Why Publishers Don't Deploy New Codecs
What are the key factors that drive and impede new and emerging codec deployment on premium OTT streaming platforms like Netflix, Disney, and Apple TV+? Jan Ozer, Principal, Streaming Learning Center, Sr. Director, Video Technology Marketing, NETINT, Contributing Editor, Streaming Media, provides a much-needed codec deployment reality check in this clip from his presentation at Streaming Media East 2023.
“When you implement a new codec, the benefits of that codec are only going to be delivered for the small number of devices that support it," Ozer says. "So Versatile Video Coding(VVC) sounds great today, it's 50% benefit over High Efficiency Video Coding (HEVC) but it might be 0.5% of all TVs on the market, which means that you're not going to get a lot of savings because it's such a small group of targets out there."
Ozer brings up a graphic titled “Reality Check – Why Publishers Don’t Deploy New Codecs,” which breaks down the points he makes. “Encoding and storage costs are always going to be additive, meaning you're going to have to continue to support H.264, you're going to continue to have to support HEVC for your legacy devices,” he says. “And the encoding costs that you undertake for the newer codecs are always going to be much more expensive than existing codecs because they're much more complex, whether you're doing it yourself or whether you're paying for it, you're going to spend a lot more to encode to the new technologies.”
“The other factors are bandwidth costs have dropped very, very significantly,” he says. “I remember 15 years ago, bandwidth was between 50 cents and a dollar per gigabyte. And now, the lowest cost structure on CloudFront is 2 cents a gigabyte. It takes a lot of savings at 2 cents a gigabyte to make up for the cost of implementation. And those are getting higher and higher."
The Risks of Deploying New Codecs
Ozer shows a graphic titled “Risk,” which breaks down the risks and costs of implementing new codecs. “What we have is working,” he says. “We're currently distributing with H.264, maybe a little HEVC. It's all working. The players work. Customers are happy, nobody's complaining. So why do we want to change it? As I mentioned, the costs are going to increase in the short term if you deploy a new codec, because you're adding storage costs and you're adding encoding costs, the codec may not deliver the expected savings. If you don't consider what I talked about in the latter, you may think, ‘Great, you know VVC is gonna give us 50%,’ but it may not deliver that benefit at all. You may see player failures or reduced Quality of Experience (QoE), which is one of the things that Facebook experienced when they implemented reels with AV1. They found that some Android phones couldn't play it back smoothly, so they had to start doing some testing before they distributed AV1 to those phones to avoid drop frames and reduce battery life.”
He reiterates that there is always a new codec in development, which increases difficulties in new deployments. “I think AV1 really stalled VP9 deployment,” he says. “I mean, if it's 2019, do we want to deploy VP9? Well, AV1 is right around the corner. And then we had three MPEG codecs released within 18 months of each other. So how do you go forward with these choices always seeming to be coming?”
The Bottom Line
Ozer displays a graphic titled “The Bottom Line,” which underscores the contemporary codec trends. “Codecs seem to flourish best when they enable publishers to enter new markets,” he says. “And we saw that very definitely with 4K and HDR. A lot of people went to HEVC because 4K [made] no sense with H.264 and it was impossible with H.264 to deliver Dolby Vision. So that was the major migration to HEVC, not bandwidth savings.”
An audience member asks a question: “You spoke about encoding and storage costs, but are there also transmission costs, or is that negligible?”
Ozer replies, “There's some of that, but I don't think that's going to be major. So there are some very notable exceptions. In the US, YouTube, Meta, and Netflix started shipping out VP9 and AV1 primarily. I think with VP9, a lot of that was banned with savings with AV1. Perhaps it was supporting The Alliance for Open Media. In China today, we're seeing a lot of use of VVC with VVC stakeholders. So, Kwai, ByteDance, and Tencent are all big VVC IP owners, and they're starting to distribute software players with VVC. And in India, we're seeing the MX player distributed. India is one of those instances where there's very low bandwidth, so a lower bandwidth codec makes sense. On the other hand, there's also a ton of cheap Android phones that really can't play back HEVC, much less VVC. So there's always those types of considerations. And as I say, the decision to use AV1 I think was politically motivated by a bunch of companies in The Alliance for Open Media, because even when I took a year to encode a minute of video, Netflix and YouTube were deploying AV1, clearly before it was commercially ready.”
Learn more about codecs and a wide range of streaming topics at Streaming Media Connect 2023.
Amazon considers six key factors when choosing and adopting streaming codecs. AWS solutions architect Kevin Yao breaks them down in this clip from Streaming Media Connect 2023.
Which streaming codecs do Netflix and Facebook most commonly use? Jan Ozer of Streaming Learning Center asks Andrey Norkin of Netflix and David Ronca of Facebook to discuss in-depth the specific codecs they employ for their organizations.
What are the current innovations in reducing the carbon emissions and the carbon footprint of streaming? Jan Ozer of the Streaming Learning Center asks Kevin Yao of Amazon Web Services and David Ronca of Facebook about the innovations and new technologies their organizations are using to lessen the environmental impact of streaming.
id3as' Dom Robinson and Help Me Stream's Tim Siglin discuss the latest Greening of Streaming developments--including taking the conservation case to Parliament--in this interview from Streaming Media East 2022.
Companies and Suppliers Mentioned