Buyers' Guide to Context-Aware Encoding 2019
There are myriad ways to approach the problem of lowering overall video delivery costs, but one fundamental area has eluded the overall value chain—encoding a video, or even portions of a video, against the content (what’s in the video) or the context (where the video will be consumed).
Get instant access to our 2019 Sourcebook. Register for free to download the entire issue right now!
2018 was the year that context-aware encoding (CAE) went mainstream, and Streaming Media covered a number of advances in that space, including presentations about major on-demand video content libraries being re-encoded using discrete parameters for an entire season of an episodic show. These title-specific encoding parameters are used by the likes of Netflix and others to optimize quality on a general level, on the assumption that an episodic television program’s dozen or so individual shows in a given year will have the same “look and feel” across the entire season.
More than likely, given the advances in machine learning (sometimes also referred to as artificial intelligence or AI) as well as the move toward mass parallelization of on-demand encoding using cloud-based transcoding infrastructures, 2019 will be the year that context-aware encoding reaches critical mass.
Elsewhere in this year’s Streaming Media Industry Sourcebook, there’s an extensive discussion on the State of AI and Machine Learning, but this Buyers’ Guide will focus on two practical areas to consider when purchasing a product or service offering either context- or content-aware encoding features.
What’s the Context?
If encoding is intended to save overall bandwidth, one primary factor is the location in which content will be consumed. The traditional approach to encoding requires creating typical settings for each video in a library.
In the context of videos that will be streamed to multiple devices (e.g., mobile phones on cellular data networks as well as set-top boxes that use a wired Ethernet connection), it’s imperative to have multiple encoding parameters, one for each type of device to which the video will be delivered.
This awareness of context results in what is often referred to as a rendition—discrete combinations of bitrates, codecs, frame rates, and image resolutions that offer bandwidth-saving results for a particular device in a particular setting. For example, a mobile device using cellular data will more than likely stumble over high-bitrate encodings, whereas the same mobile device on Wi-Fi will not. This rudimentary context-aware encoding aims at specific bandwidth savings in discrete content-device-network combinations, generating a matrix of renditions against which to encode a given video library.
Even if the approach is fine-tuned at an episodic show’s per-season level, though, it is still a best-effort approach more akin to using a bludgeon rather than a scalpel for surgery. The newer approach to context-aware encoding, rolled out by Brightcove and others over the course of 2018, envisions fine-tuning parameters on a per-show basis, or even—in more aggressive solutions—on a per-scene encoding basis.
In a white paper I authored with Brightcove’s input, the company says the use of context-aware encoding “potentially offers its users the same bandwidth-savings benefits that dedicated compression teams like Netflix have enjoyed for some time now in their own per-title and per-scene encoding solutions.”
The Network Ingredient
Brightcove’s CAE solution focuses on an often-overlooked ingredient to proper context-aware encoding—the network on which content will be delivered.
It makes sense that the company would add the networking ingredient to the overall encoding recipe, since one of the promises of context-aware encoding is to generate fine-tuned renditions. As such, in the Brightcove solution, “video assets are analyzed for their optimal bitrates, considering capabilities of intended delivery devices or networks.”
Without the network context, properly encoding videos at optimal bitrates is, at best, a blind gamble.
What’s in It for the CDN or OVP?
A question I’m often asked around context-aware encoding centers on the motivation for companies— from content delivery networks (CDNs) to online video platforms (OVPs)—to offer a bandwidth-saving service. The trend seems to be to offer flat-rate packages for enterprise video customers, with pricing based around numbers of hours of content delivered rather than overall bandwidth. As such, it’s in the best interest of the CDN or OVP to figure out ways to reduce overall bandwidth, and CAE is just one of multiple viable ways to deliver those overall bandwidth savings.
Given the longevity of H.264—also referred to as MPEG-4 Part 10 or Advanced Video Coding (AVC)— the use of context-based encoding is also one of only a few ways that bandwidth can be saved while still maintaining image quality.
The good news about context-aware encoding is that, regardless of the codec, there will always be a need to balance the content-device-network delivery matrix. For that reason, CAE will continue to thrive even with the ascendancy of H.265 (also known as High-Efficiency Video Coding or HEVC).
The Need for Speed
One key differentiator between services offering a context-aware component is the speed at which they create these renditions and encoding parameters. Some solutions use multiple analysis passes, while others generate the parameters off a single video asset and then apply those rendition parameters across the whole season of an episodic show.
The encoding ladder served its purpose, but as streaming becomes more nuanced a ladder just doesn't provide enough options. It's nearly time for the video matrix.
Once revolutionary, pre-title encoding was replaced by shot-based encoding and then context aware encoding. Here's how to evaluate vendors when choosing a solution.
Despite all the hype around the cloud, plenty of use cases still call for on-prem video encoding. Here's what to look for when choosing a solution.
As files get larger, Encoding.com does its best to ensure encoding times stay small. Ludicrous HLS processes HD and UHD movies in minutes.
Streaming Media Contributing Editor Tim Siglin interviews Brightcove's Matt Smith at Streaming Media East 2018
A new generation of encoders looks at the context of content to deliver better video playback and higher efficiency. Here's what publishers need to know about CAE.
Rather than forcing content into pre-determined adaptive bitrate ladders, this system creates a unique ladder for each piece of video.