Balancing Bitrate Optimization and Video Quality —
The video industry is in a continuous high-wire act that tends to span 2 difficult choices: reduce bitrates and risk delivering subpar video quality, or establish and meet quality thresholds without regard to budget. And even though they’ve been practicing the tightrope walk for a while, few are ready to do it without a safety net.
A survey in June 2020 on the subject of bitrate optimization and video quality drew 300 respondents and revealed 2 things about how the video industry is coping with the issue:
• First, there is no consensus in the methods used to determine the viewer experience at the consumer
• Second, even when companies understand the optimization-viewer experience balance, they often run into roadblocks in implementing solutions.
Among the findings:
• There are wide variations in the methodologies used to understand optimization, and in how viewer metrics are collected. While more than 35% of respondents relied on technical approaches such as Peak Signal-to-Noise Ratio (PSNR) (14%), Structural Similarity (SSIM) (11.3%) and Video Multimethod Assessment Fusion (VMAF) (10%), for example, 8.3% relied on viewer surveys, and 3.3% said that they had no plans to check the impact of bitrate optimization on the end viewer at all
• The top methods used by the industry at large to measure the impact of optimization on the viewer experience are: PSNR (14%), third-party platforms (13%), providers’ internal video engineers’ methodologies (12.7%), and SSIM (11.3%). PSNR (18%) was the leading methodology used by multichannel video providers, while third-party platforms (23%) was the leading broadcast option, and internal video engineers’ methodologies (27%) dominated within Over-The-Top (OTT) respondents.
• 96% of respondents are looking for ways to reduce the overall bitrate: 46% are taking the short-term route towards Content Delivery Network (CDN) cost renegotiation, 25% are taking a longer-term, more precise approach by optimizing the Adaptive Bitrate (ABR) ladder, 15% are looking at an overall approach towards reducing the overall bitrate, and 14% are building their own CDN.
• However, even though more than 85% of respondents are looking for ways to balance delivery costs with viewer quality, respondents are hitting 2 major obstacles around uncertainty and complexity. Under uncertainty issues they cite the inability to get management approval as the single biggest obstacle (20% of respondents) to optimization of video delivery. Other uncertainty obstacles cited were: “the benefits of reducing the costs will be outnumbered by the resources to implement it” (15%); “do not know how it will impact viewers” (14%) and 9% are concerned with “changing the current process”. On the complexity side they cite; “don’t have the proper tools to implement it” (15%), “too time consuming” (14%), “sounds too complicated” (7%) and “don’t have the resources to research a solution” or “don’t know a solution” (combined 6%).
The complete SSIMWAVE survey is available (see Resources and Notes) but one more finding deserves to be noted: more than a third of respondents — 35% — said that “measuring QoE success with viewers” was the top issue they were experiencing as they scale the number of subscribers. Everything else — including increased delivery costs (21.2%) and inefficiencies across the video stack (20.7%) — paled by comparison.
Peter Drucker, the 20th Century management consultant, said You can’t manage what you don’t measure. What is clear from the survey answers is that the bitrate optimization — video quality equation — is a real problem for the industry. We have too complex a matrix of optimization initiatives and an absolute lack of proper standard measurement approaches. As a result, video engineers who would be quick to implement a single solution are paralyzed by the plethora of choices.
Not All Bits Are Created Equal
What’s worse is the implication by many respondents that too much optimization actually can backfire in the event that those efforts would lead to a negative effect on the viewer experience, inciting subscriber dissatisfaction and potentially contributing to churn. When seen in this light, the difficulty of securing management approval is not surprising. Without the ability to demonstrate conclusively the impact of optimization on the viewer, it can be challenging to make a convincing argument that bitrate and cost savings will result in positive business impact despite the clear savings.
In too many situations, increases in bitrate are equated with increases in quality, when the reality is that not all bits are created equal. There is no value in treating an animated series geared to preschoolers in the same way as a first-run premium film. Such a solution — brute force bitrate reduction across all content — invariably results in dramatically uneven viewer experiences across channels, titles, geographies, video players, and viewing devices. On top of that, such an approach does not perform to its maximum potential balancing the cost-quality equation, by a long margin, due to lack of statistically significant, actionable, and accurate data on the degradation of the viewer experience due to the aforementioned modifications.
In an environment in which choice is driving viewing, and consumers have little desire to engage with sub-par video quality, achieving the right balance between bitrate optimization and viewer experience is crucial. Poorly executed bitrate optimization can result in degradation of the viewer experience that could have lasting results for streaming providers.
Published reports indicate that 1 in 2 streaming subscribers have cancelled a service because they were unhappy with the video quality. In a report earlier this year, Nielsen noted that 77% of respondents cited “Streaming/Playback Quality” as very important. And with consumers spending more time streaming due to pandemic quarantines, an increasing number of viewers may be sampling a service for the first time. A single bad experience may send them in another direction and make them unlikely to return.
• Implementation of a single, universal metric for measuring the impact of bitrate optimization on the viewer experience, so that the industry is not paralyzed by the variety of choices and outcome uncertainty that currently face it.
• Predictability that can show what can happen when bitrates are dialed up or down, allowing technical teams to map strategies to desired results to ensure optimal customer experiences and to secure the needed management buy-in.
• Achievement of bitrate optimization savings byadjusting resources according to the required quality of the content. For example, resources consumed for delivery of high-quality content would be offset by the reduction of consumption and costs in lower-impact areas.
With more and more consumers engaging with streaming content this year, there has never been a more important time for the industry to put its best foot forward in terms of experiences. Recognition that it must be done is a first step; what’s needed now are the tools that provide objective metrics and the strategic vision to implement them so that streaming providers can both reduce costs and optimize viewer satisfaction.
Resources and Notes
1. SSIMWAVE survey. “Webinar Overview: Best practices for optimizing bitrates, costs and viewer experiences”. By SSIMWAVE, July 20, 2020. Featured, News. https://www.ssimwave.com/news/webinar-overview-best-practices-for-optimizing-bitrates-costs-and-viewer-experiences/
2. Conviva. “One in Five Viewers Will Abandon Poor Streaming Video Experiences Immediately”. November 18, 2015. https://www.conviva.com/press-releases/one-in-five-viewers-will-abandon-poor-streaming-video-experiences-immediately/
3. Nielsen Media. ”Playback Time: Which Consumer Attitudes Will Shape the Streaming Wars?” Media 02-11-2020. https://www.nielsen.com/us/en/insights/article/2020/playback-time-which-consumer-attitudes-will-shape-the-streaming-wars/