02 Mar 1997 ............... Length about 3200 words (21000 bytes).
This is a WWW version of a document. You may copy it. How to refer to it.

DRAFT

Notes on video conferencing
by Steve Draper


Contents (click to jump to a section)

Introduction

This is an attempt at a short, user-centered review of available video conferencing facilities.

Basis A user-centered review should support the tasks a user needs or wishes to perform. This should be based on gathering requirements from users, both in advance and then as reactions to draft i.e. prototyped versions or the review. Here I have just estimated what may be required by users, based on my own needs and reactions to available information. Unless and until I do more work to gather requirements and reactions to this review, the structure of this review has no firmer justification.

Structure

There are 3 distinct views of the field:
a) The technological one, classifying setups by the technology used and its performance in signal technology terms.
b) A cognitive science one, in terms of theoretical measures of human communication.
c) A user centered one, in terms that matter to users considering whether to use video conferencing, and if so which technology to choose.

Users of video conferencing want to perform communication tasks e.g. they want to chat to friends, or hold a project meeting without having to travel. Hence a useful review should address how well each technology addresses these tasks. On the other hand, facilities are offered under the name of the technology used to implement them, so users have to know those names and something about what they mean. Hence the review is structured by the technologies, but indicates what each means for supporting communication. Some of the technical properties are directly important to users e.g. how easy it is to set up a call, whether the sound quality makes understanding the other person difficult. These too are therefore indicated. The cognitive science measures are harder to relate to direct user concerns. Here, some remarks are recorded about issues with each technology (e.g. difficulty in fluent switches between speaker) that may be explained by issues identified in cognitive science studies.

A major distinction for users is whether the communication is 2-person (i.e. 1:1), or multi-person. We are (nowadays) very used to 1:1 telephone conversations. In fact currently phones support multi-person calls, but few of us have any experience of this, and there is little published about it. We can expect that deciding whose turn it is to speak may be a problem, but whether that goes away with practice or not is not really known. This is also a key issue in video conferencing, but it is not clear yet whether video helps this or not.

Another important issue for users is how much trouble it is to set up a call. 50 years ago long distance phone calls had to be booked, and might take hours to set up. Now that would seem unacceptable, and the spread of mobile phones despite their extra cost suggests that ease of setting up calls is worth a lot of money for both initiator and receiver. On the other hand, face to face meetings are scheduled intensively, and booking a conference call, whether video or audio, is little different from this. It seems clear that this is an important issue, but what exactly its importance is is not clear. Perhaps we need to distinguish three classes of session: spontaneous (as when you bump into somebody, either in real life or on the net), scheduled (whether a physical meeting or a booked video conference), or unplanned but deliberate (as when you phone someone without a prior arrangement to do so).

Primer on the technology background

It is unnecessary to know about most of the technical background. This primer is for those who used to know something about communications technology but have been overtaken by recent developments.

Main alternatives for conferencing

These are arranged in approximate ascending order of expense, which is also largely ascending order of technical quality of the signal. However it is not at all clear that that is ascending order of value for users: for instance, more expense also generally goes with ascending not descending difficulty in arranging a call.

1. Telephone
2. ISDN audio phone
3. ISDN general / video
4. CUseeMe
5. "Desktop video conferencing"
6. Mbone
7. SuperJanet / ATM
8. MAN

1. Telephone

Introduction Traditional and cheap. Now available: multi-party calling. This is actually sensible e.g. for agreeing a meeting time, but we are not in the habit yet. Our university phone system provides it for free: dial the first person, when they answer warn them, press "R" (recall or break), dial the second person, when they answer press "R" again. Can do it multiple times (?). Can do it when you answer a call too.
BT now offers this service domestically, at low cost.
Video standard Audio quality: 3kHz top frequency.
Compression technique / factor None
Delay due to compression None
Point to point or multi-party (user level) Multi-party
Multicast, broadcast, or duplicate point to point (packet level) Effectively multicast
User equipment handset for £10-100, but must have a "recall" button (even for BT service).
Site system equipment Phone line installation and rental e.g. £200 for first year.
Booking needed None.
Other issues We typically aren't practised at managing the turn taking.

2. ISDN 2

Uses same copper pair as ordinary telephone, but special digitial electronics. Gives two logical channels each 64kbps, each for use in computer networks, faxes, or audio. Audio (using only one of the logical channels) is of broadcast quality if between two ISDN phones. Can interdial audio channels with ordinary phone network.

Can do video using 2 channels. The audio is said to be with no delay, just low quality video and small picture size.

Olivetti sell kit for this, (hardware card and software to run on/in a PC). It provide video, audio, plus shared computer space so you can show each other a file etc. But my informant says that the user interface was the main problem: even after a training course, people with a real use need couldn't concentrate on the human interaction because they had to worry about the interface: or perhaps how to show someone something given the delays in transmitting files etc.

Video standard Unknown
Compression technique / factor 128 kbps for ISDN2.
Delay due to compression xxx
Point to point or multi-party (user level)
multi-party
Multicast, broadcast, or duplicate point to point (packet level) PtP or duplicate: depends on what exchange equipment does.
User equipment Rent ISDN line: ordinary phone costs £200 plus calls for first year; ISDN2 costs £2,000 for first year plus calls. Video and compression equipment.
Site system equipment Small box of electronics, on wall or in handset, provided by phone company.
Booking needed No: as for phone.
Other issues xxxx

3. ISDN 6 video conferencing

Introduction Switched phone network, with chunks of 64kbps bandwidth that can be grouped up e.g. use 6 in parallel for video.
Video standard ? looked high standard
Compression technique / factor ?? 384 kbps for ISDN6.
Delay due to compression Yes, seemed about .5 sec.
Point to point or multi-party (user level) PtP mainly, but no doubt multi-party calls also possible
Multicast, broadcast, or duplicate point to point (packet level) PtP, multicast done by digital exchanges?
User equipment Video suite and compression hardware using one of the standards for ISDN
Site system equipment ISDN phone lines (and terminal electronics) installed (3 pairs of wires??): quarterly rental, somewhat more than for ordinary phones.
Booking needed None: just dial up and use. Book the suites/equipment at each end.
Other issues See SuperJanet below for comments on using video suites for conferencing. Plus problem of delay.

4. CUseeMe

Introduction Cheap and free, use on existing PCs, on top of ordinary internet.
Video standard Picture size, frame rate. Slow and small.
Compression technique / factor ??
Delay due to compression Delay, and variable loss
Point to point or multi-party (user level) multi-party
Multicast, broadcast, or duplicate point to point (packet level) PtP i.e. lots of duplication.
User equipment Own PC plus camera/mike; free software.
Site system equipment None. Or install a reflector site?
Booking needed None.
Other issues Low and very variable quality. May load local networks. Connections made by each station sending to a CUseeMe hub/reflector, asking it to send all the traffic it has.

5. "Desktop video conferencing"

Introduction xxxx
Video standard Picture size, frame rate. Slow and small.
Compression technique / factor ??
Delay due to compression xxx
Point to point or multi-party (user level) xx multi-party
Multicast, broadcast, or duplicate point to point (packet level) PtP xx
User equipment xxx
Site system equipment xxx
Booking needed xxxx.
Other issues xxxx

6. Mbone

Introduction Software layered on top of existing internet to allow multicast at the packet level; plus a relatively high video standard.
Video standard ?? Seems to be associated with a standard.
Compression technique / factor ??
Delay due to compression ??
Point to point or multi-party (user level) multi-party
Multicast, broadcast, or duplicate point to point (packet level) multicast
User equipment PC win95, camera etc.
Site system equipment A unix box on same local hub to do routing; mbone software on unix box; mbone software in each user PC; mbone software in some routers.
Booking needed None.
Other issues Can consider separate routes to avoid swamping bits of network: a) separate link from local hub to chief institution hub; b) separate bandwidth slice of MAN

7. SuperJanet

SuperJanet is a fibre optic cabling network between about a dozen UK universities. Top quality video conferencing using ATM switches on top of the network is available, between dedicated rooms with installed special equipment (typically one per site).

Video standard Seemed to be broadcast standard.
Compression technique / factor ?? 2Mbps fixed bandwidth.
Delay due to compression No delay perceptible.
Point to point or multi-party (user level) Multi-party possible
Multicast, broadcast, or duplicate point to point (packet level) Don't know. ATM could in principle to multicast.
User equipment Special room, must be on the fibre optic network. Furniture. Video equipment (~ £7,000). Min. equipment would be one camera and two monitors (for in and out). Typical equipment is 5 cameras including one for documents, 2 big monitors, control panel for switching between cameras, and moving/zooming them.
Site equipment Codecs (video to ATM interface: digitise and compress). ATM switches. (Fibre optic cable network.)
Booking needed a) reserve use of network. The booking reserves cable use, and gets the managers to set the switches at the booked time. b) The rooms are booked. Typically weeks in advance, possibly an hours' notice might work.
Other issues Main limitation seems to be the video equipment. Positioning of monitors relative to cameras is one issue. Another is just picture size: probably huge pictures would be much better e.g. a video projector. The image of a person on the monitor should be of comparable apparent size to the people beside you around the table. Lack of standardisation means you have no idea how big your image is to the other end: it depends on the size and viewing distance of their monitors. User needs a few minutes to practice controlling the cameras.

8. MAN = Metropolitan Area Networks

155Mbps links. About 10Mbps for conventional internet, rest reserved for playing with ATM links. So: optic fibre cables, fancy electronics that divide bandwidth AND do ATM. Make no difference (?) to internet; just possible ATM applications with the additional bandwidth.

The MANs could be used for video conferencing: the cables are there, but each video station costs hugely for ATM switches, codecs, fancy video rooms.

10Mbps is about what any p.c. gets on its connection these days, but is better than most long distance internet links.