comp.compression, comp.dsp, comp.speech, comp.os.ms-windows.apps.winsock.misc
Archive-name: internet-services/voice-faq Last-Modified: 1995/10/22 Version: 0.3 FAQ: How can I use the Internet as a telephone? Version 0.3 - Oct 22 1995 * This is a DRAFT FAQ. This document is new and in transition. If you notice that software for doing speech over the Internet is missing from this list, or information herein needs updating, please send e-mail to email@example.com * This document is copyright 1995 by Kevin M. Savetz and Andrew Sears. All rights reserved. More legal stuff is near the end of this file.
Can I use the Internet as a telephone? What do I need to call others on the Internet? How does it work? What do I need to call others on the Internet? How do I make calls using a modem? Is the sound quality as good as a regular telephone? Is there a noticeable delay in hearing the other user? What is the difference between full duplex and half duplex? What is multicasting? Can I talk to users of other phone software? What software is available? For the Mac: Maven, NetPhone, CU-Seeme, PGPfone For Windows: Speak Freely, CU-Seeme, Internet Phone, Digiphone, Internet Voice Chat, Internet Global Phone, Web Phone For UNIX: Speak Freely, nevot, vat, mtalk, ztalk Legal Stuff Where to Find this Document
Yes. In recent months, several programs have become available that will let you, equipped with the right computer hardware, an Internet connection, and special software, to speak (voice) in real-time over the Internet - in effect using the network like a telephone. Some of the programs available to do voice on the 'net is free, others are commercial. Many of these products are ready for prime time, but others are still very experimental.
It's magic! Audio conferencing programs work by digitizing your speech as you talk and sending the digital data over the Internet. But there's a problem: a typical modem connection has limited bandwidth - 14.4KBPS modems can send and receive a maximum of 1,800 bytes of non-compressible data each second. Telephone quality speech needs 8,000 bytes per second of bandwidth. There are two solutions to the problem: get more bandwidth, or compress the sound information before transmitting it. Although both solutions are used, most programs compress the audio. There are a variety of methods for encoding and compressing sound data, and as you might expect, the standards aren't necessarily very standard, yet. The quality of the audio you'll send and receive depends on the application you're using, the speed of your computer and the compression method used. In my tests, audio is usually understandable, albeit less clear than a phone call. Still, talking across the country or around the world for the cost of an Internet connection is kind of amazing, and easy on the wallet. You and the associate you are talking to need to be using software that can transmit and receive the same protocol. That doesn't necessarily mean you need the same software, or even the same type of computer. For instance, the GSM protocol is pretty standard. However, some programs can only use proprietary encoding methods, so they'll only communicate with the same software on the other side of the wire. Standard compression methods are: GSM: for more information: http://www.cs.tu-berlin.de/~jutta/toast.html CVSD RTP: Real Time Protocol
Besides the software, you will need an Internet connection You'll also need a fairly speedy computer(25 Mhz). Although the exact specs vary from program to program, a Mac Classic or a PC XT generally won't do the trick :-) Your computer will a sound card with speakers and microphone. All sound cards will have a microphone jack in the back of them.
All you need is a version of Winsock and a SLIP/PPP connection. The bare minimum for good quality sound is a 14.4k modem. The quality of a 14.4k connection will depend largely on the software and the error rate of your telephone line. A 28.8k modem should give excellent quality for most all applications.
Surprising to most people, if you have a good Internet connection then the sound quality is comparable to a regular phone call. However the sound quality varies dramatically between applications and depends largely on the local capacity at each end. In nearly all cases, the limiting factor will not be the speed and capacity of the Internet, but will depend on the local work at each end. If you have a good local network connection then some applications give sound quality significantly better than a regular telephone call because they use 16 bit signals rather than 8 bit signals in regular phone lines.
The answer to this also depends on the application and local connection. For most applications, you might notice a delay if you try really hard, but for good connections it seems be in the range of .01-.05 seconds. A delay over .05 seconds might be noticeable on the normal telephone system because you will hear an echo, but calls over the Internet are sent differently do not produce an echo.
In a full duplex conversation, you can speak and hear the other person at the same time. In half duplex, only one person can speak at the same time, and if you speak while the other person speaks, then you will not be able to hear each other. In these applications, you usually press a button to speak Most applications are starting to support both full duplex and half duplex modes. Your sound card will either be full duplex or half duplex, which will control which mode you can operate in. Most new Macs and many Unix stations come equipped with full duplex sound cards, but most PC's do not. When you load the application, if it does not let you choose full-duplex mode, then you probably have a half duplex card.
Multicasting allows one user to send packets to several users for conferencing. The M-Bone is overlay on the Internet that will allow one site to broadcast to many users minimizing bandwidth usage. In addition to being used for voice conferencing, multicasting can be used to make the Internet a broadcast service to allow users to hear live radio broadcast or even see live video of events like Space Shuttle launchings.
Yes. Speak freely offers full compatibility between Windows and Unix users. CU-Seeme offers full
compatibility between Mac and PC users.
Maybe. To talk to others using different software you need to have the same type of connection scheme and the same type of compression. There are two standards emerging for establishing connections that are emerging, VAT and RTP which are compatible with each other. Since VAT was the first working voice conferencing program, many other applications have mimicked its method of establishing a call. Programs using RTP will use the same connection procedure but will have added capabilities to control call quality. Those applications that conform to the RTP/VAT standard include Maven, Netphone, Vat, Nevot and soon Speak Freely. GSM seems to be emerging as the compression standard and is supported by Maven, Netphone, Speak Freely, Internet Global Phone and Nevot. There are also slight variations between the compression algorithms that might still cause compatibility problems. If you are using a commercial product with proprietary compression, then you are out of luck because you will not be able to talk with others because they want everyone else to buy their software.
Maven was the first Internet audio conferencing tool for Macintosh. Maven is free software, but it is a bandwidth hog. Maven requires a minimum of 16KBPS, so even in its lowest quality sampling mode, it requires just a little bit more bandwidth than a 14.4KBPS modem can give you. The program can talk to other Macs running Maven, as well as the UNIX vat program. For those who have the bandwidth capabilities, it probably offers the best sound quality for a Mac. Platform: Macintosh, 16BPS of bandwidth Encoding/compression: vat or Maven-proprietary Features: Full/Half Duplex Negatives: Uses more than 14.4k of bandwidth Mailing list: send e-mail: To: firstname.lastname@example.org Body: subscribe maven your name Software: ftp://sunsite.unc.edu/pub/packages/infosystems/maven ftp://ftp.univie.ac.at/systems/mac/info-mac/comm/tcp Cost: free
NetPhone, like Maven, it gives an Internet-connected Macintosh the ability to do audio conferencing. But NetPhone works over slower connections, even with a 14.4KBPS modem. At a sampling rate of 10,000 samples per second, the sound quality isn't wonderful, but the voice of my NetPhone compatriots are certainly understandable. There's a free demo version of NetPhone available that limits calls to 90 seconds. Platform: Mac LC or faster, 14.4KBPS modem Encoding/compression: vat, CVSD, GSM Features: User Directory, Full/Half duplex, Good Compatibility/VAT compatible WWW Compatible Negatives: often has poor sound quality, demo only last for 90 seconds E-mail: email@example.com Web page: http://www.emagic.com/ Software: http://www.emagic.com/ Cost: $75 per copy, or $125 for two copies. Free demo.
This product comes from the makers of the PGP encryption scheme. It is currently only available in beta format for Macs, but a Windows 95 version will be available soon. The encryption scheme used in the application is banned for US export, and because of current lawsuits against the makers of PGP, users outside the US must use a different version. The current version only supports full duplex sound cards. Encoding: PGP Features: Good encryption Negatives: Only beta version, Only full duplex sound cards supported Web page: http://web.mit.edu/network/pgpfone/ Cost: Free
(see description in Windows section)
Speak Freely is written John Walker, the founder of Autodesk. It is free and one of the best applications available. It offers many features more advanced than commercial software such as voice mail, multicasting, encryption, and usually offers the excellent sound quality. Its GSM compression routine requires a high-end 486 or Pentium processor. Will work for 14.4 modems with GSM, but sound quality is degraded. Current version is not compatible with other software except for Speak Freely for Unix. Version 6.0 will be compatible with the VAT/RTP standard and will include DVI4 and L16 compression and will include a user directory accessible from within the application and on the WWW. The complete source code is also available. Platform: Windows 3.1, 386 Enhanced Mode, Winsock, 14.4KBPS modem Encoding/compression GSM, ADPCM, PCM, PGP, DES, IDEA. Features: Voice Mail, good variety in compression/encoding, excellent sound quality, Full/Half Duplex, multicasting for conferences, WWW compatible Negatives: sound quality is marginal over 14.4k modems E-mail: firstname.lastname@example.org Web Page: http://www.fourmilab.ch/netfone/windows/speak_freely.html Software: Windows (http://www.fourmilab.ch/netfone/windows/speakfb.zip ) and Unix (http://www.fourmilab.ch/netfone/unix/speak_freely-5.3.tar.gz) Voice Chat Room: http://rpcp.mit.edu/~asears/voice.html Cost: Free
CUSeeMe is a project of Cornell University for audio and video conferencing over the Internet. Sound quality is good, but might have problems with a 14.4k connection. Viewing images is impossible at 14.4k and is slow and 28.8k, but audio conferencing will work without images. Users can transmit images using a camera that can be purchased for as little as $99, but a camera is not needed. One main advantages of CU-Seeme for audio conferencing is that users can either connect directly to each other or they can enter a conference at a reflector. In addition, CU-Seeme probably offers the best compatibility with other applications. Features: Good conferencing capabilities, and allows for video conferencing, good sound quality, compatible across Windows/Mac, WWW compatible Negatives: Need at least a 28.8k connection Web Page: http://cu-seeme.cornell.edu/ Listserv: Send message "Subscribe CU-SEEME-L Firstname Lastname" to LISTPROC@cornell.edu Software: http://cu-seeme.cornell.edu/get_cuseeme.html, Windows, Mac & Power Mac
Internet Phone runs on top of Internet Relay Chat, providing you with a list of on-line users and topics of conversation, making it easy to find new people to chat with. Quickly becoming one of the most used applications because the software makes it easy to find others to talk to. The main disadvantage is that that the software is one of the few application that only offer proprietary compression scheme, which makes it impossible for it to be compatible with other applications. The test version only allows for one minute of conversation before it times out. Platform: Windows 3.1, 33 MHz 486 or faster, Winsock 1.1, 14.4KBPS modem Encoding/compression: Internet Phone proprietary Features: Excellent user directory, full/half duplex, low bandwidth demands Negatives: Demo limited to one minute, not compatible with any other software, Cannot make connections without using IRC E-mail: email@example.com Web page: http://www.vocaltec.com/ User directory: http://www.pulver.com/ Software: ftp://ftp.vocaltec.com:/pub (one minute demo) Listserv: Send Message "subscribe iphone" to firstname.lastname@example.org Cost: $99
This is application just recently was made available. Because they do not offer a test version, little is known about how effective it works. According to news releases they had preorders totaling over $1 million. Digiphone can come by itself or in a complete package which includes a Web browser, E-mail, telnet, FTP, voice messaging and caller ID. While it seems that they have invested heavily in advertising, they offer no WWW distribution of their product. Features: Good integration with other Internet applications Negatives: not available on the Internet, not compatible with other software E-mail: email@example.com Web page: http://www.planeteers.com/ Software: None Available on the Internet Compression: Proprietary Cost: $89.95 alone or $149.94 w/ complete package
Internet Global Phone is not a clean, ready-to-use application. In fact, it is very much a work in progress. Internet Global Phone is a "technology demonstration project" providing a code platform for two way real-time voice exchange over the Internet. The details of the project are documented in an article in Dr. Dobb's Journal in December 1994. If you like hacking in Microsoft Visual C++, you too can hack on Internet Global Phone. IGP currently lacks a real user interface, but it really more useful for people to look at the source code in Visual C++. Platform: Windows 3.1, 20 MhZ 386 or faster, 14.4KBPS modem Encoding/compression: GSM E-mail: firstname.lastname@example.org (Sing Li) Software: ftp://ftp.cica.indiana.edu:/win3/demos/IGP* Source code: ftp://ftp.cs.tu-berlin.de:/pub/local/kbs/tubmik/gsm/ddj Cost: free
Internet Voice Chat provides almost - but not quite - real-time communication. Unlike other communications programs, it doesn't send sound as you speak -- it waits until you finish to compress and send the sound. The result is a slower-paced conversation with frequent gaps. (On-the-fly compression is planned for a future version.) Once the shareware is registered, the program includes features line an answering machine and call screening functions. Platform: Windows 3.1, 386 or faster, Trumpet Winsock 2.0B, 14.4KBPS Encoding/compression: Internet Voice Chat proprietary E-mail: email@example.com Newsgroup: news:alt.winsock.ivc Web page: http://futures.wharton.upenn.edu:80/~ahrens26/ivc.html Cost: shareware, $25
Webphone is an unreleased product of Quarterdeck. Their aim will be to integrate Webphone with their Web browser. Not Yet Released Web page: http://www.qdeck.com/qdeck/press/webphone.html Cost: Less than $50
(See Version information above in the Windows Section)
NEVOT (Network Voice Terminal) provides voice communications using unicast, simulated multicast or IP multicast, using the vat or RTP protocols. Requires a high-end workstation. Encoding/compression: vat, RTP, GSM, LPC Platforms: SunOs 4.1, Solaris 2, Irix, HP/UX Features: RTP/VAT compatible, good controls, WWW compatible Web site: http://www.fokus.gmd.de/step/employees/hgs/nevot/nevot.html http://www.cs.tu-berlin.de/~jutta/toast.html Software: ftp://gaia.cs.umass.edu/pub/hgschulz/nevot ftp://ftp.fokus.gmd.de/pub/minos/nevot Cost: free
Vat was the first audio conferencing tool for the Internet. Requires a high-end UNIX machine. The VAT standard for setting up a call is quickly becoming the dominant standard along with RTP, which will allow users of different applications to communicate. Encoding/compression: vat Platforms: Sun Sparcstation, Silicon Graphics and DECstation 5000 Software: ftp://cs.ucl.ac.uk/mice/videoconference/vat/ Cost: free
MTALK is an "very alpha" experimental voice-talk system for LINUX. It uses low bandwidth (approx. 1 KBPS) and claims to work even on computers with very low bandwidth connections. Requires a LINUX workstation with a Soundblaster compatible sound-card. Encoding/compression: unknown Platforms: Linux E-mail: firstname.lastname@example.org Software: ftp://sunsite.unc.edu/pub/Linux/apps/sound/talk/ Source code: ftp://sunsite.unc.edu/pub/Linux/apps/sound/talk/ Cost: free
Simple, "Extremely alpha" voice software for Linux. Includes a voice mail feature. Encoding/compression: GSM Platforms: Linux E-mail: email@example.com Source code: ftp://sunsite.unc.edu/pub/Linux/apps/sound/talk/ Cost: free
This document is copyright 1995 by Kevin M. Savetz and Andrew Sears. All rights reserved. Permission for the following types of distribution is hereby granted, provided that this file is distributed intact, including the above copyright notice: - non-commercial distribution - posting to Internet archives, BBSs and online services - distribution by teachers, librarians and Internet trainers - inclusion on software/FAQ/Internet-oriented CD-ROMs Permission for commercial distribution may be obtained from the editor. SHARE THIS INFORMATION FREELY AND IN GOOD FAITH. DO NOT DISTRIBUTE MODIFIED VERSIONS OF THIS DOCUMENT. This document is new and in transition. If you notice that something important is missing, or information herein needs updating, please contact the editor. The editor and contributors have developed this FAQ as a service to the Internet community. We hope you find it useful. This FAQ is purely a volunteer effort. Although every effort has been made to insure that answers are as accurate as possible, no guarantee is implied or intended. While the editor tries to keep this document current, remember that the Internet and its services are constantly changing, so don't be surprised if you happen across statements which are obsolete. If you do, please send corrections to the editor. Corrections, questions, and comments should be sent to Kevin Savetz at firstname.lastname@example.org or Andrew Sears at email@example.com Please indicate what version of this document to which you are referring.
This file WILL BE BUT ISN'T YET posted twice monthly (on the 5th and 19th of each month) to Usenet newsgroups You can find it on the WorldWideWeb: http://www.northcoast.com/~savetz/voice-faq.html http://rpcp.mit.edu/~sears/voice-faq.html You can receive each new edition of this document automatically via electronic mail, if you are so inclined. This is a low-volume list, with updates every few weeks. To subscribe, send e-mail: To: firstname.lastname@example.org Subject: subscribe voice-faq Body: <ignored> You can also receive it once via electronic mail (without subscribing to automatic updates). To: email@example.com Subject: archive Body: send voice-faq You WILL BE ABLE TO BUT CAN'T YET receive it via anonymous FTP: ftp://rtfm.mit.edu:/pub/usenet/news.answers/internet-services/voice-faq ###end of document###