“Generations come and generations go,
but the earth remains forever…”(Ecclesiastes 1)
In almost every other post in this blog, and maybe in any tech blog, in salesmen’s pitches, the news, magazines and commercials – it’s High Definition this, High Definition that. And ever since LifeSize announced the first ever High Definition endpoint in December 2005, the video conferencing market has gone HD-loony with “high definition” becoming a must-phrase in every discussion.
Q: So what, really, is High Definition video conferencing?
A: It’s easy – “a video conferencing system using a high definition video”.
Q: And what is “high definition video”?
A: Well, “High definition video” is a video stream of “a higher resolution than standard definition (SD)“…
The above Q&A might sound like something out of a Monty Python sketch, but this is our reality. Advances in technology come and go, just like the generations King Solomon wrote about in Ecclesiastes, and yet the video remains forever. And just like the “next generation” product I am working on will soon become the “present generation”, the “present generation” will become “legacy” and yesterday’s “high definition” will be today’s “standard definition”. And “high definition”, then, will always be the highest definition, the newest technology.
Not long ago (only three years!) CIF (352×288 pixels) ruled the video conferencing market. Back in the days of H.263, but even later when H.264 was introduced, QCIF and CIF were the standard and 4CIF (704×576 pixels) was dubbed “high resolution” or “full resolution”. Thereby positioning itself as “the next big thing”, with four times the “standard” resolution.
With advances in codecs (mainly H.264), networks, cameras and displays, LifeSize and its followers shook the ground under the feet of the market which slowly pacing towards 4CIF. It at once turned 4CIF into “standard definition” and crowned the next ruler of the video conferencing industry – 720p.
A picture is worth… 720p words? Source: LifeSize.
The video conferencing industry has long claimed that a picture is worth a thousand words. Yet it seems that just like TV was never a match to the cinema until HD screens and content came to the homes of everyone, visual communications in CIF were never a real alternative until high definition video became a reality.
But how high is “High”? Or in other words: what’s high enough to be “High”?! When some endpoints from leading vendors (take HTC’s Touch HD, for instance) use 480p (848×480), they are dismissed as “not really high definition”. On the other hand, Skype has called its VGA (640×480) video “high quality“, and got by. Microsoft’s RoundTable also uses a 640×480 view for the active speaker, and a five camera setup which gives a panoramic view of the room at 1056×144 @ 15fps.
Remember George Ou from ZDnet arguing that “standard” definition at 1Mbps can actually look better than “high” definition at 1Mbps, as it is compressed less, and unless the pipes become fatter, HD won’t win over SD? Well, the plumbers are working, and high definition has gone higher. The endpoints have been improved as well, and high quality 720p can be achieved at a little bit more than 1Mbps nowadays.
Original specs for present HD date back to the early 1980s when Japan’s NHK developed Hi-Vision, an 1125-line TV standard at 30 frames per second. Current high definition video standards in North America are a technical descendant of the original NHK system and were developed by a FCC process, led by the Advanced Television Systems Committee (ATSC). FCC eventually adopted for HDTV an interlaced 1080 line video with a maximum frame rate of 30 fps and a 720 line video, progressively scanned, with a maximum frame rate of 60 fps.
Today’s High Definition should be therefore defined threefold by the following:
- Display resolution: at least 720p.
- Number of frames per second: at least 25 or 30 frames per second (fps).
- Bitrate: at least 1Mbps
Today, most HD video conferencing (HDVC) equipment supports 720p. Telepresence endpoints support 1080p. Most HDVC vendors will release 1080p endpoints this year (2009 that is; Happy New Year, folks!), and by the end of 2009 it will probably be the standard “high def”. That will, of course, make 720p “standard”, and so one can wonder what will be the “high definition” of tomorrow…
But what about 4-1080p?! The California Institute of Telecommunication and Information Technology (CalIT2) in San Diego already has a 100 Million pixel display wall and projectors that run at four times the 1080p resolution. This system requires, BTW, a bandwidth of more than 10Gbps.
The next big thing?! An illustration of the X-Seed 4000.
Generations come and (next) generations go, display resolutions go high and high-er, and there is never enough bandwidth or processing power. Consumers will always be on the hunt for the next big thing, with big being the main issue here, and so I guess the exact definition of “high definition” will keep forever evolving.
I read your article with great interest. I agree with all your points, but I found one factual error that needs correcting. You mention the Quad-HDTV (also known as 4K) imaging systems installed at UCSD/Calit2 requires 10 Gbps to operate. I am very familiar with this system and have used it on several occasions on Calit2 projects and as part of various CineGrid projects . While its true that the uncompressed base-band signal for the 4K image format is 6 – 7.6Gbps depending on specific configurations, it is rarely operated in this mode for telepresence applications. For 4K telepresence, the system typically uses NTT’s JPEG 2000 streaming codec which reduces the required bandwidth to a user-selectable bit-rate of 160 – 800 Mbps for images up to 4096x2160x60fps at 10 bits dynamic range (4:2:2). If you are interested in more details, please write me at
I read your article with great interest. I agree with all your points, but I found one factual error that needs correcting. You mention the Quad-HDTV (also known as 4K) imaging systems installed at UCSD/Calit2 requires 10 Gbps to operate. I am very familiar with this system and have used it on several occasions on Calit2 projects and as part of various CineGrid projects . While its true that the uncompressed base-band signal for the 4K image format is 6 – 7.6Gbps depending on specific configurations, it is rarely operated in this mode for telepresence applications. For 4K telepresence, the system typically uses NTT’s JPEG 2000 streaming codec which reduces the required bandwidth to a user-selectable bit-rate of 160 – 800 Mbps for images up to 4096x2160x60fps at 10 bits dynamic range (4:2:2). If you are interested in more details, please write me at
Laurin – Thanks for setting the record straight regarding 4K. This sounds very interesting indeed.
You’ve mischaracterized my position a bit. I did not say SD at 1 Mbps is better than HD at 1 Mbps. Here’s my exact quote in the blog you linked to.
“Since most modern displays up sample SD video sources to HD resolutions anyways, it would seem to be a wash between the SD and HD solution. But this clearly isn’t the case since the HD image is noticeably superior in detail but this sounds like we’re getting a free lunch so what could explain this? What’s happening is that lower resolution of the SD image indiscriminately throws away finer detail by starting with 3 times fewer pixels whereas the H.264 video compression algorithm has the luxury of picking and choosing what it wants to discard if it had a higher resolution stream to work with.”
George – thanks for clearing your position and adding the exact quote.
I guess that, with the world going completly HD, using HD video sources, and HD-able codecs, the debate between SD and HD is no longer a debate.
Now it’s 720p vs. 1080p and 1080p vs. god knows what else.
There are never too many pixels, I guess…
Sagee,
Vimeo, YouTube, and Facebook have all moved to 720P video at around 2 Mbps. While it’s always better to have higher bitrate, it’s always better to have higher resolution too. What I have stated in the past was that 8 Mbps 480i SD video can be better than 2-4 Mbps 720P “HD” video under normal motion/image complexity.
Most people will be satisfied with 720P because that’s the display resolution their HDTVs natively supports. But as new “Full HD” 1080P LCD HDTVs increase market share as most new TVs are like this now, people will begin to prefer 1080P.
5 years from now, we’ll begin seeing a shift towards Quad-HD 2160P displays. 15 years from now, we’ll likely see a shift towards 4320P displays.
Initially the concept of video conferencing was new to me but due to the change in technologies I felt that work and the performance of business can be effectively be increased positively due to the investment I made in more than one telepresence room. However, many telepresence solutions are now standards based, so they are compatible with other desktop.
The idea which I got from a conference room environment that delivered an experience which is so true to life an executive could use the solution instead of spending time and money traveling to meetings. The investment which I made on time the returns of the company justify the expense. The profits margin has increased tremendously.
Many telepresence solutions have compatibility with the standards, but it’s limited. So while you can open up a single-screen connection to other telepresence solutions, you cannot (that I’m aware of) open up a 3-screen connection with other solutions. This is unfortunate because it effectively limits the size of the network and reduces the value of the solution.
George – I agree with you. I dedicated a whole post for that titled “Telepresence Interoperability – Deal with it!”: http://blog.radvision.com/videooverenterprise/2008/08/21/telepresence-interoperability-deal-with-it/