Virtual reality (VR) was one of the most buzzworthy topics at the CES event earlier this year, and both telecom and television industries are experimenting with its potential. One of the challenges any new technology faces is interoperability, where development initiatives by individual companies can go off in different, incompatible directions. This results in format wars and expensive integration efforts, and often stalls progress and consumer adoption.
Well, that's not going to happen with VR. At least that is the hope and mission of the newly formed Virtual Reality Industry Forum (VRIF), which was officially launched just a few weeks ago.
Telco Transformation caught up with Paul Higgs, officer and treasurer at the VRIF, to get more insight into the objectives of the organization and its current status.
Telco Transformation: What is the VRIF, and why was it created?
Paul Higgs: VRIF started about a year ago, in 2015. Initially, it came out of discussions between the DTG [Digital TV Group] in the UK and various commercial entities, and centered on a desire to avoid fragmentation in the development of virtual reality. We have seen this happen in the past, where different entities develop different technologies or approaches and then it takes time to agree to a common, interoperable approach.
Historically, technology has developed but there are challenges when you go to deploy it. We've seen that with UHD, for example, and they have now created the UHD forum to fill any holes and create deployable technology. We know there is a lot of development going on in the area of virtual reality, and we want to avoid delays in being able to deploy it -- avoid gaps between the development of technology and the availability of deployable technology.
TT: Can you describe the creation of the organization -- how was it formed?
PH: It started with conversations within the DTG, but also with other entities. They were seeing development and wanted to get their members' opinions on VR. But they found that members had different opinions, so they started to look more widely, expanded the pool of people and companies. This developed into the VR Interest Group, which had no formal structure but just brought together constituents that had a stake in the development of VR. Eventually they decided there was a need for an industry group of this sort to develop specifications and drive interoperability.
TT: Who are the members today, and how is it organized?
PH: We started the formal process only in October [2015] so it's all very new. Companies like Ericsson, Huawei, Technicolor, Intel and Qualcomm were the founders. Developing such a group takes a lot of effort -- it includes developing a charter, bylaws, etc. -- and these companies have the experience having developed such initiatives in the past. It took them three to four months to get it in place and then officially launch at CES 2017.
There are 28 companies signed up at the moment, and everyone has "skin in the game" -- they are involved with developing or using VR technology in some way. Companies like DTS, Dolby, Fraunhofer are also involved with partner activities and contributing to the discussion.
TT: What are the key issues/areas that the VRIF will be concentrating on?
PH: It's still early days to really commit to a specific focus. But device interoperability is certainly a very important goal. We need to create content that can be viewed on the maximum number of devices with the fewest format challenges. A lot of this will depend on how the content is encoded and passed through the delivery pipeline. There are lots of technologies and standards bodies involved in between, and our goal is to try and reduce, simplify these so that interoperable services can be deployed relatively easily.
We also have to look at the creation of VR content. It's a different experience [from linear 2D TV], and we need to create guidelines for the right way to create content for VR.
Next page: Device interoperability
TT: What are the key issues/areas that the VRIF will be concentrating on?
PH: It's still early days to really commit to a specific focus. But device interoperability is certainly a very important goal. We need to create content that can be viewed on the maximum number of devices with the fewest format challenges. A lot of this will depend on how the content is encoded and passed through the delivery pipeline. There are lots of technologies and standards bodies involved in between, and our goal is to try and reduce, simplify these so that interoperable services can be deployed relatively easily.
We also have to look at the creation of VR content. It's a different experience [from linear 2D TV], and we need to create guidelines for the right way to create content for VR.
TT: Today, there is a loss in resolution for VR, but if you increase resolution, there's a trade-off in bandwidth. From a telecom perspective, we are concerned about the bandwidth impact of VR -- is there a solution?
PH: Yes, the visual quality is an issue. Generally, you are looking at about a quarter or less of an entire scene in a VR feed. Three-quarters is being delivered but not being seen, which is a huge waste of bandwidth. Different entities and companies are experimenting with different approaches. For some companies, a "tiling" approach is very relevant. [They] deliver only those tiles that are in the user's view at high quality and not others.
Also, VR content is being post-produced and then unicast. We may also look at multicast when the creative guidelines are in place and sorted.
There is also an approach that uses edge-rendering. This utilizes the cloud-edge, which picks up the right pieces and sends them. But latency is also an issue -- if you look to one side and the visual imagery doesn't change fast enough… so there's no one hard and fast way at the moment. Much of this remains to be seen, it's early days.
TT: Is there an approach you are most optimistic about?
PH: Well, I'm an engineer so I like the most elaborate option. Cloud-based rendering is the most interesting for me. But it has to come down to price and performance.
Also, these are all basically workarounds -- we need to persevere with this till we figure out how to get three-dimensional content to headsets. There's a little bit of evolution that is required to happen. Standards bodies such as MPEG need to look at 3D encoding specifically.
TT: How is the VRIF structured?
PH: We have a liaisons group that works with the major standards bodies, such as the W3C, MPEG, ITU -- these are the bodies that do the heavy lifting when it comes to developing technology standards, which is a key issue for us.
We are also recruiting new members. We've got a lot of members from the content production/development sector and technology vendors. We could probably add some display vendors, but we are engaged with them already.
We are likely to set up two sets of working groups -- one focused on technology and technology guidelines; and the second on the human element, looking at things like storytelling in VR, motion sickness etc. We also have people working on various other factors: content capture, quality metrics and various others. The major players are also implementing these today, so we are seeing fast progress [in our understanding of these areas.
TT: Are there concerns VR may go the same way as 3D TV?
PH: Well, that's why we are creating this organization. To make sure there is an interoperable ecosystem, which was lacking with 3D TV. Content wasn't developed with 3D in mind, it was shoehorned. They tried to make it work, but couldn't. By trying to engage with content creation, distribution, technology etc., we hope to avoid that.
TT: Do you see headsets as another challenge?
PH: Yes, it is a challenge. We need comfortable, high-quality headsets for entertainment especially. Video entertainment is a different mindset compared to gaming, which is more active. We need to find ways to get more immersion and quality. Phone-based devices offer a low resolution while we need to create a good, immersive experience. Four thousand horizontal pixels per eye is seen as offering a near-realistic experience. It seems like a good target for media technology.
TT: VR seems to cause motion sickness as well, in some people.
PH: Yes, there's a percentage that suffers significantly, and it can be immediately after seeing any VR content. For most people though, it's not a problem unless its poor quality.
VR images have to be stitched to create a 360-degree experience, and if they are not stitched properly they can cause disorientation. Similarly pixel loss or brightness or color mismatches can affect viewers. Macroblocking can be a problem -- if something is missing, the eye is immediately drawn to it and it can confuse the brain.
There's also issues with latency -- how quickly the scene moves when you turn your head. If you shift your viewpoint and nothing changes, that can upset the brain. There are other challenges -- for example, with stereoscopic 360-degree video, if you lift your head it can throw people off. There's only a single viewing perspective and that's horizontal.
So with different people, different things can cause problems. But if the quality is high and the VR content is properly created and delivered then it’s a very small percentage of people who will suffer motion sickness.
— Aditya Kishore, Practice Leader, Video Transformation, Telco Transformation