How We Help Hudl "Up" Their Video Quality Game
Is your streaming business burdened by escalating storage and CDN delivery costs? Is this affecting your profitability and threatening the sustainability of your enterprise?
Continue reading to discover how Hudl, a premier sports video and analytics provider, navigated a significant storage cost challenge using Visionular’s revolutionary AI-transcoding technology.
Hudl, a leading provider of sports video and analytics technology, enables teams, athletes, and coaches with the tools they need to analyze game footage and improve their performance.
Hudl’s product finds use across a diverse array of customer segments, such as –
- Coaches: for team review purposes, to analyze formations, and evaluate player performances.
- Athletes: to examine individual plays, highlighting their strengths and weaknesses, compile their best plays, create custom content for brand-building, and enhance their recruiting profiles.
- Fans: to view and share videos of their favorite athletes.
Hudl’s Product Portfolio
The Problem
Hudl’s razor-sharp customer focus and product innovation led them to serve 200,000 teams and 6 million users worldwide. However, this meteoric growth resulted in a massive increase in game footage and huge storage costs!
Hudl’s storage footprint grew to over 100PB, consisting of thousands of hours of game and practice films for 1000s of teams worldwide.
As Hudl dug into their problems, it became painfully apparent that their current technology built on AVC (H.264) needed to be improved.
And, it had to be done sooner than later, given their scale and growing needs.
As their internal investigation continued, Hudl re-evaluated its encoding ladder, driven by three critical questions.
If we change encoders,
- can we maintain and improve end-user device compatibility?
- can we reduce the bitrates and still maintain our product quality?
- what will be the impact on our turn-around times?
- can we reduce the cost of encoding and use our cloud resources more efficiently?
The answers to these questions could greatly influence Hudl – from product and business development to the user’s experience.
It was indeed time to rethink and reinvent their video compression stack.
Solution
Hudl was acutely aware that any change (major or minor) to their video codec stack, choice of bitrates, and encoding architecture would directly impact their business and, most importantly, their end-user experience!
And, in their careful search for a significantly better transcoder solution, Hudl spoke to Visionular and was impressed by Aurora4 (H.264) and Aurora5 (H.265) – Visionular’s AI-powered transcoders.
However, to be doubly sure of their choice, Hudl undertook the arduous task of multiple video codec implementations, including VideoLAN x264 (H.264), MultiCoreWare x265 (H.265), and another leading codec company (H.264, H.265) in their extensive testing process.
Hudl’s automated testing tool ran each video file through the candidate encoders thousands of times with various settings, resolutions, bitrates, and speeds. They compiled the results to highlight the most important factors while choosing a codec –
- Quality (objective and subjective)
- File size / Bitrate
- Encoding speed
- Encoding efficiency (more encoding per CPU core)
Results
The outcome of Hudl’s extensive testing made it clear that Visionular’s Aurora4 and Aurora5 stood a class apart from the competition in several areas.
Aurora4
- Aurora4 was 50% faster than x264 and the other leading encoder company.
- Comparing VMAF scores, Aurora4 delivered an ~ 17% improvement over x264 and a ~ 10% improvement over other competitors.
- 30% higher encoding FPS at the same number of vCPUs.
Aurora5
- Aurora5 was ~50% faster than x265 and the other leading encoder company
- Comparing VMAF scores, Aurora5 showed a ~19% improvement over x265 and a ~10% jump over other competitors.
- 2x – 3x encoding FPS at the same number of vCPUs.
The results stood out because the performance remained consistently high even with fewer vCPUs (i.e., more efficient per core), indicating the superior efficiency of Visionular’s libraries.
Encoding efficiency: measured as a function of the encoding FPS vs. the number of vCPUs used. Source: Hudl
Here, the performance of an encoder is judged by the size of the file it produces, its quality (objective and subjective), and its speed. Visionular blew away the competition by excelling at all these parameters while still being more efficient per core!
Subjective Testing
Visionular’s encoders exhibited dramatic improvements in visual quality as well.
Manual visual testing revealed that Visionular’s Aurora4 and Aurora5 generated clearer, sharper, and more vibrant images at lower bitrates – a significant win for Hudl and its users.
x265 4K Main@5Mbps (left) versus Aurora5 4K Main @ 4.4Mbps (right). The grass, jerseys, colors and field details are much better with Aurora5. Source: Hudl
Benefits
Better Visual Quality: Visionular’s encoders enhanced aspects critical in sports analysis, such as clarity of jersey numbers and field details.
Reduced Storage Costs: Visionular’s encoders delivered lower-bitrate and size videos than Hudl’s current stack, thus resulting in significant storage cost savings.
Reduced CDN Usage: By generating the highest quality videos at the fewest number of bits using Visionular’s encoders, Hudl was able to reduce their CDN delivery costs.
Faster Turn-Around Time: Extensive testing proved that Visionular’s technology could compress video much faster than the competition, thus reducing Hudl’s turn-around time on new content.
Device Compatibility: Hudl used Visionular’s standards-compliant HEVC encoder to produce high-quality streams that play seamlessly on end-user devices and browsers without any difficulties.
Superior Encoding Efficiency: Visionular’s encoders allow Hudl to make better use of their cloud computing budgets by being able to encode more videos per core – and faster!
“We’re excited to be in business with the Visionular crew for Hudl’s video encoding solution. Aurora4 is now a welcome part of our encoding suite. We look forward to implementing HEVC and Aurora5 into our offering very soon!” – Hudl.
Green Streaming and Encoding
In a world where millions of individuals consume billions of streaming hours daily, the need for energy efficiency for service providers cannot be overstated.
This requires a concerted effort from every entity in the video streaming pipeline – from storage and transcoding to delivery and playback to data collection.
Simply put, the “Greening of Streaming” demands serious consideration from the industry.
Visionular’s AI-powered encoders have been tuned to generate the smallest possible files with the highest video quality, at the fastest speed achievable (in terms of time and per-core performance).
This means that every video encoded using Visionular’s patented algorithms uses
- less storage
- requires fewer bits to be delivered
- uses fewer cloud computing resources
These interventions lead to less cloud computing resources for encoding, alleviate the burden on Content Delivery Networks (CDNs) for delivery, and ultimately, lead to higher energy efficiency and a significantly reduced carbon footprint for service providers.
In the larger scheme of things, this makes Visionular not only a contributor to streaming efficiency, but also a champion of sustainable, green technology in the streaming industry.
Conclusion
Impressed by the results, Hudl implemented Visionular’s encoding solutions as a part of their technology stack. Aurora4 has been integrated into their encoding suite, and they look forward to implementing Aurora5 soon.
Partnering with Visionular, Hudl has successfully enhanced the quality of their video analytics, creating a superior experience for their coaches and athletes and setting a new benchmark in their field.
At Foothill Ventures, we believe in startup companies that ride the transformative power of major technology shifts such as deep learning in computer vision. Visionular’s founders are world-class technologists in their field of video codec and AI-driven optimization. We feel privileged to support their adventure with our resources and experience.
I invested in Visionular because the team is at the forefront of innovations in video encoding and image processing for real-time low latency video communications and premium video streaming applications.
