The Key Streaming Video Challenges the Industry Faces 今天

我们已经看到流媒体视频在现代工作和娱乐的各个方面迅速崛起. 然而, 尽管在过去的二十年里取得了重大的技术突破, there remain specific hurdles that video providers continue to face. 特别是, 带宽限制, 延迟的问题, 设备兼容性方面的挑战阻碍了观看者体验无缝视频流. With continued innovation across devices that can deliver 4K, immersive or augmented reality, 360°视频, how can we make the technology powering video more reliable?

带宽限制

在过去的二十年里, we’ve seen major advancements in broadband technology, 高清视频, 流媒体. 然而, 更高分辨率的视频和更大的带宽能力提出了新的挑战, as networks are more likely to experience a bandwidth crunch.

For example, it’s common for a 720p video to be streamed at 2.5 to 5 Mbps, or for a 1080p video to be streamed at 5 to 10 Mbps. 然而, with 4K resolutions and beyond, that ratio doesn’t scale. If video providers are using H.264年AVC, they need the capacity to stream 40 Mbps. 结果是, the newer codecs of HEVC, VP9, or AV1 are critically important in today’s streaming video landscape, since they can compress to half the size of H.264年AVC.  

随着最近网络中立性的倒退和整体的掐线运动, 带宽限制 are only going to become more prevalent. 我们看到越来越多的人每天在各种设备上消费视频内容, 消费者希望能够以最好的质量观看这些内容.

在任何时间任何地点提供最高质量视频的唯一方法就是保持紧密, 无损压缩. And that’s not an easy task by any means. 就在不久之前，狄拉克和其他类似的压缩是可行的，但现在已经不可行了. 今天’s two-pass compression addresses all of these constraints, unlike the old two-pass VBR compression. Two-pass compression provides an overall package compression, 其次是二次压缩，以帮助交付更高压缩视频内部. 你可以把它想象成在传输时删除几行隔行视频, or removing 信息 that is similar, instead of transmitting unique 信息 once for every frame. 整体, 这允许提供者发送所需的内容，然后根据已知信息重建帧.

There havealso been recent advancements in new compression technologies, 但许多编解码器还需要很长时间才能被采用或合并成更成熟的编解码器. 一个例子是最近发布的一种编解码器，它可以从中心向外处理视频. 它基本上是从一个粗略的图片开始，然后从那里填充它, versus the traditional bottom-up or top-down approach. 这种方法很有趣，因为它允许更快地开始运动估计和量化处理, thus reducing the overall encoding time and improvinglatency.

Latency and Reliability Issues

With higher resolutions and bitrates, 正在交付大量的数据，而这种活动量可能会导致延迟和可靠性问题. 当观众, 我们都经历过非常熟悉的场景，即视频流在关键时刻出现延迟或完全中断. And as video professionals, 我们站在另一边，亲眼目睹了向观众大规模提供高质量视频是多么困难.

防止延迟的解决方案在于使用多cdn或SD-CDN方法. By tapping into more than one content delivery network, 流媒体提供商可以在他们最需要的时候以最短的视频包时间访问最好的网络. 该解决方案允许智能玩家分析用户活动并确定何时切换网络将改善观看体验. 将多cdn和sd - cdn视为跨cdn和服务器的负载平衡. 最终, this solution allows for more ingest points and optimized delivery, helping to make the overall process extremely resilient and robust.

Metadata and 搜索 Challenges

Human error accounts for most of the mistakes we see in cataloging video. That miscategorized 信息 then sits in archives forever, making searching for specific content a nightmare. 另外, inmostcases, 元数据标记仅限于视频的基本信息，如标题, 演员, 和流派. This lack of detail can ultimately lead to bad search results, andlead to poor qualityrecommendations. 人工智能的进步, 然而, 是否使元数据标记自动化成为可能，从而提供增强的搜索和发现功能, which can feed intofacial recognition, 对象识别, and closed captioning capabilities.

With AI you can train the system on people, places, spellings, and more. Useful metadata leads to fruitful searches for your viewers, and helps with recommendations, 所有这些都为供应商带来了更好的盈利机会.

Device Compatibility Challenges

确保视频流与市场上无数设备兼容并为之优化是提供商面临的另一个挑战. 今天, fully adaptive streams can adjust to any bitrate and resolution, making it easier to deliver the right stream to every screen. 然而, 视频提供商仍然需要决定是将多个流分发给服务提供商，还是发送一个流并对其进行转码.

To effectively deliver the highest quality video to multiple devices, 最佳解决方案是将一个流发送到服务提供商，并根据需要为设备进行转码. 这种方法确保所需的带宽可用来匹配最高质量的流. 然后，经销商可以采用该流并对其进行调整，以满足各种质量要求, which ultimately saves bandwidth and prevents overall latency, while reducing the cost for customers, 也.

今天的技术和进步已经可以克服这些问题，随时为用户提供高质量的流媒体视频, 在任何地方. 通过使用较新的编解码器, a multi-CDN or SD-CDN approach, 和人工智能技术, 也 as intelligently transcoding video streams, providers can improve the user experience and keep viewers engaged. In today’s crowded media landscape, 为了留住观众和提高盈利能力，必须解决技术上的挑战. 虽然我们都知道，创造高质量的内容是流媒体视频提供商的一个关键区别, 如果观众找不到这些内容，或者找到后无法获得流畅的观看体验，那么这些内容就毫无价值. 确保大量视频文件能够快速传输，没有延迟，并在任何设备上观看，对于流媒体提供商的成功至关重要.

 [Editor's Note: This is a contributed article from IBM. 流媒体 accepts vendor bylines based solely on their value to our readers.]

Announcing the State of Streaming 2019 Survey: We Need Your Help

StreamingMedia.Com正在进行有史以来第一次调查，计算整个流媒体行业的规模和健康状况. 花几分钟回答我们的问题，你就有可能赢得奖品!

8月26日2019

IBM Watson的斯科特变成灰色谈论实时字幕和IP视频的兴起

流媒体's Tim Siglin interviews IBM Watson Media & 气象高级解决方案经理斯科特变成灰色在流媒体东部2019.

2019年5月17日

The State of WebRTC and Low-Latency Streaming 2019

It's not a standard yet, but that will likely change. Here's a detailed look at the state of WebRTC, 这个项目最终可以大规模地提供即时视频流.

4月19日2019

教程:如何利用IBM Watson Media的最新交互式网络广播功能

以前被称为Ustream的直播平台现在允许用户在视频中添加幻灯片和字幕. 我们将深入研究这些功能，并讨论其投票和注册门的持续问题.

2019年4月17日

Latency and Device Playback Are 今天's Biggest Pain Points

Bitmovin的年度调查显示，视频流媒体生活在一个高度碎片化的世界, and no one codec will take over the market.

2018年9月11日