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.
Might be the killer application to finally get MPEG4 or OpenDivx codecs into hardware. This alone could probably get higher bitrate, higher quality video transmitted over the lines.
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
It seems that alot of criticism is being directed at the choppy video feeds. There will always be a trade-off of quality and compression that is limited by bandwidth. I really don't see the bandwidth problem being solved in the near future. But, who says that these feeds really need to be in real-time? Yes, there are certain instances where having a real-time feed is useful, but it would also be good if they could capture some high quality feeds then "squirt" them to the receving stations. It wouldn't be instantaneous, but you could get a better quality feed.
Where the wind blows, the tumbleweed goes.
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!
This might be brute force, but how about add the capability to transfer data over two or more phone lines simultaneously, in parallel, if they exist at a location?
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
...broadband sat uplinks require a big, bulky satellite rig by comparison and can be a liability if you have to move in a hurry.
I am not merely a "consumer" or a "taxpayer". I am a Citizen of the State of Texas
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"?
On many large yachts, they have small satellite dishes inside fibreglass domes which automatically rotate to track the satellite as the boat moves along, so you can watch satellite TV whilst moving, and also access satellites for comms and phone. These only cost a few hundred dollars, so it can't be that hard to put one of these on a satellite videophone...
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?
Tell me about it, i think it is too fill in the spaces when they have nothing to say. Yesterday I watched on MSNBC for eleven minutes as they marveled at how bright the headlights on a pickup truck looked at seven miles with their new infrared cams...
Lame lame lame lame lame, live broadcasting.
Oh and here is another thing, how secure are these reporters? Why the hell should we be delivering state of the art communications equipment to a country where the foreign press has far more advanced tech the standing government?
they should hire ventriloquists as reporters: ventriloquists can talk without moving their lips and this will save a ton of compressed bandwidth!
I think it is much more important that they fix the latency more than the bandwidth problem. The picture quality right now is acceptable. And it will improve.
But latency is a much harder problem.
I wonder how many satellites this has to bounce off of? Won? Too?
Each satellite is abou 23,000 miles out. And 22,300 miles back. Then the reporter gives an answer. Then the answer goes another ~50,000 miles. Round trip distance: about 100,000 miles.
Now let's see, at the speed of light, this is how many seconds? 0.6? Now add in all the processing time of video compression latency. This is probably even more time than just the distance to the friggin satellite(s).
No wonder they ask a question and it takes 3 seconds before the remote reporter's lips start moving. And they get into "interruption wars" and "courtesy wars" due to the extreme latency.
Those who would give up liberty in exchange for security and DRM should switch to Microsoft Palladium!
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.
And H.263 encoding is effortless, whereas MPEG-4 would require the reporter on the scene to recompile the Linux kernel before he could transmit!
No, but using an appliance is simpler than using a computer. The videophone is an appliance (although a complex one). Everyone in the message threads suggesting they just hook up a PC and hack out some software to get better codec quality is suggesting a solution that won't work in the field because it requires using a more complex system.
There is no such thing as one codec which is more "technically complicated" to the user than another
I never claimed there was. What I said was that the near-real-time use of video that would require recording, compression, and then transmission (a multi-step process), would be more complex for the user than a real-time method with lower quality.
The point is that quality is not the ultimate goal here -- reliability is. Using a real-time standardized codec guarantees that if you can get a connection to the satellite that your video will get out.
yes, MPEG4 would be wonderful, but the standard was finalized literally days ago. Once we have hardware that can compress it in real-time and be sure that they'll be able to connect to other systems using the same standard, then someone will build that into the videophone, but not before.
Say what you will about the quality of the h.xxx videoconferencing codecs, but the fact that you can get a windows PC, a mac, a unix box, a videophone, a teamstation system, and a picturetel system all in a videoconference together, over WIDELY divergent bandwiths and topologies is FAR more important than getting a few more FPS...
Recursive: Adj. See Recursive.
I am a news videographer, and as a man that does the occasional sat live, there are a few things that you should know. This is really interesting technology. The old way took a load of equipment, time, and money. Time is the problem... in a war, the stationary time is the dangerous part in a hostile country.
I know that everyone is complaining on Slashdot about the picture quality of these new suitcase devices that can transmit anywhere in the world, and they are very impressive. The issue here with these transmitters is that they had to sacrifice something to get the video image in, so it was compressed to the point of massive lossiness. It is acceptable in the news business, because, well, in a situation like this, you need to be able to get out very fast.
But to compare to current inconvenience, they are incredible. Even the newest full band KU band digital transmitters are usually packed in the size of a SMALL TEN WHEEL TRACTOR TRAILOR. Woof. Granted, the viewing of the shot on a full bandwidth is like that in the studio. But in the field without the giant tractor trailer, to get the full signal requires an engineer with a nights prep, and a Ford Econoline-size van of equipment to do it right in the field. Not less important, a rather large amount of electricity which in those situations is often hard to find. So many times you had to bring your own generators. I am not kidding the difficulty of full quality broadcasts... many of the field engineers are ex-military comm school types. It is a tough business. Matter of fact, all of news is a tough business.
I occasionally get to speak with some of the network engineers who travel overseas in hot zones, and they say that some of them keep about 4-thousand US dollars cash on them at all times just to bribe all of their equipment into the country. When Bosnia took off the engineers were some of the first ones in, and they had to weld steel plates outside their dishes so that the snipers wouldn't destroy their transmitters. They were sandbagged in. And they had a military guard.
I can only say that a device that does the work of a nights engineering and a truckload of equipment on a 12V DC source is amazing... AND IT DOES IT LIVE. This will save lives of newsmen by keeping them on the move, and it will keep us in touch in the world. This will soon change everything. I assume that very, very soon that the whole thing will go studio quality, and when it does, it will change the whole nature of live television. Imagine network cameras with this technology built into the camera itself. The world will not miss a thing. It sounds scary and Big Brother like, but for newsmen, we will be able to SHOW you, without the unbelieveability of us TELLING you what is going on.
Better communication. Perhaps more people will understand the truth out there when they see it. It is a good thing... really.
Right now MPEG4 can't be done in real-time with any consumer-level chips (of course the MPEG4 standard was only ratified a few days ago, so nobody could have manufactured them confidently anyways).
Right now most things you see are MPEG2 -- digital cable and DVDs use MPEG2 which can be done in real-time with the right hardware, but generally requires at least 2-4 megabits/s to have full-screen quality. So if you REALLY pushed a real-time board you could do a 320x240 MPEG2 at under a Mb/s, maybe even down to a few hundred kb/s.
But we're talking about a sat connection that is generally 64kb/s (sometimes 128kb/s). You can add in overhead for some kind of IP (because we're no longer using an ISDN video connection -- we're sending data), then you have to leave room for audio, which DOES have to go both ways (though you can do audio on a separate voice phone).
Regardless, you have to make a codec that works well at ~50 kb/s. A LOT of codecs (real, Windows, MPEG4, Sorensen) can do well at that low data rate, but the h.xxx protocols have been doing it successfully for several years now, and have a huge base of compatible equipment.
Right now the only way to use MPEG4 would be to compress in near-real time and transmit after the fact. Most other codecs would require the same kind of pre-processing, or basically running a streaming server from the phone (to do something like Real or WiMP). Using a computer that isn't hard, but again its the difference between an appliance and a computer.
We've done the remote streaming server trick to get better one-way video quality over sat, but honestly it wasn't so big of an improvement that it would be worth the hassle for non-techs to worry about it. There's only so much codec trickery you can pull off with 40-45k/b of bandwidth for video (we leave 16kb/s for full duplex audio).
Recursive: Adj. See Recursive.