Boosting Streaming Profitability with IMAX StreamSmart
Video streaming providers are under tremendous pressure to make streaming TV more profitable. Expenses related to acquiring and producing original content, marketing in a fiercely competitive market, and sustaining the complex technology behind streaming have significantly surpassed subscription revenues for many organisations. While raising subscription prices may be inevitable, few things irritate viewers more and directly lead to churn than subscription increases.
One way to combat churn is to offer content worth watching and provide exceptional viewer experiences with excellent video and sound quality for those titles. That’s been IMAX’s bread and butter for decades in the movie business. If your content is worth streaming, then maintaining high video quality is crucial for customer retention in the fiercely competitive direct-to-consumer streaming market.
Organisations can increase the quality of their streams by simply sending more data, but finding the righ tbalance of quality versus bandwidth can be challenging. Sending more data incurs higher costs, further compounding the desire to reduce operating expenses. Sending insufficient data will reduce quality, risking a dreaded drop-off from frustrating viewers, especially if it’s a fan favourite—and every fan has a different favourite.
Additionally, streaming infrastructure is often complex and highly interconnected, involving numerous processes and unique steps that must work together seamlessly. As a result, many organisations are hesitant to modify their existing systems due to the perceived risks associated with any potential changes.
Streaming organisations must be able to accurately measure and swiftly assess the impact of workflow changes to validate their effectiveness. Therefore, they must have access to the right metrics that can reliably predict viewer perceptions of quality for achieving bitrate savings without compromising the viewing experience.
The Solution-—Minimising Operational Costs Without Risk
With so much focus on gaining new subscribers and reducing churn, it can be easy to overlook the low-hanging fruit of minimising operational costs. Content Delivery Network (CDN) fees represent one of the most significant expenses for streaming providers. If an organisation can successfully reduce CDN fees without impacting the viewer experience, it represents an excellent path to enhanced profitability. To minimise risk, any proposed solution must be easy to implement – ideally integrating with existing workflows to minimise disruption – and be additive to existing optimisation methods.
The Approach—Exploring Optimisation Methods
There are several established methods to control and reduce streaming bandwidth while using the same codec to create compressed video streams. These techniques, also known as rate control, include Variable Bit Rate (VBR) encoding, Content-Aware Encoding (CAE), and Per-title Encoding. Using CAE as an example, there are certain settings that can be tweaked to save bitrate, but there is no way to accurately measure output quality. Settings that work well for some content may perform much worse for other content. While these methods can be effective in reducing bitrate, they do not guarantee that quality is maintained for viewers.
To reliably reduce bitrate while balancing quality, a perceptual quality metric is essential. Such a metric should correlate with human visual perception and work seamlessly across content types, focus areas, types of impairments, and the devices on which the content is displayed. This can verify that any bitrate reductions made won’t negatively affect the viewer experience. However, measuring quality alone doesn’t solve the problem of intelligently and automatically optimising streams.
Pairing a perceptual metric with an optimisation decisioning engine enables real-time feedback during the encoding or packaging process so that content is continually optimised against a trusted human visual quality standard. In this way, the standard becomes a guardrail for consistent quality output during any bitrate reduction ensuring that quality never dips below a preset standard.
To address these challenges, IMAX has developed StreamSmart™, a cloud optimisation product that leverages its VisionScience™ AI technology. This innovative solution enables streaming providers to reduce bitrates and corresponding CDN costs significantly without compromising video quality for viewers.
StreamSmart—Reducing Bitrates Without Impacting Quality
StreamSmart isn’t an encoder; it is a software platform designed to integrate seamlessly with existing encoder workflows. It employs IMAX’s patented VisionScience AI to dynamically optimise video bitrates for popular encoders, their formats, and streaming manifests. Unlike traditional encoding methods that often result in noticeable quality degradation, StreamSmart uses advanced perceptual quality metrics that mimic human visual perception, ensuring that any reductions in bitrate are imperceptible to viewers.
Implementing StreamSmart is straightforward and designed to cause minimal disruption to existing workflows. The software communicates with encoders through an API, which allows it to integrate and enhance current optimisation techniques seamlessly without requiring extensive modifications. StreamSmart is designed to be compatible with multiple codecs and resolutions, making it a versatile tool for streaming providers. Its ability to integrate with leading third-party encoders and support advanced formats up to UHD, including AVC, HEVC, AV1, and HDR workflows, ensures that it can meet the diverse needs of modern streaming services.
StreamSmart utilises advanced machine learning algorithms to generate multiple lower-bitrate encode candidates, which are then segmented and scored for quality using the IMAX ViewerScore™ (XVS™). This score, highly correlated with human visual perception, guides the selection of the optimal segments for the final video output, pulling a lower bitrate segment only when the quality remains indistinguishable from the original.
Key Features of StreamSmart
- Dynamic Optimisation: StreamSmart analyzes every video frame, optimising compression efficiency to ensure minimal file sizes without compromising quality.
- Seamless Integration: The software overlays existing workflows and encoders, minimising disruption and enhancing current optimisation methods.
- Real-Time Processing: Capable of making real-time decisions about video quality, StreamSmart supports various content types and adapts to different devices and viewing conditions.
Understanding IMAX VisionScience AI Technology
Human visual perception is complex, and traditional quality metrics often fail to capture the intricacies of how viewers perceive video quality. IMAX VisionScience™ technology is designed to emulate the human visual system, recognising areas where bitrate reduction would be imperceptible to viewers. This allows for targeted bitrate reduction without noticeable degradation in video quality.
To reliably balance quality and bitrate reduction, it is essential to measure and track video quality accurately using quality metrics. These metrics must correlate with human visual perception and work seamlessly across all content types (including high-dynamic-range (HDR) content), viewing conditions, types of impairments, and the devices on which the content is displayed. VisionScience operates on the understanding that humans perceive changes in quality in distinct “steps,” creating zones where quality loss goes unnoticed by the human eye. This insight provides a valuable opportunity for targeted bitrate reduction without noticeable degradation in video quality in a way that remains indistinguishable to humans.
IMAX VisionScience showing human perception and zones of opportunity for bitrate savings
Because impairments are perceived differently by device and screen size, it’s crucial to know exactly how low the bitrate can go by device type, as a larger screen shows more impairments than a smaller screen does.
IMAX continues to invest in the development of VisionScience technology, ensuring that StreamSmart remains at the forefront of video optimisation. This commitment to innovation is reflected in the numerous patents and awards, including two Emmys, that the technology has received.
Measuring Quality with IMAX ViewerScore (XVS)
When evaluating quality with IMAX VisionScience, content is scored from 1-100 using the IMAX ViewerScore (XVS) metric, with 100 being the best perceptual quality possible. The XVS metric has an over 90% correlation to audience perception as measured by Mean Opinion Score (MOS). Unlike other metrics, XVS can detect quality differences that were previously undetectable.
IMAX ViewerScore XVS metric 1-100
IMAX VisionScience supports all types of content and includes both full-reference and no-reference metrics for measuring quality when there is no source for comparison. This becomes particularly important when measuring the quality of live events when the source is unavailable. If there was a problem with the original source quality, a full-reference metric might mistakenly score downstream content highly because it matches the (poor quality) source. A no-reference metric is immune to these types of errors.
Having a no-reference metric enables StreamSmart to be used effectively on live streams as well as on-demand content.
Business Value for Streaming Providers
The primary business value of StreamSmart lies in its ability to reduce CDN costs significantly. It is not unreasonable to expect savings of 15% to 25%, with little to no difference in perceived quality. This can translate to millions of dollars in annual savings for larger streaming companies, making it a compelling strategy for reducing operational costs. For example, a major streaming service provider has successfully integrated StreamSmart, optimising over 22,000 titles and streaming 160 billion GB of data. This integration has led to significant cost savings of over $25 million annually while maintaining the highest quality standards.
Typical Workflows for VOD and Live Streaming
StreamSmart integrates seamlessly into existing workflows by communicating via an API with existing encoders, minimising disruption to infrastructure and the tech stack.
IMAX StreamSmart On-Air Workflow
StreamSmart | ON-DEMAND Workflow:
- Source Encoding: The source video is encoded using the existing encoder configuration, creating the “Anchor.”
- Candidate Generation: A set of alternate encodes (Candidates) are generated using advanced machine learning.
- Quality Scoring: Each segment is scored with IMAX Viewer Score (XVS) perceptual quality metrics.
- Optimisation: The software dynamically segments and selects the best candidate for final output.
StreamSmart | ON-AIR Workflow:
- Source Encoding: The source video is encoded using the existing encoder configuration, creating the “Anchor”
- Candidate Generation: In parallel, alternate encodes (Candidates) are generated using advanced machine learning.
- Quality Scoring: All segments are scored with IMAX Viewer Score (XVS) perceptual quality metrics.
- Optimisation: The software dynamically selects the optimal segments and sends an optimised manifest file to the Origin.
StreamSmart On-Air side-by-side
Competitive Advantage
While other quality metrics are available, the IMAX ViewerScore stands out for its unique ability to accurately correlate with human vision. IMAX StreamSmart is differentiated from other optimisation methods like VMAF and content-aware encoding by offering a more accurate and faster quality measurement that aligns closely with human perception. This results in better bitrate savings and a higher ROI.
Advantages Over Competing Metrics:
- Accuracy: VisionScience provides a more precise measure of video quality than methods such as VMAF, especially for HDR content. Unlike VMAF, which is limited to reference metrics, VisionScience offers both no-reference and full-reference metrics.
- Speed: The real-time processing capability is significantly faster, making it suitable for dynamic production environments.
- Computational Efficiency: Compute resources costs money, so processing efficiency is crucial. StreamSmart cost-effectively works in the lab and in production for live and extensive libraries of on-demand content
- Comprehensive Quality Control: Unlike content-aware encoding, which adjusts bitrates based on content complexity without measuring quality, StreamSmart ensures consistent high quality across all content types.
Ensuring a Positive ROI
A critical question for streaming providers is whether the ROI of optimisation justifies the investment. The best pricing model for optimisation technology is based on a percentage of the savings achieved. StreamSmart pricing ensures that the technology's cost is always less than the savings it generates, making it a no-brainer for streaming platforms.
Conclusion
In a competitive and mature streaming market, finding ways to differentiate and remain profitable is essential. Reducing streaming costs by lowering bitrates without compromising quality is a viable strategy. IMAX’s StreamSmart, powered by IMAX VisionScience technology, offers a compelling solution for streaming providers looking to reduce costs while protecting quality. By leveraging perceptual quality metrics and advanced optimisation techniques, StreamSmart ensures significant bitrate savings, enhances viewer experience, and provides a strong ROI.
StreamSmart— Summary of Key Benefits
- Guaranteed ROI: Save 15%–25% on distribution costs with pricing tied to a fraction of the savings achieved.
- Seamless Integration: Easily integrates with existing workflows with minimal disruption.
- Enhanced Quality: Maintains high video quality across all devices and viewing conditions.
- Operational Efficiency: Streamlines encoding workflows and optimises video content in real-time.
- Proven Results: Demonstrated success with leading streaming providers
Related Articles
The newly released Wirecast Gear 3 combines a change from Intel to AMD CPUs and a rearchitected software engine that Telestream claims "uses up to 60% less computer resources than previous versions." In this review, I tested the new version of hardware and software to its limits and came away very impressed.
22 Aug 2022