The Reality of Virtual Reality: Despite the Hype, VR Isn't Viable

It seems like you can’t browse the web or read a magazine without something popping up about VR. 确定, there’s a lot of hype around the technology (probably as much as, 或者甚至更多, 20世纪90年代中期围绕手机的炒作), 但消费者的接受度相对较低. Sales of VR headsets number in the low millions (according to UploadVR, 6.3M在2016年被出售), barely a drop in the bucket compared with the proliferation and market penetration of smartphones and their ilk.

Gaming is probably the most viable use case for VR, but consuming video content is right up there. The promise of immersive 360° video experiences is crucial to VR’s future growth. But what’s the reality of providing those VR 360° videos right now?

Before I answer that question, let’s get a few things out of the way. First, there’s a difference between run-of-the-mill 360° video and its immersive, VR cousin. 前一种情况在Facebook这样的地方大量出现, where users can pan around a video using a mouse and keyboard. 正是后者承载了所有的希望, 使用头戴式显示器(hmd)和, 通常, 一部手机.

Second is the quality required to deliver an immersive VR video experience. 来自Sky VR的Josh Courtney表示 keynote presentation he gave at 流媒体 East in 2016), delivering a quality VR experience with 360° video requires 4K per eye. 那是8K的内容. 第三是所需的带宽. 流式传输以H编码的8K内容.264(仍然是主要的编码格式), 一个查看器可能需要大约50mbps的吞吐量. 最后，还有延迟. 随着观众的移动, their heads to look at different perspectives in a VR video experience, any lag in rendering that content can potentially induce nausea.

你可能已经猜到了, the current outlook for delivering a top-quality VR video isn’t that great today. Not only do we not have the base of high-efficiency encoding (through codecs like HEVC, VP9, 或AV1), 但是传输它的带宽也没有, especially considering that the most popular consumption means would be through 一部手机 and an HMD like the Samsung Gear VR. 想象一下，在LTE连接上传输50mbps的数据流! The lack of 带宽 coupled with high throughput requirements lead to only one end—latency in the viewing experience.

Thankfully, lots of companies are working to solve the problems. Not only is Facebook (which acquired Oculus several years ago) pioneering new methods for immersive VR rendering, but Apple recently announced support of HEVC for its devices (perhaps a precursor to a new iPhone with VR capabilities). 其他公司, 比如Beamr和Harmonic, are making it easier to provide high-quality bitrates with lower throughput requirements. The confluence of a number of different industries working on solutions to some of the critical problems associated with VR video today—quality, 带宽, and latency—promises to accelerate the adoption and availability of VR video. 当这些挑战被克服, and as the cost to produce immersive VR video content continues to drop, 这只会刺激优质内容的生产, 因为有更多的内容可用, 消费者的接受度.

So, to answer my initial question, VR video isn’t really viable today. The reality is that premium content owners and other content creators are experimenting with the medium, 但此时不能期望ROI. Poor video quality is already a leading contributor to subscriber churn. Imagine how consumers would feel if they were paying for VR video content and the experience was laggy and pixilated, 并吸走了他们所有的带宽?

There’s no doubt that VR video experiences will improve over the next few years—带宽 requirements will lessen, 内容将更具沉浸感, and computational rendering power (especially on smartphones) will continue to climb. But if I were a content owner thinking about producing VR video? I might experiment a little to make sure my workflows were capable of supporting it, 但我可能会把钱花在别的地方.

[本文发表于2017年9月号。 流媒体杂志 作为“VR和视频的现实”."]