your goals?  Netmeeting is a very simple implementation of
 > voice over IP type of stuff.  Are you looking to use your same data
 > infrastructure for voice (i.e. replace your current expensive, proprietary
 > PBXs in the future).  While the data networking industry has been growing by
 > leaps and bounds, the voice industry is still that same closed-off,
 > proprietary, call Lucent or AT&T for support infrastructure that it was 10
 > years ago.  Cisco's plans are to dominate this market with voice-data
 > networking integration.
 >
Pete

 > Cisco's voice primer (a great place to start) document is available at
 > http://www.veryComputer.com/
 > _c/vcoview.htm#xtocid82636
 >
 > -Dennis Hartmann
 > CCNA/MCSE
 >
 > P.S.  I have a beta CTI book on my ftp site:
 > ftp://24.3.131.142/Tech_Books-White_papers/
 >

 > ----- Original Message -----


 > Sent: Thursday, January 14, 1999 3:13 PM
 > Subject: Re: Voice Quality


 > >> we  tried some days ago Microsoft Netmeeting between two ethernet LAN
 > >>interconnected by an ISDN link ( two B-Channnel in PPP multilink ). We had
 > >>two Cisco2503. We have tested sharing application and voice, no video. The
 > >>demo has been successful but the quality of the voice was not as good as I
 > >>expected.
 > >> What kind of bandwidth management and queuing do you suggest me for this
 > >>environments?

 > >The problem is that NetMeeting uses G.711, which is not as high quality of
 > >a voice codec as G.729 or G.729a.

 > >Try replacing the 2503's with 2600's with VoIP cards.  Also set, the IP
 > >precedence of the voice traffic to something other than 0 so WFQ can
 > >prioritize it properly.

 > >Wade
 > >---------------------------------------------------------------------------
 > >Wade Williams                      "Put your message in a modem, and throw
 > >Systems Engineer, CCIE #3373        it in the * sea."
 > >Cisco Systems, Inc.                       - N. Peart
 > >Brentwood, TN
 > >615-221-2918

 > >---------------------------------------------------------------------------

 
 
 

Voice Over IP/Voice-Data Integration

Post by Paul Ferguso » Sat, 16 Jan 1999 04:00:00



Quote:>Paul,

>Cisco is marketing MPLS as the solution for quality Voice over IP.  Do you
>disagree?

>Scott

My earlier point was that "quality" is extraordinarily hard to quantify,
especially with regards to toll quality voice over IP.

- paul

 
 
 

Voice Over IP/Voice-Data Integration

Post by Scott Poretsk » Sat, 16 Jan 1999 04:00:00


Paul,

Cisco is marketing MPLS as the solution for quality Voice over IP.  Do you
disagree?

Scott



>>    I believe G.711 is the highest quality voice compression you can get
>>(Cisco's web site has a listing of various codecs, their data rates, and
>>their MOS scores) .  This represents regular 64k PCM (Pulse Code
>>Modulation).  G.729 and G.729a, as well as others offer much lower bandwidth
>>requirements, but at the expense of quality (their quality is acceptable
>>though).

>Ah. "Quality" is difficult to evaluate, is pretty much left to human
>interpretation. Craig Partridge eloquently calls it the "human-in-the-
>loop problem."

>For the most part, packetizing voice traffic, as opposed to representing
>voice in a TDM environment, is quite a tricky business -- quoting standards
>doesn't really buy you anything more than BS assumptions -- delivering
>voice traffic in a packetized environment comparable to TDM "quality" is
>virtually impossible. Anyone that tells you differently is yanking your
chain.

>These are diametrically opposed technologies.

>Having said that, I assume you had some sort of point? If so, I didn't
>get it.

>Cheers,

>- paul

>--
>Paul Ferguson                                           ||        ||
>Consulting Engineering                                  ||        ||
>Internet Architecture                                  ||||      ||||
>Herndon, *ia   USA                            ..:||||||:..:||||||:..


 
 
 

Voice Over IP/Voice-Data Integration

Post by Grant Hartlin » Sat, 16 Jan 1999 04:00:00



> Paul,

> Cisco is marketing MPLS as the solution for quality Voice over IP.  Do you
> disagree?

I think MPLS will be a factor more for QoS guarantees than strict voice
quality.  Certainly one tends to affect the other, but the voice quality
nut has more to do with how packetized voice sounds when transported
over an otherwise quiet network.  The goal of MPLS when it comes to
VoIP, at least as I understand it, is the delivery of that same voice
quality over "non quiet" networks.

As to measurements of voice quality, I agree with Paul that it's
difficult to measure (but see ITU's work with PSQM, et al,
http://www.itu.int/publications/itu-t/itutp10.htm), though I disagree
that the technologies are so opposed that you'll "never" see TDM style
voice quality over packet networks.  In fact, cisco to cisco, we've
already seen it in our labs, both with PSQM scores and with "blindfolded
taste tests."

My biggest problem right now with VoIP state is interoperability.  Cisco
to cisco sounds great; cisco to netmeeting (or any software codec, or
ipj card, or pretty much anything I've tried), sucks.  My understanding
is that Cisco and VocalTec are working on interoperability projects that
may address that down the road.  Perhaps H.323 V2 will help as well.

Grant

 
 
 

1. Is this possible: voice->data->voice->data ...

I want to do this :

        1 phone line with a modem and phone at each end.

main()
{
        Dial a number from my phone.

        while(connected)
        {
                human voice blab,balb balb;

                switch voice to data transfer;

                switch back to voice;
        }

Any modem out there which can do the above.  With a practical peripherals
9600fxmt, you can dial from the telephone and then switch to the modem.
You can't switch back and forth.

The simultenous voice/data modems are way to expensive for me.

Any info would be appreciated.

--thanks
--kim

2. Antic Magazine

3. Logitec TrackMan - 6 pins?

4. voice-overs

5. Developing with eVB

6. Voice and data Integration

7. sql solution

8. Data Voice Video Integration - MC3810

9. Comparative Evaluation of Voice, Data Integration Over LANs

10. Voice and data Integration

11. ** Query: Voice features of data/fax/voice modems?

12. Need help getting voice to work with 56k PCI data/fax/voice