How Video Encoder Computing Efficiency Can Impact Streaming Service Quality Mark Donnigan Vice President Marketing Beamr
Get the original LinkedIn article here: How Video Encoder Computing Efficiency Can Impact Streaming Service Quality
Mark Donnigan is Vice President of Marketing for Beamr, a high-performance video encoding technology company.
Computer system software is the bedrock of every function and department in the enterprise; appropriately, software video encoding is vital to video streaming service operations. It's possible to enhance a video codec implementation and video encoder for two but seldom 3 of the pillars. It does say that to provide the quality of video experience customers expect, video suppliers will require to evaluate business solutions that have been performance enhanced for high core counts and multi-threaded processors such as those available from AMD and Intel.
With so much upheaval in the distribution model and go-to-market organisation plans for streaming entertainment video services, it may be appealing to press down the priority stack choice of brand-new, more effective software video encoders. With software eating the video encoding function, compute performance is now the oxygen required to flourish and win versus a significantly competitive and congested direct-to-consumer (D2C) marketplace.
How Video Encoder Computing Efficiency Can Impact Streaming Service Quality
Up until public clouds and common computing turned software-based video operations mainstream, the procedure of video encoding was carried out with purpose-built hardware.
And after that, software application consumed the hardware ...
Marc Andreessen, the co-founder of Netscape and a16z the famed venture capital firm with investments in Foursquare, Skype, Twitter, box, Lyft, Airbnb, and other similarly disruptive business, penned a post for the Wall Street Journal in 2011 entitled "Why Software application Is Consuming The World." A variation of this post can be found on the a16z.com website here.
"Six decades into the computer system transformation, 4 years considering that the development of the microprocessor, and 20 years into the rise of the contemporary Web, all of the technology required to transform industries through software lastly works and can be extensively delivered at worldwide scale." Marc Andreessen
In following with Marc Andreessen's prophecy, today, software-based video encoders have actually nearly completely subsumed video encoding hardware. With software applications devoid of purpose-built hardware and able to work on ubiquitous computing platforms like Intel and AMD based x86 machines, in the data-center and virtual environments, it is entirely precise to say that "software application is consuming (or more appropriately, has actually eaten) the world."
But what does this mean for an innovation or video operations executive?
Computer system software is the bedrock of every function and department in the business; accordingly, software application video encoding is necessary to video streaming service operations. Software video encoders can scale without needing a direct boost in physical area and energies, unlike hardware.
When handling software-based video encoding, the 3 pillars that every video encoding engineer must deal with are bitrate efficiency, quality conservation, and computing efficiency.
It's possible to optimize a video codec implementation and video encoder for two however hardly ever three of the pillars. Most video encoding operations therefore focus on quality and bitrate efficiency, leaving the calculate effectiveness vector open as a sort of wild card. However as you will see, this is no longer a competitive approach.
The next frontier is software application computing efficiency.
Bitrate effectiveness with high video quality requires resource-intensive tools, which will lead to slow operational speed or a significant boost 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 outright quality is typically needed.
Codec intricacy, such as that needed by HEVC, AV1, and the upcoming VVC, is exceeding bitrate efficiency improvements and this has produced the requirement for video encoder performance optimization. Put another way, speed matters. Generally, this is not an area that video encoding specialists and image researchers require to be worried about, however that is no longer the case.
Figure 1 illustrates 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 specific 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 offered this image
For services requiring 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 private streams on a single AWS EC2 C5.18 xlarge circumstances while x265 operating in 'ultrafast' can not reach 60 FPS at 4K. As you can see in this poignant example, codec performance is directly related to the quality of service as an outcome of less makers and less complicated encoding structures required.
For those services who are mainly interested in VOD and H. 264, the right half of the Figure 1 graphic shows the efficiency benefit of an efficiency optimized codec execution that is set up to produce really high quality with a high bitrate performance. Here one can see approximately a 2x advantage with Beamr 4 compared to x264.
Video encoding compute resources cost genuine cash.
OPEX is considered thoroughly by every video distributor. Suppose entertainment experiences like live 4K streaming can not be provided reliably as an outcome of a mismatch between the video operations capability and the expectation of the customer.
Since of performance limitations with how the open-source encoder x265 uses compute cores, it is not possible to encode a live 4Kp60 video stream on a single device. This does not indicate that live 4K encoding in software isn't possible. It does say that to provide the quality of video experience customers expect, video suppliers will require to evaluate industrial options that have actually been performance optimized for high core counts and multi-threaded processors such as those offered from AMD and Intel.
The need for software application 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 for the versatility and virtualization choices they supply will experience overly complicated engineering obstacles unless they select encoding engines where multi-processor scaling is native to the architecture of the software encoder.
Here is an article that shows the speed benefit of Beamr 5 over x265.
Things to think of concerning computing effectiveness and efficiency:
Do not go after the next advanced codec without thinking about initially the complexity/efficiency ratio. Dave Ronca, who led the encoding group at Netflix for ten years and recently left to sign up with Facebook in a comparable capability, just recently released an outstanding short article on the topic of codec complexity titled, "Encoder Complexity Hits the Wall." It's tempting to think this is only a concern for video banners with 10s or hundreds of millions of customers, the exact same compromise considerations need to be thought about regardless of the size of your operations. A 30% bitrate cost savings for a 1 Mbps 480p H. 264 profile will return a 300 Kbps bandwidth savings. While a 30% savings at 1080p (H. 264), which is encoded at 3.5 Mbps, will give more than triple the Click Here return, at a 1 Mbps cost savings. The point is, we must thoroughly and systematically consider where we are investing our compute resources to get the maximum ROI possible.
A business software service will be constructed by a devoted codec engineering group that can stabilize the requirements of bitrate efficiency, quality, and compute performance. Precisely why the architecture of x264 and x265 can not scale.
Firmly insist internal groups and experts perform calculate performance benchmarking on all software application encoding solutions under consideration. The 3 vectors to determine are absolute speed (FPS), private stream density when FPS is held consistent, and the total number of channels that can be developed on a single server using a small ABR stack such as 4K, 1080p, 720p, 480p, and 360p. All encoders should produce equivalent video quality throughout all tests.
With so much upheaval in the distribution model and go-to-market business plans for streaming entertainment video services, it may be tempting to push down the concern stack selection of brand-new, more efficient software application video encoders. With software application eating the video encoding function, compute efficiency is now the oxygen needed to flourish and win versus a progressively competitive and crowded direct-to-consumer (D2C) market.
You can experiment with Beamr's software application video encoders today and get up to 100 hours of free HEVC and H. 264 video transcoding each month. CLICK HERE