Video Optimization and Your Business
If you've got enough content, video optimization can help your viewers and your bottom line. We take a look at the offerings from three players in the space—Beamr, Cinova, and EuclidIQ—to see what each one has to offer.
Learn more about the companies mentioned in this article in the Sourcebook:
More and more video goes online each day, so it’s no surprise that we’re seeing more compression companies competing for a foothold by offering ways for content publishers to lighten their delivery load and save money in the process. While virtually all transcoding companies advertise optimizing content as one of their key features, a number of companies have hung their hats on this claim, making it the very core of their value proposition. Though these companies might have all settled on a common feature to promote, they take slightly different approaches to the problem.
But between marketing hype, a lack of technical details, and differences in business models, it can be nearly impossible for a customer to make an informed decision about which (if any) company is right for them. This article sets out to describe this marketplace of companies, explain their business models, and identify the problems they’re attempting to solve for customers. Additionally, we’ll clarify and compare some of the different technologies at work, as well as provide honest feedback on their viability.
Optimization is a bit of a code word, referring to a reduction is file sizes with no (or very little) sacrifice in quality. And now is a good time, marketwise, to be promoting yourself as an optimizer. As an industry, we’re at a stage where the codec of choice today for most, H.264, has matured, and its successor, H.265, has yet to achieve broad adoption. So optimization of the current video—offering the quality and savings promised in the next generation, but without any compatibility issues—well, that’s just downright appealing.
There are two primary reasons a company would want to focus on optimization. The first is improvement in user experience: By lowering the bitrate of the video streams, a company improves the key metrics of end user experience—faster stream start, fewer rebuffering events, and higher-quality video available on the user’s current bandwidth. The second is, of course, cost savings. By lowering bitrates, a company enables a significant reduction in delivery (CDN) costs to content providers and over-the-top service providers as well as lower overall storage costs.
How Video Optimization Works
By and large, the companies in this space are attempting to differentiate themselves through tweaks made to the quantization of the video. In video compression, quantization is a process that attempts to determine what information can be discarded safely without a significant loss in visual fidelity. Primarily this work is done as a pre- or postprocess step to the actual encoding of the content (though as you will see below, this is not always the case). Because of this aspect, the benefit of optimization will be much more relevant to on-demand assets than live streaming content. In the case of a live stream, either there’s not enough material available to make a good analysis, or the process itself is too labor intensive to be done reliably in a real-time or faster scenario.
It can become hard to precisely calculate the benefit reaped by work done in this area. In part, this is because many of the improvements in user experience are more subjective than objective. In short, optimization falls under quality of experience (QoE). QoE provides an assessment of human expectations, feelings, and perceptions about a particular product, service, or application. Video playback over the internet becomes subjective because of all the various moving parts. Obvious ones, such as type of device, connection to the network, and the resolution of the source content, are easy to attribute to the viewing experience. But more subtle issues, such as the specific CDN and its traffic at the time of playback, as well as the overall resources available to the device attempting to playback, also impact the viewing experience significantly. The fuzzy nature of QoE means results delivered in optimizing the content are hard to measure and hard to guarantee across all content.
As I mentioned, there is a wide range of companies, technologies, and customers in this conversation. Let’s quickly run down three of the companies that come up often in the conversations around video optimization.
Beamr has perhaps the most traditional offering. The company touts itself as a provider of media optimization, powering some of the world’s top web publishers, social networks, and media companies. It offers patent-pending perceptual video solutions, which reduce the bitrate of H.264 and HEVC streams by as much as 50%, preserving their full resolution and quality. It takes files that have already been encoded, including files that are encoded in several layers for adaptive bitrate streaming, and optimizes these files. The Beamr Video output files are in the same format as the input (MOV or MP4 container, H.264 or HEVC video, etc), so they remain compatible with any player that supports the original files.
The Beamr dashboard has a quick and clean “at a glance” view that tracks a variety of important details for the user.
Beamr’s optimization solution is offered both as on-prem software and as a cloud service that is accessible both as a REST API and through direct cloud storage integration (on Amazon Web Services). To date, Beamr has applied for 60 patent applications, five of which have already been granted. The on-prem software runs on Linux and is not connected to the cloud, except for the license mechanism, which is web-based. The Beamr Video cloud service is accessible both through REST APIs and through direct cloud storage integration. So a customer with a video workflow on any cloud platform can connect to the Beamr Video cloud service through the REST API, or provide credentials to his cloud storage, and Beamr Video will retrieve the video files from the user’s cloud storage and return back the optimized files.
As with any transcoder, processing times will vary based on the system the solution is installed on. Beamr can purportedly optimize 1 hour of 1080p video in 1 hour when running on a 16-core server. To further enable efficient processing of multiple files, Beamr has also built in a segmenter that divides the video file into multiple segments and processes them in parallel on different cores, ensuring the maximum performance and fastest turnaround times for a user’s video optimization jobs. Once the optimization
is complete, Beamr Video "stitches" the segments back together to create the output file. "Split and Stich" is not a new feature, but it's certainly a good one to have in place to ensure the server is kept running at 100%.
Though the product only launched 2 years ago, Beamr has gained some traction. Beamr’s target customers are primarily content publishers and OTT service providers and include Sony Crackle, Interlude, and M-GO, a joint venture between Technicolor and DreamWorks Animation.
Company Name: Beamr
Product/Service: Beamr Video
Target Customer: OTT service providers
Proposed Savings: 35–50% reduction in file size
Standout Feature: Works on adaptive bitrate content in addition to progressive files
Presenting a weather report on the industry's accelerated move toward cloud-based video acquisition and delivery.