[comm] Telecommunications video meeting Tuesday Feb 18 at 1030EST/0830MST

[Gareth Hunt]

People,

  [At time of writing, this meeting is now weather-dependent.  According to
  reports, CV has had 8-12" and I have at least 6" at my house.  Side streets
  in town have not yet been plowed.  Tomorrow may therefore be a snow day here.
  GB is usually worse.  Nevertheless, I personally intend to try to make it to
  work if only to escape cabin fever.]

Fred Lo gave Ted Miller the mandate to improve the video support for
observatory wide functions, especially for colloquia and "town hall" meetings.
After a brief discussion with Alan Bridle, Gene Runion, and myself, Ted
solicited a proposal for a formal project to carry this out.

When the project is formally approved, Ted will approve the funds to proceed.
Our first submission is appended.

As you will see, much of the proposed effort is to improve the *sound* systems
in the auditoria.  Alan and Gene have been working on the audio in the CV
auditorium.  I understand from James that he and KScott have been working on
the audio system in the AOC auditorium.  We should consider combining these
efforts and, if appropriate, modifying the document to combine them into a
single proposal.

Cheers,
Gareth.
------- start of forwarded message (RFC 934 encapsulation) -------
From: Gareth Hunt <ghunt at nrao.edu>
To: tmiller at aoc.nrao.edu
Subject: Video proposal
Date: Wed, 12 Feb 2003 19:15:25 -0500

Ted,

Please find appended our proposal for progress on the video communications.
This is perhaps more detailed that you had intended.  However, I believe that
we have outlined most of the issues and provided you with a concrete timeline
and budget, plus the personnel assistance and cooperation that will be needed
from other sites.

Cheers,
Gareth.
===============================================================================

Upgrade and maintenance of infrastructure for observatory-wide video events

G.Hunt, A.Bridle and G.Runion

February 13 2003



1. Introduction

This document is a proposal for coordinated improvements and maintenance to the
NRAO auditoria, conference rooms and network infrastructure to provide better
support for observatory-wide video events such as colloquia, workshops and
"town meetings".


2. Current Status

By January 2003, the NRAO had deployed ten Polycom video conferencing systems.
Three are in Auditoria (GB, CV and AOC), the others are in conference rooms.

Those in the 3rd-floor conference rooms in Charlottesville and Socorro contain
video "hubs" which allow them to host up to four-way video meetings.  An older
stand-alone video hub (NRAO-hub) can also host four-way meetings, but with
inferior audio quality and less efficient use of bandwidth; it is being phased
out and is no longer on maintenance.

Experience gained with these systems in 2002 has allowed (a) routine use of all
of them for video conferencing in small and medium-sized groups, and (b)
exploratory use of the larger rooms for presentations, with up to three other
sites attending.

Meetings such as the five-way December 2002 "town meeting" require the use of
two video hubs simultaneously; the overhead of relocating the hubs into the CV
and AOC Auditoria on these occasions is significant, and must be reduced if
such meetings become more frequent.  Full participation by the VLA site on
these occasions also requires video capability in a large room at the site,
e.g., the cafeteria.  To support such meetings better in future, we have
authorized extending the VLA site's data network to the cafeteria ($1k) so that
video can be viewed there when appropriate.

We have also found (a) that the acoustics in the CV Auditorium interact poorly
with the audio delay in the NRAO-hub when this hub is heavily loaded, and (b)
that the topology of the NRAO intranet does not allow four-way hubbing in
Socorro of events that originate in the GB or CV Auditoria.  We have therefore
addressed these problems and part of the "overhead" issue for 5-way meetings by
upgrading the CV-Aud video system to one containing a modern video hub.  This
new hub was installed on February 10 2003.

As currently configured, the data network infrastructure meets present needs.
SIMULTANEOUS use of both video hubs in Charlottesville for independent
multi-way meetings that include Socorro will however require the CV-SO intranet
connection to have greater guaranteed bandwidth.


3. Deficiencies and plans to remedy them

The biggest remaining deficiency for interactive multi-site video (as opposed
to simple presentations) between the auditoria, is the auditorium sound
systems.  None of these was engineered for the interactive conferencing
situation that should occur during multi-way discussions after colloquia,
within workshops, or in "town meetings".

While "presentation" audio is relatively straightforward, large-group inter-
auditorium discussions require (a) knowledgeable use of the video and enhanced
audio systems by local chairs, (b) personnel at all sites to be tasked with
configuring and monitoring their auditorium sound systems as needed to capture
such discussions, and (c) backchannel communications between the system operators
at all sites during these events, via headset telephones or PCs with instant
messaging.

The major work remaining to be done is (a) to outfit the auditoria in CV, GB and
the AOC with echo-cancelling microphone and mixer systems capable of supporting
observatory-wide video discussions in these large rooms, (b) to train personnel at
all sites in their use, and (c) to ensure that meeting organizers become aware of
the system capabilities and plan their meetings accordingly.

3.1 Audio in auditoria

We have no staff with experience of large-room audio systems, so we need to
become familiar with this technology as it applies to multi-site video
conferencing.  We will have to try solutions and iterate until we find and
adopt one that works well in the presence of the significant audio delays
intrinsic to multi-way video conferencing.  [Both Green Bank and Socorro have
used audio consultants to help them plan modifications to their auditoria, but
these consultants were unfamiliar with the special problems of audio in
multi-way video conferencing systems.  Excellent echo cancellation across a
significant time-delay window is essential for interactive video such as "town
meetings", but not for local sound quality within each auditorium.]  This is a
specialized topic where we believe that we need more hands-on experience before
we can commit to an observatory-wide solution.

We will spend time researching options and looking for knowledgeable
consultants.  We will then acquire equipment and install it in the
Charlottesville auditorium.  Testing will be done while providing video of
colloquia to the other sites, wherein we usually have some minutes of
discussion with a large group in the hosting room and a multi-way hookup to the
other sites.  Experience has shown that you cannot test everything in
almost-empty rooms with just a few participants, or with only one other room
hooked up by video.  It is possible that we will need a second iteration to
find a good solution.

When we are satisfied that we have a suitable configuration in Charlottesville,
we will work with local staff in Green Bank and Socorro to install similar
systems at those sites, allowing for their different acoustic characteristics.

3.2 Audio in conference rooms

At present, the audio is acceptable in most conference rooms.  However, when
there is a full complement of people in the larger conference room, it is
sometimes inadequate because people screen each other from the table
microphones.  The problem appears to be most acute in Socorro, and we will
investigate changes needed to this room to improve the quality of its audio in
smaller inter-site meetings.  Again, iteration and experiment with real meeting
situations is our preferred approach.

3.3 Personnel

We propose the following group to be responsible for achieving these goals:

 Project Manager    Gareth Hunt
 Project Scientist  Alan Bridle
 Project Engineer   Gene Runion

This work will be done as a minor part of their NRAO duties.

We will need ongoing cooperation and time from the engineering staff in Socorro
and Green Bank as appropriate.  We will also need the cooperation of local
business managers with respect to setting up their auditoria.

During testing in live colloquia, we must have a dedicated person in Socorro
and Green Bank available before, during, and after (for post mortem) the video
feeds.


4. Technical Issues

First, it should be noted that the video conferencing system is voice activated
with no provision for studio-style selection of the video feed, which would
require use of significantly more expensive video hubs. Second, centralized
camera operation implies that the operator can only respond to what is actually
received in the video feed from the other sites; they cannot anticipate events
in an unseen room.  To achieve "professional studio" performance in the
observatory-wide events would require both additional equipment and dedicated
camera and audio operators at all sites.  We believe that these are not cost
effective approaches if these events will be relatively infrequent.  A more
practical solution, given better sound systems, will be for the organizers of
these events to plan them in advance so that cueing between the sites can be
done by local chairs who have been made aware of the important characteristics
of the video system in multi-way meetings.  Basically, that to be seen you must
first be heard, and that a little composition time is then required if your
camera is being operated remotely.  A small investment in meeting planning and
rehearsal for these occasions might then be a viable substitute for a large
investment in "studio-like" technology that would not be needed for routine
use.

The NRAO video traffic shares bandwidth with data on the NRAO intranet.  The
intranet was designed to provide the connection topology needed for optimal
internal and external data communication.  The links from Charlottesville to
the other major sites are therefore specified for higher bandwidth use than
other inter-site connections, with a unique path capability on which queuing
priorities can be set.  This restricts the configuration of the video system,
particularly for locations of the video hubs.  While significant upgrades to
the intranet are technically straightforward, they involve increased monthly
recurring cost, so they are strongly constrained by the available budget.


5. Timeline

 2003

 Jan         Acquire FX for CV auditorium
             Fund the upgrade of the link to VLA cafeteria
 Feb         Begin investigation of audio system for CV auditorium
 Apr         Acquire equipment for CV auditorium
 May         Testing
 Jun         Final decision on acceptability
 Jul         (if necessary) 2nd iteration
 Aug
 Sep
 Oct         Acquire and install equipment in Green Bank auditorium
 Nov         Acquire and install equipment in AOC auditorium

We believe that this is an achievable timeline, subject only to the possible
impact of construction activity in and around the Charlottesville Auditorium.
If we make the correct choice for initial tests in April, we should be proceed
at the other sites.  Although final configuration of microphones in
Charlottesville should wait until its Auditorium has been fully refurbished,
this should not prevent us providing basic proof of concept there.


6. Budget
                                           $k
                                           ==
 Expenses already committed (CV & VLA)     12
 Consultant and/or conference attendance    2
 Audio equipment for CV auditorium         10
 Audio equipment for CV auditorium(2)      10
 Audio equipment for GB auditorium         10
 Audio equipment for AOC auditorium        10
 Audio equipment for AOC conference room    5
 Travel                                     5
                                           ==
                                           64

We believe that this is the maximum needed in the worst case scenario.  It
should be noted that the costs of items 4-6 will depend on the outcome of the
testing of item 3.  A more accurate budget estimate should be available at that
time.


7. Further development

Not included in the timeline or budget are ongoing efforts to improve the
quality of the inter-site meetings and extending video to other NRAO
installations:

o   Inter-site distribution of computer screen presentations at higher quality
    than NTSC;
o   Video for the Green Bank Science Center;
o   Upgrade of video systems as they become obsolscent;
o   Video to remote operations sites (VLBA);
o   Video to individual desktops.
------- end -------