Netflix Isn't the Only One Rethinking ABR
Adaptive bit rate (ABR) technology helps make multiscreen video work, and work reasonably well. Depending on the bandwidth connection of a user, an online video provider can, with ABR technology, automatically tune the bit rate of a video stream to match the resources available. So if a viewer has a high-capacity fiber connection, the quality of the video stream -- and its bandwidth load -- will ramp up. While if a user is connecting to the web over a slower mobile Internet link, the quality of the stream will drop (often to a level that that's still reasonable on a smaller screen) to accommodate the lower throughput.
As Variety's Janko Roettgers reports, however, Netflix Inc. (Nasdaq: NFLX) is trying to improve on the idea of standard ABR. Instead of encoding videos for delivery strictly based on bandwidth availability, Netflix is working on encoding video more intelligently based on the type of video being shown. An animated TV show, for example, is far less visually complex than an NFL football game. So Netflix has engineers encoding these different types of video streams in different ways. The same theory can also be applied not just show by show, but scene by scene. An action sequence contains more data that has to be carefully encoded than a scene of straight dialog.
In Netflix's case, the company is encoding video streams in new ways both to decrease bandwidth load -- in some cases up to 20% -- and to improve streaming quality.
But Netflix isn't the only one working in the encoding space.
At the SCTE Cable-Tec Expo two months ago, Arris Group Inc. (Nasdaq: ARRS) VP of Marketing Duncan Potter also talked about smarter ABR technology, including the use of quantization parameter (QP) codes to assign what is essentially a complexity value to video content in order to encode it accordingly. Arris also has papers published on the subject, and at least one (and probably many) patent applications filed on related technology.
For example, three Arris engineers -- Wei Zhou, Venkatram Vajjhula and Wayne Michelsen -- filed a patent application at the end of 2014 describing a dual pass encoding system. The system uses both a primary encoder and something called a lookahead encoder that measures the complexity of a video scene based on at least one of several factors including "spatial detail measurements, correlation of motion vectors, quantization parameters, number of coded bits, color detail and buffer fullness." The lookahead encoder communicates the complexity value back to the primary encoder, and that encoder compresses the video using the information provided.
As more video processing moves into the cloud to be managed by software, there will also be more opportunities to dynamically optimize encoding as the composition of a video stream changes. Like other New IP technologies, smarter ABR will drive significantly greater efficiency in the network and improve performance at the same time.
In a few years, ABR 1.0 will look like a relic. Through the work of companies like Netflix and Arris -- not to mention Amazon.com Inc. (Nasdaq: AMZN), Ericsson AB (Nasdaq: ERIC) and InterDigital Inc. (Nasdaq: IDCC), among others -- encoding technology is guaranteed to evolve to support even higher volumes of video online as well as advanced video outputs like Ultra HD TV, and perhaps even virtual reality. (See also Video Pre-Processing Promises Bandwidth Boost and IP Video Services Explode at IBC.)
— Mari Silbey, Senior Editor, Cable/Video, Light Reading