Read the original LinkedIn article here: How Video Encoder Computing Efficiency Can Impact Streaming Service Quality
Written by:
Mark Donnigan is VP Marketing at Beamr, a high-performance video encoding technology company.
Computer software is the bedrock of every function and department in the business; appropriately, software video encoding is important to video streaming service operations. It's possible to optimize a video codec execution and video encoder for 2 but rarely 3 of the pillars. It does say that to provide the quality of video experience customers expect, video distributors will need to examine industrial services that have actually been performance enhanced for high core counts and multi-threaded processors such as those available from AMD and Intel.
With so much turmoil in the distribution model and go-to-market service plans for streaming home entertainment video services, it may be appealing to press down the top priority stack selection of new, more effective software video encoders. With software consuming the video encoding function, compute performance is now the oxygen needed to grow and win versus a progressively competitive and congested direct-to-consumer (D2C) market.
How Video Encoder Computing Efficiency Can Impact Streaming Service Quality
Until public clouds and ubiquitous computing turned software-based video operations mainstream, the procedure of video encoding was carried out with purpose-built hardware.
And then, software consumed the hardware ...
Marc Andreessen, the co-founder of Netscape and a16z the well known equity capital company with financial investments in Foursquare, Skype, Twitter, box, Lyft, Airbnb, and other equally disruptive business, penned a post for the Wall Street Journal in 2011 titled "Why Software application Is Consuming The World." A version of this post can be found on the a16z.com site here.
"6 decades into the computer revolution, four years since the innovation of the microprocessor, and two decades into the increase of the modern-day Web, all of the innovation needed to transform industries through software application lastly works and can be commonly delivered at global scale." Marc Andreessen
In following with Marc Andreessen's prediction, today, software-based video encoders have actually nearly totally subsumed video encoding hardware. With software applications devoid of purpose-built hardware and able to run on common computing platforms like Intel and AMD based x86 devices, in the data-center and virtual environments, it is completely precise to state that "software is eating (or more properly, has eaten) the world."
What does this mean for a technology or video operations executive?
Computer software application is the bedrock of every function and department in the business; accordingly, software application video encoding is vital to video streaming service operations. Software video encoders can scale without needing a linear boost in physical space and utilities, unlike hardware. And software application can be moved around the network and even whole data-centers in near real-time to satisfy capability overruns or momentary surges. Software application is much more versatile than hardware.
When handling software-based video encoding, the three pillars that every video encoding engineer should address are bitrate effectiveness, quality preservation, and computing efficiency.
It's possible to optimize a video codec application and video encoder for 2 but seldom 3 of the pillars. Many video encoding operations hence focus on quality and bitrate performance, leaving the calculate performance vector open as a sort of wild card. However as you will see, this is no longer a competitive technique.
The next frontier is software application computing efficiency.
Bitrate efficiency with high video quality requires resource-intensive tools, which will result in slow functional speed or a significant increase in CPU overhead. For a live encoding application where the encoder should run at high speed to reach 60 frames-per-second (FPS), a compromise in bitrate efficiency or absolute quality is frequently needed.
Codec intricacy, such as that needed by HEVC, AV1, and the upcoming VVC, is surpassing bitrate performance advancements and this has produced the requirement for video encoder efficiency optimization. Put another method, speed matters. Generally, this is not an area that video encoding practitioners and image researchers require to be worried about, but that is no longer the case.
Figure 1 shows the advantages of a software application encoding implementation, which, when all characteristics are normalized, such as FPS and objective quality metrics, can do twice as much work on the exact same AWS EC2 C5.18 xlarge instance.
In this example, the open-source encoders x264 and x265 are compared to Beamr's AVC and HEVC encoders, Beamr 4, and Beamr 5.
No alt text provided for this image
For services needing to encode live 4Kp60, one can see that it is possible with Beamr 5 however not with x265. Beamr 5 set to the x264 comparable 'ultrafast' mode can encode 4 specific streams on a single AWS EC2 C5.18 xlarge circumstances while x265 operating in 'ultrafast' can not reach 60 FPS at 4K. more info here As you can see in this poignant example, codec performance is directly related to the quality of service as a result of less makers and less complicated encoding frameworks required.
For those services who are mostly worried about VOD and H. 264, the ideal half of the Figure 1 graphic programs the efficiency benefit of a performance enhanced codec application that is set up to produce really high quality with a high bitrate performance. Here one can see as much as a 2x advantage with Beamr 4 compared to x264.
Video encoding compute resources cost genuine cash.
OPEX is considered thoroughly by every video distributor. Expect home entertainment experiences like live 4K streaming can not be delivered reliably as a result of an inequality between the video operations ability and the expectation of the consumer.
Due to the fact that of performance constraints with how the open-source encoder x265 uses calculate cores, it is not possible to encode a live 4Kp60 video stream on a single device. This does not imply that live 4K encoding in software application isn't possible. But it does state that to deliver the quality of video experience consumers anticipate, video distributors will need to evaluate business solutions that have been efficiency optimized for high core counts and multi-threaded processors such as those offered from AMD and Intel.
The need for software to be enhanced for greater core counts was recently highlighted by AMD CTO Mark Papermaster in an interview with Tom's Hardware.
Video distributors wishing to utilize software application for the flexibility and virtualization alternatives they supply will come across overly made complex engineering hurdles unless they select encoding engines where multi-processor scaling is native to the architecture of the software encoder.
Here is a post that shows the speed advantage of Beamr 5 over x265.
Things to consider concerning computing efficiency and performance:
It's tempting to think this is just a concern for video streamers with 10s or hundreds of millions of customers, the exact same trade-off factors to consider should be considered regardless of the size of your operations. While a 30% cost savings at 1080p (H. 264), which is encoded at 3.5 Mbps, will give more than triple the return, at a 1 Mbps savings. The point is, we need to carefully and systematically consider where we are spending our calculate resources to get the maximum ROI possible.
A commercial software application solution will be developed by a dedicated codec engineering team that can stabilize the requirements of bitrate efficiency, quality, and calculate performance. Exactly why the architecture of x264 and x265 can not scale.
Firmly insist internal teams and specialists conduct compute efficiency benchmarking on all software encoding services under factor to consider. The three vectors to measure are outright speed (FPS), individual stream density when FPS is held continuous, and the overall number of channels that can be produced on a single server utilizing a nominal ABR stack such as 4K, 1080p, 720p, 480p, and 360p. All encoders should produce equivalent video quality throughout all tests.
The next time your technical team prepares a video encoder shoot out, make sure to ask what their test strategy is for benchmarking the compute performance (performance) of each solution. With a lot turmoil in the circulation model and go-to-market service strategies for streaming entertainment video services, it may be tempting to press down the top priority stack selection of new, more efficient software video encoders. Surrendering this work could have a genuine effect on a service's competitiveness and ability to scale to meet future home entertainment service requirements. With software consuming the video encoding function, calculate efficiency is now the oxygen required to prosper and win versus an increasingly competitive and crowded direct-to-consumer (D2C) marketplace.
You can experiment with Beamr's software application video encoders today and get up to 100 hours of totally free HEVC and H. 264 video transcoding each month. CLICK ON THIS LINK
