performingarts - RE: ConferenceXP update
Subject: Performing Arts Advisory
List archive
- From: "Feghali, Jose" <>
- To: "Michael C. Wellings" <>
- Cc: <>, <>
- Subject: RE: ConferenceXP update
- Date: Sun, 9 Mar 2008 10:25:47 -0500
Ooops… I should have said that our
tests with Juilliard were with DV25 and HDV video inputs - not uncompressed
video. Apologies for the error. From: Feghali, Jose
[mailto:] I should clarify a couple of things about
my thoughts and experiences on this issue. Of course I believe that “pushing
the envelope” with uncompressed video is a very important endeavor. For people
working in computing sciences and research, the state-of-the-art must surely be
one of the most important priorities. I am always excited to see reports about
the envelope being pushed farther, and have the hope that one day when
bandwidth considerations become less of an obstacle (and less costly) that we
will be able to use that technology. Considering however the difficulty that we
have encountered even finding DVTS partners, and after talking to a lot of
people in the music area of the performing arts, it became clear that one of
the most important features missing with normal “off the shelf”
video conferencing equipment like Polycom (which a lot of them are using) and
other non-DVTS solutions was the lack of uncompressed audio. Obviously, for music,
this is a much more important priority than perfect video. ConferenceXP allows
for multiple users, which also opens the door for multiple speakers/teachers
from different locations to participate in the same session, like some flavors
of videoconferencing and AGrid. We can for instance have a violinist, a cellist
and a pianist in different places coaching a piano trio over here, or a
composer from one location and a conductor from another helping out with a
contemporary work’s rehearsal. If this was to be done with uncompressed
video, then of course the bandwidth requirement would be virtually impossible
to handle. I think that ours is somewhat of a
“niche” requirement compared to most commercial and other needs, in
that we value better sound than perfect video – with the exception of the
Opera area, where the best quality video would also be quite important. Most
companies and developers appear to emphasize audio codecs which work well for
voice, echo cancellation and so forth (understandable, considering that this is
what the majority of their market needs). Although these are important
considerations, they can lead to not provide the best sound possible. It seems
that only fairly recently has “transparent”,
full-frequency-spectrum audio cancellation been possible for uncompressed
audio. Even an institution like the Juilliard
School of Music does not currently have access to Internet2, and during our
tests 2 days ago we were able to do some uncompressed video and audio tests
with them only because they had a spare OC line which hadn’t been
disconnected yet. Previous tests, even at 1 or 2 Mbps, where a disaster. They
are working towards having Internet2 access, but meanwhile the best they can do
is 2 “bonded” T1 lines with a max useable bandwidth of around 3
Mbps. If anyone cares to chime in with their
experiences, specially if different from ours, I for one would very much like
to hear your input. Best, JF From: Feghali, Jose Agreed. Unfortunately, most performing
arts institutions do not have anywhere near 250 Mbps available (most still run
100 Mbps Ethernet connections in their building) and such a connection here,
for instance, would take up half of our available bandwidth with the Gigapop
providing our Internet2 connection – assuming we replaced our current 100
Mbps routers in the music building with Gbit ones. What ConfXP does now that no other
solution appears to do (save maybe Vic and Rat, which are very complicated for
non-technical people to use) is actually to allow for LESS video bandwidth, not
more, but including uncompressed audio – thus filling a gap which exists
at the moment between DVTS and normal video conferencing. Until institutions
have updated their network connections and acquired Internet2 or similar
connections with high available bandwidth, a lot of them here in the US but
specially abroad will continue to be “bandwidth-challenged” and
cannot work with DVTS, let alone dream about SD or uncompressed HD.
ConferenceXP opens up the possibilities of video conferencing now with CD and
higher quality audio. For most of what we do in music
(masterclasses, lectures, demonstrations, etc…) the 350 ms or so of
latency does not really matter, and the “doubly-compressed” video
(DV + WMV) is also not a concern. I am pleased however that CXP appears to have
reduced the latency of the “straight” DV25 stream to about half of
what DVTS displays. José From: Michael C.
Wellings [mailto:] The DV CoDec (http://www.nssn.org/search/DetailResults.aspx?docid=280521&selnode=)
does frame-based compression at 25mbps. Uncompressed SD video varies
between 100 and 250 mbps depending upon the sample format. Current
standard broadcast/production data rate is 250mbps. The point of
transporting uncompressed video (SD in this case) is to lower the latency.
DV compression adds 100 or so ms per end. For arts-related use lower
latency would be pretty important. The point is to get ConferenceXP to
transport video at uncompressed rates…indeed at a suite of rates...to
make it more scalable and useful for multiple applications. I really appreciate
your continued interest in ConferenceXP and ip video…it’s really
important for us to understand the needs of the community and to engage others
in continued development mike From: Feghali, Jose
[mailto:] What I am saying is that ConfXP appears to
do what DVTS does – i.e. send the firewire DV stream from the camera out
via IP uncompressed by a codec on the computer. I know that the DV codec has
“theoretically” got some compression on it, but I got the same data
rates for both DVTS and ConfXP. I have not dealt with any *really* uncompressed video format and do
not intend to experiment with uncompressed HD or other “cutting
edge” technologies anytime soon. From: Michael C.
Wellings [mailto:] That data rate is too
low to be uncompressed SD video. What is your input format? Must be
compressed… From: Feghali, Jose
[mailto:] About the same as DVTS – almost 30
Mbps with video and audio (around 25 Mbps just video) using DV. From: Michael C.
Wellings [mailto:] What is your output
data rate? From: Feghali, Jose
[mailto:] Under video settings, you can choose
“uncompressed video” in the advanced box. It switches off the
Windows Media encoding engine. From: Michael C.
Wellings [mailto:] I understand...fred
mentioned you. What hasn’t been there is support for multiple
codecs, and for uncompressed HD video. We are looking at the ConferenceXP
front end for iHDTV. ConferenceXP
natively supports Windows Media encoding…what support for uncompressed
video are your referring to? From: Feghali, Jose
[mailto:] Michael, I have been in touch with Uncompressed video has always worked, and
HDV also worked yesterday. In order to get DV and HDV to work properly with
ConfXP, one needs to “downgrade” to Windows Media 10, as there
appears to be a problem under Windows Media 11 which doesn’t allow for
compressed video in CXP. José From: Michael C.
Wellings [mailto:] Jose Along with our CS
Department we got the grant from Microsoft Research to continue development
work on ConferenceXP. In a mtg the other day we discussed this prospect
as well as plugging live HDV into ConferenceXP, and opening it up to other
CoDecs and – Uncompressed video. If there is interest in the
community to do this is may help with the decision process mike From: Feghali, Jose
[mailto:] I tried RAT a while back and
couldn’t get it to work right uncompressed. Do you know what is the max resolution
that VIC supports? Is there a FAQ that you know of that explains how to setup
RAT and VIC for IP-to-IP conferencing? Thanks, José From: Petr Holub
[mailto:] RAT from MBone Tools
(or AccessGrid) can do also
uncompressed audio at this level. Petr ================================================================
Petr Holub CESNET z.s.p.o.
Supercomputing Center Brno Zikova 4
Institute of Compt. Science 162 00 Praha 6, CZ
Masaryk University Czech Republic
Botanicka 68a, 60200 Brno, CZ e-mail:
phone: +420-549493944
fax: +420-541212747
e-mail: From:
Feghali, Jose [mailto:] Colleagues, We have know successfully tested ConferenceXP with
uncompressed stereo audio (16/44.1 and 16/48) using both compressed and uncompressed
video. The latency with all uncompressed appears to be lower than DVTS, but
there is some question as to the video quality. One of the testing sites
reported the video being slightly less sharp with ConfXP using an HDV camera.
It appears that an optimum bandwidth with compressed video/uncompressed audio
is in the order of around 2.5 to 3.5 Mbps, but we have gotten good results down
to around 2 Mbps. If there are any other performing arts institutions or
others that would like to test ConfXP in full uncompressed audio with or
without compressed video, please get in touch with me off-list at (j DOT
Feghali AT tcu DOT edu). If anyone here knows of any other conferencing solution
currently available that allows for full, 44.1/48 KHz uncompressed audio to be
paired with compressed video I would appreciate the input. José Feghali Artist-in-Residence Coordinator of Internet Technologies |
- ConferenceXP update, Feghali, Jose, 03/08/2008
- Message not available
- RE: ConferenceXP update, Feghali, Jose, 03/08/2008
- Message not available
- RE: ConferenceXP update, Feghali, Jose, 03/08/2008
- Message not available
- RE: ConferenceXP update, Feghali, Jose, 03/08/2008
- Message not available
- RE: ConferenceXP update, Feghali, Jose, 03/08/2008
- Message not available
- RE: ConferenceXP update, Feghali, Jose, 03/08/2008
- Message not available
- RE: ConferenceXP update, Feghali, Jose, 03/08/2008
- Message not available
- RE: ConferenceXP update, Feghali, Jose, 03/30/2008
- RE: ConferenceXP update, Feghali, Jose, 03/08/2008
- Message not available
- RE: ConferenceXP update, Feghali, Jose, 03/08/2008
- Message not available
- RE: ConferenceXP update, Feghali, Jose, 03/08/2008
- Message not available
- RE: ConferenceXP update, Feghali, Jose, 03/08/2008
- <Possible follow-up(s)>
- FW: ConferenceXP update, Feghali, Jose, 03/09/2008
- RE: ConferenceXP update, Feghali, Jose, 03/09/2008
Archive powered by MHonArc 2.6.16.