Update --
Due to heavy opposition by members of IMC-Video and the US Newsreal, combined with a general collapse of Indymedia's tech infrastructure, this project is unlikely to happen. The information below is still relevant as a completely doable project for a non-Indymedia group.
Overview
Video segments will be submitted to an Indymedia
server
and
combined into finished half hour programs
consisting of a
bumper, 5-9
segments
with a brief
transition clip between each segment,
and
credits at the end. The video segments will be
submitted in a high bitrate
encoded form in PAL or
NTSC and the finished program can be exported in either PAL or NTSC format.
Submitted segments must be transcribed in the program's native language, a
volunteer team will
translate the segment into
other languages. Several distributors may
distribute
physical tape or VCDs for their reigions, and local IMCs can provide
assistance to help video producers encode and submit their segments.
Lastly, this document mentions the
funding and
volunteers it needs to work.
This project grew out of a
proposal made for the US
Newsreal in Spring '03.
A server (currently berkman.indymedia.org) will host the website for the
Global Newsreal, which will include accepting submittions, displaying
program status, allowing specific users to manage segments (such as removing
duplicates and segments deemed unsuitable for the program), compiling the
finished program, and being the primary source for distribution.
Each program may be up to 256megabytes in size, which means that utilising all
of berkman's bandwidth (roughly 100megabit/sec, or 720megabytes/minute) we can
send over 4000 copies a day. We of course can't max out the bandwidth for a
day, the server is used by other projects (such as liveradio.indymedia.org) so
we probobally want to cap ourselves at a quarter to half of the server's max.
We will use the
Iceshare and
BitTorrent P2P systems to greatly reduce this load. Every party receiving a copy of a show will be expected, prehaps required if they want to download in the first 24-74 hour "high demand" window, to send that same volume to atleast one other user. Many will be available to send to more than one, thus increasing our available bandwidth several-fold.
Working entirely within the Ogg framework gives us many benefits, such as
being able to chain logical bitstreams into a single piece of media without
transcoding (decoding and then re-encoding into a lossy format, thus producing
greater loss). Each logical bitstream carries it's own encoding values and
metadata so we don't have to worry about chaining a NTSC clip to a PAL clip,
the player will see the change and handle it properly.
Segment producers will upload their work to
the server
and can immediatly see it included in the next program. They're told that
their segment may be removed if it doesn't meet some editorial criteria, but
in general segments are included "first come first serve". If the current
program is full it'll get added to the following, and producers can opt to
float their segments to any upcomming month that's lacking segments. Many
such segments can be submitted by frequent segment providers to encourage new
producers to submit work.
When a segment is submitted it's checked for sanity, bitrate, metadata,
supported codecs, and to ensure that there's atleast one audio, video,
and subtitle codec included. Once a segment has been added it gets
listed on the translators section of the website. The native language
of the segment is provided with text entry fields for translators to fill
in. Translation work is added to the segment as it's completed, the
status pages will show which segments lack translation to which languages.
Segments can be submitted up to 72 hours prior to release, at which time
the program's timeline will be filled with "floating" segments. Any
pending translation and editorial work (removing segments) must be finished
by 24 hours prior to release.
A 24 hour window is reserved to distribute the finished program to our
distributors. This may include mirror servers, tape/vcd duplicators,
and cable networks (such as FSTV). This method allows these groups to
download a copy while the
server bandwidth is not
burdened by local groups attempting to each obtain a copy for screening
or syndication on
local public access. It also
gives us some breathing room in case some translation is missing, etc.
A well designed bumper is of course needed. This will be on the server
in an encoded form and will be chained at the beginning of each program.
Please add your ideas on what this could look like here.
Video clips of any size will be accepted by the system, but only segments
which are 3 to 5 minutes in length will be automatically added to the
next program.
Segments under 3 minutes, or "shorts", may include brief informational
videos on an upcomming events or political commentary, cartoons, etc. We can
suggest that they incorporate the text "Global Newsreal" into the segment
so that it can be a transition clip, but our own transition clip can be used
before or after the short. This can be a good project for newly trained
video producers who may become frustrated with a longer segment at first,
plus seeing their work get distributed to such a wide audience can inspire
them to continue working on Indymedia video projects.
Segments over 5 minutes, or "longs", may include extended documentaries and
collaborative projects around big events or topics. We may place a cap at
15 minutes, so that a single segment doesn't consume more than half a program,
and longs may only be used in programs that don't have enough segments at the
time of processing. Many longs will likely be submitted by unexperienced
video producers who may be contacted by a member of the editorial team with
suggestions on how it can be cut down to be under 5 minutes.
To make the finished program look more polished transition clips including
the (((i))) logo and "Global Newsreal" can be used between segments. There
can be one transition clip that's used repeatedly or we can encourage many
people to create their own that will be used randomly.
Shorts may include their own themed transition clips, prehaps at the bottom
of the frame or a video effect used (ie, crossfade) to the "official" logo
by itself. Other segments may also begin with their own themed transition
clip, for example, an "Earth Day" segment could have the logo with trees
and plants growing into it and it's theme music fades into the segment's
audio of outdoor sounds. Allowing flexibility like this creates space for
a large amount of creativity on the part of segment producers.
Rolling credits at the end of each program may include the names of the
editorial team, segment producers, and information on how to get involved.
Video segments must be submitted in a patent-free Ogg codec. This currently
includes Theora for video, Writ for subtitles/captions, and the producer's
choice of Speex Vorbis or Flac for the audio layer. This limitation is
included because of the strong desire to keep the software which processes
the Newsreal programs entirely free of control by corporate/proprietary
software and available for techs to constantly develop and improve.
Using the Ogg framework also gives us the flexibility we need for some of
the
processing we'll be using to produce each
program. This is important because repeated transcoding of video becomes
very noticable after only a few generations, and most video recording
devices (ie, DV, miniDV, D8) and editing systems (Final Cut Pro, Premier,
Cinelerra) transcode atleast once prior to encoding for our purposes.
The producer does have a large amount of flexibility over the video format.
We'll be able to accept PAL and NTSC, any perspective (16:9 widescreen?),
any resolution, any framerate, any colorspace, stereo or mono, any samplerate,
or any other eccentric encoding choice the codecs permit for. Ogg players,
which will all utilise the common
OggFile? library, will play the file as one
file but see them internally as seperate pieces of media, thus while sending
fullscreen to tape it'll handle different encoder values seamlessly. Also,
since Ogg is normally VBR (variable bitrate), each producer can choose a wide
range of bitrates depending on their ability to process them, their bandwidth
for submitting them to our
server, and how picky about
audio/video quality they may be.
The producer is responsible for transcribing their own segment in it's
native language prior to submittion, other languages may be included at
the producer's choice but are not required. There will be free tools
available to do this soon, but we'll see if a web-based alternative is
needed to lower the technical barrier for newer video producers.
A translation team will have access to view the native language transcript as
text, without needing to view the video (or even listen to the audio, tho
this is available), and can translate the native language to another language
on the website. They will need no extra tools to do this, just access to a
web browser and the time to translate the video phrase for phrase.
We will need to decide which languages to provide subtitles for in order to
be reliably available for distribution in reigions that speak those chosen
languages. This is limited only by the availability of translators to do
the job, and it'd be good to have multiple translators for each language to
ensure sustainability and prevent burn-out (translation may not be easy!).
Our primary source of distribution will be the
archive server using the
Iceshare P2P system.
Many IMCs and community groups that are involved with video work already
have the ability to download and process video, including screening it
directly from a laptop or exporting it to VHS tape for a screening event.
However, not everyone has the equipment or knowledge, or even just doesn't
have the time to go this route. Many IMCs also lack the bandwidth to
download high quality video, including many in the global south which we're
especially interested in getting involvement from.
One way to handle tape distribution is to leave it to reigional autonomous
reigional volunteers/groups to dub to tape and handle the physical
distribution, and they are left to set their own fees for this service.
Tape distributors will, however, be given access to obtain a copy before
the general public and will be listed on the main website (see
processing).
Leaving tape distribution to independent reigional distributors, vs including
it in the scope of this project, also gives the local distributors flexibility
in the form of the distribution. It may make sense, for instance, for some
reigions to use VCD instead of VHS tape, or prehaps some other tape format is
popular in certain reigions.
Tape distributors will be able to choose which language they want to use,
and to print the subtitles for their language right in the video frame for
segments who's native language not in their chosen language.
Submittion of physical tape can be handled through a local IMC, a
reigional tape distributor, or any other local entity. A local IMC
may instead provide the equipment and training for the producers to do
so themselves, or any number of other arrangements.
Many cable companies limit programming which is not produced by a group
or resident in their local area. This limits the amount of syndication
we can do through public access of the "official program".
Traditionally, local video editors for many groups have worked around these
limits by taking clips of syndicated work, editing them, giving credit where
it's due, but putting a local name and local info in the program. This is
of course time consuming and due to the extra stages in the editing chain
produces lower quality results than the origional.
To make this task (much) easier, facilities for generating a localised
Newsreal will be provided. A local editor can produce their own bumper,
transition clips, throw in a few shorts announcing local events, and even
choose to replace some of the segments with locally produced segments.
This function will allow local IMCs to produce their own localised Newsreal
suitable for screening on their local public access, or even just screening
at an event, and have enough editorial control over each program to be able
to state that they are the producer for each program - all on the website.
The project will rely entirely on patent-free media, free software, and
server resources which are already available. Any monetary resources we
receive (ie, through syndication) can be put into further development of
our server resources, getting video equipment to "global south" IMCs,
and advertising (full color posters, etc).
Unlike the current Newsreal projects (US, NYC, and European) this system
does not require a central source to mail tapes to, mixdown, encode,
etc. These functions are instead provided by software on an IMC server.
Editorial control over segments will be open to many people, from
different parts of the globe, who will be able to preview and discuss
segments on a mailing list and make decidions collaborativly.
This responsibility will not be in the hands of one individual.
Translation will be made easier through the web-based interface, the
team will not need to each obtain a copy of the video or have specific
software, only use a web-based form and spend roughly an hour (as I've
been told) to translate each half hour program for their language.
Most of the effort needed for this project is the non-recurring development
work by Audio/Video techs to write the software and design the interfaces.
Overall, it'll require far less effort than any current Newsreal project
plus having what editorial workload we do have spread amoung more people.
--
ArcNLN - 11 Sep 2003
to top
