Video conferencing has a weight problem. The signals that send your image to me and my image to you are often too big for the networks that run between us to handle. Like digital Winnie the Poohs stuck in tiny rabbit holes, the data we send over the internet gets jammed up. The result is delayed and low-quality images.
These delays, or latency as they’re known in technical circles, are among the most frustrating aspects of video conferencing, and a legitimate barrier to the technology’s growth. The video calling vendors of the world are aware of the problem, however, and the race has been on for awhile now to find a better way to video conference.
The latest entrant comes from the pure research end of the industry, and their focus is on the underlying mechanics that govern the flow of data through a video call. These engineers have accepted that it’s hard to enlarge all the internet’s rabbit holes, so instead, they’re trying to make the bear caught in the middle smaller.
Stanford University’s Better Way to Video Conference
The research is being led by Ph.D. candidates at California’s Stanford University. The research team has developed a model, dubbed Salsify, that streamlines the encoding and transportation systems commonly used in video conferencing into a single process. Effectively, they’re combining two otherwise independent but simultaneous data streams into one transaction. The result is a smaller, more dynamic signal that can be adjusted to suit the networks it’s flowing through.
If the research team is proven right, then video conferencing speeds and reliability can be improved without a costly and improbable overhaul of the architecture of the internet.
The difference Salsify introduces is a frame-by-frame delivery, rather than one based on the overall average size of an entire data stream. It is constantly adjusting the size and speed of data delivery so that no single piece ends up causing unnecessary congestion in the system.
You can see a visual representation of how Salsify adjusts to the networks around it in this video from the Stanford team:
It’s all academic at the moment, and we end users won’t see any new Stanford engine in our video calls–at least for some time–but the theoretical implications are important. If the research team is proven right, then video conferencing speeds and reliability can be improved without a costly and improbable overhaul of the architecture of the internet. It’s the more economical option of the two broad ways we can make better video conferencing possible: smarter signals or smarter networks.
Better Video Calling Networks
There are examples of both solutions currently up and running. The latter “build a better network” model is, of course, aimed at the internet at large, but video conferencing stands to be one of the biggest beneficiaries. That’s because the data files needed for video are huge, many, many times heavier than those required to make an audio-only call, for example.
You need deep, deep pockets to cast aside decades-old civic infrastructure and start fresh.
Google is currently rolling out its own super-fast fiber-optic network that will shatter the speeds currently offered by the major domestic suppliers, and there’s a German model under development that would provide internet usage 20,000 times faster than the U.S. average! That involves physically laying down new networks, which will take some time. Not one to wait, SpaceX is already building a new network in the skies above us. The company recently placed into orbit the first of more than 12,000 satellites that should one day improve internet connections across the globe.
You need deep, deep pockets, though, to cast aside decades-old civic infrastructure and start fresh. That’s why the Stanford model, and others like it, are the more likely source of improvement.
Making Better Video Apps
Stanford’s team isn’t the only one working on adaptable, data-shrinking video technology. Video calling and messaging app Airtime has tapped into browser-based, WebRTC video calling to make its own dynamic call structure. The Sean Parker-backed app uses a central hub to optimize each data transmission to suit the device it is serving, even across a large group video chat. Video conferencing vendor Vidyo has also won praise for its unique method of compressing data down to its extremes in order to speed up video calls. These apps already provide better latency performance than older offerings such as Skype and Google Hangouts.
Then there are the outliers, those offering an entirely different approach to maximizing video conferencing performance. Chinese researchers, for example, have experimented with unhackable data transmission based on quantum mechanics, while a number of video chat apps have cropped up using the blockchain technology that underpins cryptocurrencies like Bitcoin.
All these advances will ultimately have one outcome: a better way to video conference. As end users, there’s nothing we can do to speed up their delivery, but it’s good to know that the video conferencing vendors of the world are at least aware of the problem of slow and poor-quality video delivery.
It may take a while, but better video conferencing is coming.