Historical Notes

Efforts in transmitting audio over the Internet protocols IP and ST-II go back to the early 1970's [ Magill73, RFC741, Cohen76 Cohen77 Cohen78 ].

Danny Cohen and Peter Kirstein:

August 1974
Realtime Packet Voice demonstrated between USC/ISI and MIT/LL, using CVSD and NVP.
December 1974
Realtime Packet Voice demonstrated between CHI and MIT/LL, using LPC and NVP.
January 1976
Live Packet Voice-Conferencing demonstrated between USC/ISI, MIT/LL, CHI, and SRI, using LPC and NVCP.
approximately 1976
First packetised speech over SATNET between Lincoln Labs and both NTA (Norway) and UCL (UK).
April 1977
J. Flanagan (of BTL) applied for a patent on Packet Transmission of Speech.
July 1978
US patent 4,100,377 granted to J. Flanagan.
Voice conference YouTube demonstration
May 1988
US patent 4,748,620, Time stamp and packet virtual sequence numbering for reconstructing information signals from packets granted to Harry W. Adelmann and James D. Tomcik
ITU Recommendation G.764 (Voice packetization - Packetized voice protocols)
February 1991
DARTnet voice experiments.
August 1991
LBL's audio tool vat released for DARTnet use.
March 1992
First IETF MBone audiocast (San Diego, CA)
June 1996
USA patent 5,526,353, granted for System and Method for Communication of Audio Data over a Packet-based Network.

While these experiments used special-purpose hardware and low-bit-rate encodings such as LPC, new experimental interest for transmitting voice over packet networks arose after Sun Microsystems introduced the SPARCstation 1 with a built-in mu-law audio codec.

Versions of software for Sun workstations from the early 1990's include vt and vtalk.

Experiments leading to the MBONE tools and RTP were conducted on the ARPA test network called DARTnet, a transcontinental IP network of about a dozen research sites currently connected by T1 trunks:

Date: Tue 12 Feb 91 00:43:28-PST
From: Stephen Casner 
Subject: Packet voice over DARTnet
Cc: perkins@ee.udel.edu, deschon

I would like to report a successful test with packet voice carried
across DARTnet today.  The experiment was set up by Erik Perkins at
UDel and Annette DeSchon at ISI, using the VT (Voice Terminal) program
recently released by ISI.

Generally the sound quality was quite good and it was easy to carry on
a normal conversation.  Occasional dropouts were heard, but most of
these were apparently caused by effects other than packet loss or
delay.  Unexpected activity (e.g., compilation) on the test host at
the ISI end was correlated with some of the dropouts.  Other dropouts
heard were actually the result of noise from the listening end causing
the "speakerphone function" to momentarily cut off the talker's
microphone to prevent echo.  A third source of dropouts was observed
during a segment of continuous music; the music came through
uninterupted until the dynamic silence threshold eclipsed the music
level just as it is supposed to do to cancel background noise.  We
concluded that it would help to add commands to defeat the silence
threshold and echo suppression functions for test purposes.

Bob Braden told me there was also a test of packet video last week by
the SRI folks connecting to ISI using one of the new Sun boards.
Let's hear more about it!
							-- Steve

To: braden
Cc: Dartnet
Subject: Re: Audio teleconference
Date: Mon, 12 Aug 91 14:49:05 PDT
From: Van Jacobson 


If it would be agreeable to all the participants, I'd like to
try using LBL's new audio tool, vat, instead of vt for the
teleconference.  Vat is functionally equivalent to vt (except
the vat speakerphone mode works better) but has [what I
consider] an easier-to-use, window-based user interface.

The only problem I could see is that vat requires whatever
host you're using for conferencing understand X.  But, since
I think all the hosts we use run either X or Sunview (which
can speak both X and NeWS), this shouldn't be an obstacle.

 - Van

ps- unfortunately, the current version of vat won't interoperate
    with vt so it's an either/or choice of which to use.


See also the Network Bibliography, under packet voice.
D. T. Magill, "Adaptive speech compression for packet communication systems", Conference record of the IEEE National Telecommunications Conference, pp. 29D-1 -- 29D-5, 1973.
"Specifications for the Network Voice Protocol", ISI/RR-75-39, USC/Information Sciences Institute (4676 Admiralty Way, Marina del Rey, CA 90292), March 1976, Available from DTIC (AD #A023506).
"Issues in Transnet Packetized Voice Communication", Proceedings of the Fifth Data Communications Symposium Snowbird, Utah, September 1977 (IEEE Catalog No. 77CH1260-9C), pp. 6-10/13.
"A Protocol for Packet Switching Voice Communication" Proceedings of Computer Network Protocols Symposium February 1978, Liege, Belgium. Also in Computer Networks, Special Issue, Vol.2, No. 4/5, September/October 1978, pp.320-331.
D. Cohen, "Specifications for the Network Voice Protocol NVP", RFC 741, Nov. 1977.
IEEE Journal on Selected Areas in Communications, Special issue on Packet switched voice and data communication, 1(6), Dec. 1983.

Last modified: September 17, 1996 by Henning Schulzrinne