Last changed
20 June 1998 ............... Length about 1,800 words (21,000 bytes).
This is a WWW document maintained by
Steve Draper,
installed at http://www.psy.gla.ac.uk/~steve/resources/VidConfTips.html.
You may copy it.
How to refer to it.
(Up
to my links on video conferencing)
Tips on running a video conference
Written by Margaret Brown and
Steve Draper.
Comments welcome (email
Steve and Margaret).
Here are some "tips" for people running video conferences in MANTCHI.
(This applies to using
the ATM link i.e. high performance link between special video suites.)
These tips originally came from
the tutorial run on 4 Feb 1998, as part of the
MANTCHI project, which was evaluated at both ends; but also from
subsequent experiences.
If you want to try out this new technology for the experience, then of course
it is worth it; and probably worth travelling to do it.
If you believe the point of video conferencing is to save travelling, then it
is fundamentally silly to travel to use another video suite. After all,
video conferencing is inferior in many ways to face to face meetings (e.g. no
social or private business with others "on the side"), and must have a strong
saving in time to be worthwhile. But difficulties in booking can easily lead
to only some sites being available at the time picked. If someone asks you
to travel to a video conference, consider refusing: perhaps they would like
to travel to your office instead?
Telephone conference calls are even more convenient (in your own office), and
may do just as well.
Before the conference:
- Booking the conference call.
It is advisable to check and re-check the booking at both ends.
Considerable difficulties have been encountered booking the Scottish MAN video
conferencing network. The first time we tried,
even though a blank slot was shown on the
booking system so we arranged this time with each other,
we weren't allowed to book it (apparently a secret booking by the
technologists had been made). On another occasion, a planned conference
between two classes was cancelled at the last moment when it turned out the
booking people had just not got back to the organisers, and had in fact
reserved the network for their own use. It is therefore advisable to keep on
phoning, and to contact people at all sites to see whether it is really going
to happen. (In fact, even with a valid booking, on one occasion we only got
connected by phoning up the centre to remind them 10 minutes after it was
supposed to have begun.)
Thus problems with the booking system from a user's viewpoint include:
- Even when all agree a booking has been made, the connection is not always
made without prompting them.
- It is almost impossible to know whether a booking has been accepted.
Confirmations may or may not be sent, may or may not be accurate. The long
chain of people involved makes this very unreliable (e.g. only one end of a
video conference will do the booking on behalf of all; they will go through
their local contact, who will contact Edinburgh.) A failure at any point of
this chain results in people not knowing the state of the booking.
- The
public web record of bookings is not kept up to date and does not
reflect what conferences are booked and what slots are free.
- The notation for sites in that record is not comprehensible by users.
It doesn't use the normal names of the places connected, and doesn't provide
a glossary. Someone has to tell you that "gla-cs1" means University of
Glasgow: yet there are 3 universities in Glasgow, some with more than one VC
suite.
- The description of each booking can be sufficiently meaningless
(e.g. "Also Strath-cs1 smvidconf@gla.ac.uk") that in a recent case a user
couldn't tell whether there was a clash or not: couldn't tell whether two
bookings were related or not.
(Even though for this facility to be used in teaching means it must
be available at class times that were often fixed many months ago, the
impression is that the system is controlled by technologists who put their own
activities ahead of any real applications. Furthermore it is an inherently
difficult issue, since to set up a conference requires exclusive use of
dedicated rooms at both end-sites, plus exclusive use of the wires ("network")
between those sites and Edinburgh, plus use of the switches there.
It is going to be like flying by charter airplane rather than the shuttle to
London: cheap, but very unreliable even after forcing you to arrange
everything around the time dictated by the "service" providers. At best, it
will be like using videotape in a lecture theatre in the days before there
were facilities in every room: inflexible and unreliable. In these days of
teaching quality assessment, it could be regarded as irresponsible to depend
on this in real teaching.
It is therefore unwise to rely on using the MANS video network except as an
optional amusement.)
- Organise a parallel computer link (as an equivalent to an OHP) if talks
rather than discussion are to be presented. The video link will only transmit
one video channel: typically a picture of the present speaker. To give a
talk, the equivalent of an OHP is needed to transmit "slides" to another
monitor in each video suite. Audiences say they quickly get tired of hearing
without seeing the speaker (this was the comment by students on a 10 minute
monologue with slides I gave in one of our sessions), so the main video
channel cannot be used for "slides" successfully. This extra link is not
(yet) provided as standard, but can be done by having a computer with an
internet connection provided in every suite, linked perhaps by Netmeeting.
You are likely to have to organise this equipment yourself: certainly
independently of booking the video conference. You need to:
- Arrange to have the hardware (computers) set up at every site for the
conference. In a big room, you then need to have the computer display
projected on a big screen so everyone at the site can see it.
- Arrange to have them connected to the internet there.
- Decide how to link them e.g. if you use Netmeeting, then all the machines
need to be PCs.
- Decide how to prepare your "slides". Powerpoint is easy, maybe web pages.
Too bad if you wanted to do slides by hand or using a photocopier.
- You will probably need to know the IP address of those machines (or
rather the network ports in the rooms) and to tell the other participants
what they are.
- Agree and publish an agenda (see below)
- Consider introducing yourselves in advance by another medium e.g. email, web pages.
- Use email to prepare everyone for the meeting.
An example is here
- Other preparations for one occasion are listed
here. These might include:
- Every site should have a written list of phone numbers: those of every
other video suite, the Edinburgh switching centre, and the phone extensions
of local technical assistance.
- If a computer internet link is being used in parallel, then each site
should have to hand a written note of their own IP address (to tell other
sites as required).
- Every site needs someone familiar with the video controls: these cannot be
learned simultaneously with having a meaningful conference. If you don't have
an experienced user, then someone needs to practice in advance (see next
section).
The controls are not effortlessly usable. Therefore:
- You need to have a practiced person at each site to operate the controls,
organising their training if necessary.
- If possible the person "chairing" the session should not also be
operating controls. (The real constraint is that the controls take
considerable time and attention, so the person doing it cannot also
contribute a lot verbally or in close management of the dialogue. On the
other hand if their "chairing" is a simple matter of indicating speakers in
turn without having to mediate the content, then chairing and camera
management have been combined successfully.)
Arranging for a practiced user
The controls are not effortlessly usable. Therefore at each site you
- either need a user with previous experience of THAT suite (the
controls are different at every site);
- or you need to arrange a little practice for a designated person.
Groups get restless very quickly when someone is practicing or fumbling while
they wait (another student criticism of one of our cases): after all,
they can't learn anything because it is not their hands on the
controls. So having someone turn up and do it for the first time with a
group causes dissatisfaction and the perception of a bad meeting.
A new user can practice a lot of it without a connection (operating the
cameras and looking at the result on a monitor), but the best thing is to book
the conference 30 mins earlier and have one person at each site turn up then
to practice and to check the arrangments.
During such a setup, you could:
- Ask if they can hear you comfortably; and vice versa
- Ask if they can see you; find camera shots that THEY say suits them.
- Ask them to look you in the eyes (in their monitor) so you know what
direction they are looking when they are looking at you.
The position of the microphones should be taken into account when positioning
the participants. You cannot judge what sound you are transmitting (unless
you have a sound meter). You must ask the other end and believe what they
say. The fact that you can hear OK is, unlike in face to face, no clue at
all about what they can hear.
QA should test the sound coming from speakers in different parts of the room.
The Heriot Watt site seems to have persisting sound problems, with the sound
they transmit being too quiet for others to hear comfortably:
at least there have been problems in 3 out of 3 video conferences I have been
involved in involving that site. This may well be a trivial issue
technically, but it substantially reduces the quality of a conference and
exposes the fact that we don't have the organisational, social, and perhaps
technical mechanisms in place to fix it during a conference.
Having all the chairs facing one way, towards the cameras and monitors seems
to work well.
One issue is giving everyone a good view of the screens (and being in view of
the cameras). Another is that if a group are in a circle, it is easy for them
to feel a group and the person at the far end to feel not part of it.
Effective resolution is bad.
What matters is the size of objects at the user's eye (in, say, centimetres
per radian, or inches per degree). Thus it doesn't directly matter how big
text is at the far end: a lot depends on the display at the receiving end.
To get the most out of a video channel, every user needs to be near enough to
the screen that they can just or almost see the individual pixels or scan
lines. However in many video suites, although the monitors look big, in fact
users are much further away. For instance, sitting at my office computer,
the monitor fills 20-30 degrees of my field of vision, but in the video suite
at Glasgow, it fills perhaps 5 degrees.
Just as in giving a talk at a new place, you cannot be sure how big you need
to make the text on your OHPs, so in video conferencing you cannot be sure
what the display conditions will be at the far end (and you cannot see them
yourself either); but our experience is that this is a concern.
This means, as a rule of thumb:
- Only one face can be recognised at a time. Wider shots show bodies,
but not who they are. The camera should mainly focus on one or two people at
a time and not just a distant view of all the participants. This allows the
remote person or audience to gauge reaction etc. and feel "part" of the whole
activity.
- If you have name plates or hold up printed material, the letters need to
be over 2 inches high (255 point print) in a shot framed to show a person.
- It is useful to have a visualiser available. (You can then, but only
then, use smaller print. Smaller means say 24 point, NOT 12 point.) I.e.
Bring printed "slides": with font as big as OHPs require.
(A "visualiser" is a "rostrum camera" i.e. lights and downward
pointing camera set up to do closeups of bits of paper. Probably looks like
an OHP with a video camera where the projector lens should be.)
All the explicit techniques that help make any meeting and/or tutorial work
well apply, but can turn out to be more important (as some habits don't work).
We can talk about 3 scales of social action in connection with video
conferences:
- The large scale: considering the purpose of the
meeting, and organising the overall joint task. In education, this will be
the level of pedagogical success or failure.
- The medium scale: things you can do in any meeting
to make it go better e.g. start with introductions, begin by agreeing an
agenda.
- The small scale: issues of turn taking, asking the
other end to give you a different camera shot (or not, and being
frustrated).
All the preparation that can help any meeting and/or tutorial apply.
Basically, having a clear idea about the main purpose of the meeting, and
having all participants prepared for it. Thus if it is to be a tutorial, the
students need to have done the work and be prepared to present in some definite
way.
- Agenda. A definite agenda for the conference is useful and should be
agreed and circulated beforehand particularly insofar as it informs
participants about what each needs to prepare, unless it is so simple that no
separate document is needed.
- Alternatively, an electronic agenda (e.g. a web page, a powerpoint
screen) could be made available during the meeting if an extra internet
connection (e.g. using Netmeeting between PCs in every video suite) is being
used. This would have the advantage that it could be edited during the
meeting, yet still be shared by all participants.
- Participants should have access to all the material for the conference
and time to read it before the conference takes place. Materials which are
on the Web can be accessed easily by both sites and shared, discussed etc. --
that is one method, but faxing paper can equally work for small numbers.
- One recipe that works (has worked) is for the student to have
written an essay, the tutor to have read and commented on it, and preferably
to have sent the written comments in advance. Then the discussion can
consist of going through the comments.
- Another is for students doing group work to prepare a short presentation of
their results or what they have done, including electing which student will
represent the group. The tutor can then discuss these presentations.
- But just as in face to face seminars, a general discussion may flop unless
all participants know they will be speaking (and what about)
and prepare some ideas to offer.
All the things that help run any meeting and/or tutorial apply, but are more
important.
- Agenda. A definite agenda for the conference is useful and should ideally
be visible to all participants during the meeting.
If small/short, visibility may not be necessary.
If fixed in advance, then it can be separately pre-printed at each end and
brought on paper.
If agreed interactively at the start of the meeting, then one possibility
would be to put it up on a web page which is separately displayed at either
end (rather than transmitted by the video link) if there is a spare screen
(and computer and browser) at each end.
- Chairing. If a group of participants are present a "chairperson" is useful.
- When using a long powerpoint presentation, many overheads etc.
it would help if the audience at the other site could occasionally see
the lecturer/tutor instead of just hearing him/her. Either organise a second
channel (e.g. Netmeeting over the internet, to give 2 screens of
communication), or have the person in charge of the equipment at the
speaker's end switch regularly between the visualiser (shot of a slide) and a
shot of the speaker.
- Unless the group has already met before, it should begin by going round in
turn with each person introducing themself, including a statement of what they
hope to gain from this meeting.
- In multi-site conferences, it is important to go round each site at
the start so that everyone gets at least a glimpse of the rest of the
audience. Remember that you will then only see one other site at a time.
- Each person should construct a nameplate in front of themselves. Few
people can remember more than 2 names from introductions. However the
lettering must be very large e.g. 144 point (1.5 inches high).
- In multi-site conferences, it is also important to have clear labels
for each site, as the picture will jump between sites, which often look like
anonymous rooms. The best solution is to have a caption inserted
electronically on the outgoing image, as is now done by the University of
Glasgow; otherwise a name plate with enormous lettering.
- A good way to promote discussion, particularly if it is the first
discussion the group has, is to ask each person to say how the topic relates
to a personal experience. This is not only effective in getting discussion
going, it is also in accordance with the best educational theory (relating the
two main levels in
Laurillard's model).
You have to control the camera shots.
And because (see below) this doesn't do everything you want, you have to do
small scale social actions
to compensate e.g. ask the other end to change the
camera shot, nod in an exaggerated way to compensate for low resolution, etc.
- What is wanted, but you can't have, is to control the cameras at the other
end, just as in face to face you turn your eyes and head to see what you want
when you want. This is not offered you currently.
- Because of this, you have to tell them what you want: normal tacit
practices won't work. For instance, if their sound is too quiet it is no good
talking louder. They hear fine, and won't talk louder to suit you,
particularly if they have several people in their room who can hear each other
fine.
- Probably it is best to begin by explicitly asking each other if you can hear well.
- Then, ask them to look at their monitor that shows your face(s): so you
will know what it looks like when they are looking at you. In most setups,
their eyes will not meet yours, but be looking downwards (cameras are often on
top of the screens). You have to ask what "eye contact" will look like.
- The person controlling the shots probably needs to have little else to do,
so they can concentrate on what is wanted and how to operate the controls.
- What the other end will want is both to see the room as a whole, and
the speaker's face and reactions, and what the speaker is pointing to
e.g. a slide on the visualiser. This isn't possible.
- Probably participants should train themselves to give feedback about being
still "there" explicitly. Just as on the phone you have to say "uh uh" more
often than face to face, so you probably need to do this on video conferences
AND have the cameras show all bodies/faces periodically.
- Similarly, probably we should get in the habit of explicitly asking them to
change the camera shot ("show me what the others are doing now").
- Find Computing Services (James Watt building)
A2 on map
- Enter over the stone bridge
- Go straight ahead through the lobby, then through glass doors marked
"staff only".
- First door on right. To get it unlocked, phone Steven Jack ext.4847