Slashdot Mirror
Beyond The Cell -- Journalists' Video Phone
dimitri_k writes: "This article from poynter.org gives some information about the video phone that has become standard in reporting recently. It uses H.263 for compression, and a satellite phone to call into ISDN lines. Maybe people on Slashdot can brainstorm ways to increase the bandwidth of these things in the short term (i.e. cost-ineffective combination of lines) so that the cable news networks can turn the grainy, live, night-vision shots in Afghanistan clear." This setup looks a little chunky, but when you consider the capability to beam video information from anywhere in the world, it's very impressive.
I hope someone will think of using this to broadcast pr0n videos live from remote places, like "antarctica upside down" or "easter island statue-hard III" or something.
"A door is what a dog is perpetually on the wrong side of" - Ogden Nash
This made the lack of resolution less apparent. Scaling the image up to fill the screen produces a very pixelated image. Also it seemed that the low framerate was less noticable this way. It wasn't nearly as annoying as the video phone footage that I've seen in the past.
Perhaps if they don't want to transmit in real-time and can afford a minute or two of delay they could record some footage at a higher resolution and/or framerate and then send it to the network and have them assemble it at the network. It might take 3 minutes to transmit 1 minute of footage this way. You lose the realtime aspect of the current setup but you could get better quality.
Lasers Controlled Games!
Well, first off, cut out the full duplex operation. Send voice only out to the field, and use the extra bandwidth for more frames. The reporter on the other end rarely needs to see what's happening in the home office, while the whole world would appreciate a clearer picture.
They're using H.263 compression algorithms... some dismal figures (it was made to be used at 10 fps, for instance!) Here's a nice page detailing the standard and some comparisons to MPEGs...
Here's a great page comparing H.263 to MPEG-4... Hmmm... Jurassic Park encoded in High Quality MPEG-4 beat the 64 Kbit/s rate of H.263 by nearly %20... the video phones are, according to the article, 112Kbit/s... Anyone have any clue about using MPEG-4 to do this? Sounds to me like it'd be a much better compression algorithm...
I am disrespectful to dirt! Can you see that I am serious?!
The guy who mentioned DivX and MPEG-4 above wasn't far off at all. The problem here is definitely the compression technique. The one they're using is utterly ANCIENT by today's standards, which can produce better framerates and image quality with lower bandwidth. Anyone remember the similar i.263 codec that used to be used in AVI videos traded over the Net? No? Now you know why nobody uses it anymore.
;-)
I too have been seeing those video phones in use, on the Fox News Network. But I had no idea ancient software was to blame, I just thought it was all the bandwith's fault. But they're not using that bandwidth to its full potential. They need to use an MPEG-4 based codec instead. Make their own, or use Microsoft's little AVI-based implementation, or anything--just use a modern compression technique.
I'd also imagine they could improve quality substantially by interpolating any lost frames, back up to the NTSC standard or a flat 30FPS. Surely a big news conglomerate can afford the hardware and software to do that relatively simple, though horsepower-intensive-in-realtime, chore.
Cheap bastards.
Chasing Amy
(We all chase Amy...)
"The more corrupt the state, the more numerous the laws"-Tacitus
turn the grainy, live, night-vision shots in Afghanistan clear.
Even if they had more bandwidth, it wouldn't help that much. The low bandwidth causes blockiness. The graininess and the poor color comes from the fact that the cameras just don't work well in low light.
Now, a while ago I saw something on the Discovery channel where a guy had a low-light camera that he was using to capture the aurora borealis in real time. They could use something like that.
Of course, I could go on about how there isn't really any need for us to see explosions at night in full technicolor, but that's beside the point.
For all intensive purposes, "whom" is no longer a word. That begs the question, "who cares"?
The night vision they are using is probably where the grainy-ness first comes in. It's not to say that the News Corps. arround the world cannot afford some highend night vision equipment, it's that the US and NATO will not allow anything above a certain level to be exported to non-NATO approved country, such as Afganistan. They want to keep the nice equipment out of the terrorist hands. Ever look for Night vision online? A lot of dealers will say "cannot be exported outside the US" for this exact reason. They are selling everything from Gen 1 to what some are calling Gen 3+, but only Gen 1 and maybe some Gen 2 can cross the boarder.
I have an old Ukranian Gen 1 scope that looks similar to the footage you see on TV, but when I use my newer Gen 3 scope from ITT, it's like daylight. Hell, I've even used it to read stuff in the dark, and navagate boats with it. Gen 1 scope... uggg.... New boat anchor. Gen 3 scope... I'm hunt'n wabits... on the other side of the lake... at 3 AM... with no moon light.
I aplaud the idea of enhansing the video, but realise, when the daytime footage come through, it's fine, night vision feed from an exportable scope, looks like crap.
you may try to point out the military's footage looks just as bad, but you think they are going to let the enemy know they can spot an untied shoelace at a mile and a half?
Who wants Pork Chops?
they should hire ventriloquists as reporters: ventriloquists can talk without moving their lips and this will save a ton of compressed bandwidth!
I can tell most of the folks commenting on this thread have not used high bandwidth sat phones or done much live video (or both).
The InMarSat system is a geostationary constellation, and requires a pretty decent amount of power to transmit.
It requires a directional antenna, which is part of the reason the phones are as large as they are. The smallest are the size of a small briefcase, and these videophones are not much larger than that.
You can mux together multiple dishes to get 64k, 128k, 192k, 256k, etc, but each 64k requires another dish, another power supply, and more space.
Yes, the codecs are less than perfect, but they are standard, and allow you to connect virtually anywhere in real-time.
We've experimented with live encoding into more efficient formats and quite frankly you don't get much better quality, and the lack of built-in videoconferencing smarts on the part of the codecs costs as much as you gain in efficiency.
Yes, if you can record, encode and transmit in near-real time the quality could be better, but then you're talking about a much more technically complicated setup that a reporter with limited resources has to manage.
Operating a computer in your office is much simpler than doing it on a frozen rock with bombs falling nearby and a poor power supply. If you have a connection, you transmit because you never know when it may go down or your power will die. Getting a few extra FPS for extra time sounds nice in theory, but getting the story out ASAP is more important because 30 seconds from now things could change.
The videophones are an amazing package, and little can be done to improve them much more than the simple march of technology. They'll get smaller, we'll get better sat systems with more bandwidth, the codecs will improve, but for what resources exists now, these things do an AMAZING job of wringing out all the performance possible.
Recursive: Adj. See Recursive.
First off, I watched one of the news reports via videophone and I was quite impressed by the audio clarity and the video quality. M$'s NetMeeting can't even compare at the same data rate.v ideo-streaming.html
MPEG4 is an outgrowth of H.263.
The reason H.263 is chosen over MPEG4 and other similar streaming codecs is because the latency from video capture to transmission of the encoded image is better under H.263. During some informal testing, latency of H.263 video conferencing on a LAN was well under 2 seconds. The best I could do with Real's RealProducer using their G2 codec was around 4-5 seconds. The best I could do with Microsoft's Media Encoder with the MPEG4 codec was around 7-10 seconds.
Because of the way that MPEG2 and MPEG4 take advantage of the time domain to achieve higher compression also makes them unsuitable for 'live' 2-way video.
Here are some links to chew on:
http://myhome.hananet.net/~soonjp/vclinux.html
http://archive.dstc.edu.au/RDU/staff/jane-hunter/
http://mpeg.telecomitalialab.com/
The H.263 spec is available at http://www.itu.org for a fee.