Announcing the Swift Death of Per-Title Encoding, 2015-2019
The per-title encoding ladder as we know it is on the way out, 它将很快被动态和上下文感知编码所取代. An encoding ladder, of course, 这是一组由单一直播或视频点播(VOD)输入创建的文件，以服务于在各种设备和连接速度上观看的观众吗. 苹果在Tech Note 2224 (TN2224)中正式定义了编码阶梯，为编码界做出了巨大贡献。, 它成为编码和打包HTTP直播(HLS)和其他自适应比特率格式文件的罗塞塔石碑. Many, many, 生产商只是简单地使用TN2224梯子，或者进行了适度的修改, and it worked just fine.
标题编码由Netflix于2015年12月发明/首次公布. 这个理论简单而令人信服:所有的视频都呈现出不同的运动和复杂程度, so it makes little sense to use a single ladder for all of them. Rather, 最好是衡量视频的复杂性，并为该内容创建一个独特的阶梯. 以低比特率编码说话头视频并节省带宽成本, 并以更高的比特率对高动态视频进行编码，以保证高带宽观众在大屏幕电视上观看时的质量.
Within weeks of Netf lix’s announcement, YouTube展示了自己的基于机器学习的逐标题技术, and within 3 years, 大多数主要的编码供应商都推出了各自版本的标题编码. Suddenly, TN2224 disappeared, replaced with the HLS Authoring Specification and a new ladder, 苹果的建议是“你可以根据你的具体内容和编码流程来评估(阶梯中的比特率)，然后进行相应的调整。.”
Of course, 保持领先的最好方法是淘汰自己的技术, Netflix对每个标题编码做了哪些动态优化, essentially a form of shot-based encoding. Rather than divide your video up into arbitrarily sized chunks, 将内容划分为场景或镜头切换，这些场景或镜头切换在几乎所有类型的镜头中都有规律地出现. This might produce segments of varying durations, 但只要片段和关键帧长度在阶梯之间保持一致, this won’t cause a problem.
Meanwhile, the benefits are many. 基于场景的编码在每次场景变化时自动插入关键帧，从而提高质量. You can adjust the bitrate for each scene as needed, 细微的差异将不太明显，因为它们发生在场景变化中. Most importantly, according to Netflix tests, 动态优化使x264的比特率节省了约28%, nearly 38% for VP9, and close to 34% for HEVC. 相当令人印象深刻的数字，至少对Netflix来说，按标题编码已经死了.
最近的进展被称为(由Brightcove)上下文感知编码(CAE). 这个概念很简单:QoE信标和网络日志提供诸如查看器的有效带宽之类的详细信息, the devices they’re using to watch the videos, and the distribution of viewing over your encoding ladder. 您的编码阶梯应该考虑这些数据以及内容的复杂性.
Brightcove has deployed CAE for more than a year, while at the 2019 NAB Show, 至少有两家公司——epic Labs和mux——透露了他们的类似想法. Epic Labs在一款名为LightFlow的产品中提供了自己的版本，目前至少有一个高容量用户, 而Mux在4月份将观众自适应编码添加到其编码堆栈中.
有趣的是，虽然每个标题甚至基于场景的编码都可以独立运行, CAE under any name needs data. As an OVP, Brightcove是一个端到端解决方案提供商，可以访问必要的数据, as does Mux, which offers both encoding and QoE monitoring. Epic Labs可以从Nice People at Work等第三方QoE供应商那里导入数据.
So when evaluating vendors, you have several additional questions to ask, 比如“你们是提供基于场景还是固定片段长度的编码??和“能否将网络和观众数据集成到编码阶梯结构中?, and if so, what data sources do you support?”
[This article appears in the June 2019 issue of Streaming Media Magazine as "Per-Title Encoding Is Dead."]
Related Articles
在标题编码中，编码服务为每个视频文件创建一个唯一的编码阶梯. 虽然可用于直播，但按标题编码最常用于视频点播.
24 Mar 2022
视频编码最初是一维数据速率调整，反映了一个简单的现实，即所有视频的编码方式不同，现在是一个包含帧率的复杂分析, resolution, color gamut, and dynamic ranges, as well as delivery network and device-related data, along with video quality metrics.
29 Jun 2021
上周，p2p视频编码公司Livepeer宣布获得800万美元的A轮融资. We ask the company's CTO for details on his company's unusual, possibly revolutionary, business model.
24 Jun 2019
The encoding ladder served its purpose, 但随着流媒体变得越来越微妙，梯子无法提供足够的选择. It's nearly time for the video matrix.
21 Jun 2019
随着Brightcove和其他公司证明了更全面的方法的价值，按标题编码的做法即将过时. 流媒体的Jan Ozer在NAB 2019上采访了Brightcove的Yuriy Reznic.
26 Apr 2019
2018年是上下文感知编码(CAE)成为主流的一年, 但2019年将是上下文感知编码达到临界质量的一年.
03 Apr 2019
新一代编码器着眼于内容的上下文，以提供更好的视频播放和更高的效率. Here's what publishers need to know about CAE.
23 Feb 2018