TB Research - 360 Video and the Problem with Shiny Tech

Alan Melling
Alan Melling - March 23, 2018

Why 360 Video isn't Right for Real Estate

360 video cover image.jpeg

We live in amazing times.

It doesn’t seem too long ago that stitching a single panorama took an hour of processing and maybe an engineering degree or two. It was tedious, but the best panographers took a great amount of pride in making it perfect. Cut to today, and all of that has seemingly been compressed into a tiny multi-lens device that shoots 360 pre-stitched video! With only a press of a button! It’s almost miraculous.

I hope to convince you not to press that button. Don’t buy that button and the camera it’s attached to. There’s a problem with that shiny new tech and we’re going to get to the bottom of it.

 

A Promise, Broken

One of my guiding principles for pursuing innovation is maximal-interestingness. When making new things, we don’t always have a precise objective — sometimes we have to follow our nose, and if we’re lucky, it’ll guide us to making products that stick.

Which is why 360 video is such a huge bummer, because the idea itself is so interesting: capture all of the light coming into a single point, removing the artificial bounds of the frame. Capture more so the viewer can see more. Do that 30 or 60 times a second and you might just capture something with real meaning.

And it works really well, especially for journalism, sports, and even family videos: for video containing human movement. But that’s precisely why it doesn’t work for real estate.

 

The Problem is Attention

Video is a sequential medium: it’s meant to capture movement, but houses don’t move. There may be some gentle swaying of trees outside the window, or slight change in the light as the sun dapples through the shades, but on the whole, houses don’t move.

A house can’t tell its own story, instead they are a context for the stories of the lives lived inside them. And so it’s up to us as photographers to suggest those potential moments. Traditionally, we would use camera movement to direct the viewer’s attention: a slow dolly-in here to build dramatic intent, or a slider shot over there to build depth.

These narrative techniques have developed over generations, but do they translate to immersive video? How do we direct their attention if we can’t point that camera in the direction of the subject? With immersive video, we can't make sure that the viewer of our content is seeing the things we want them to see.
Unless you have a human in your shot (Agents count as humans too 😉) literally pointing to different areas in a guided tour, you might find it difficult in the directionless nature of 360 video.

Researchers at the BBC are grappling with this. “Since some of the key benefits of 360° video are a result of the viewer's control over their own gaze, the filmmaker must allow the viewer to retain that agency and direct gaze using subtler, unobtrusive techniques,“ such as audio and visual cues and character interaction.


In the video below, the brilliant VR filmmaker Jessica Brillhart explores how to direct attention in 360 video and how to unite each shot into a cohesive narrative. But her examples feature humans — our task of capturing and directing the viewer’s attention in real estate is even harder.

 

Can’t we just move the camera during the shot?

I love moving cameras! I spent a decade learning how to move cameras – a decade for all of the subtle tricks to guide the human eye and build a world in the mind.

With 360 video, be prepared to unlearn that decade. No sliders, no jibs, no steadicams or fancy new gimbals. You don’t even get to be in the room unless you want to wear a green-man suit (please send us pictures if you do).

This is the ultimate irony of these cameras and something that doesn’t find its way into the B&H ads: the 360 cams themselves are simple to use, but hard to shoot with day-to-day.

  


      Imagine pulling this out of your trunk at every shoot. 🙅

 

Robots to the rescue? Don’t count on it.

If we can’t be in the room with the camera, can we rig a monopod to a roomba and have it move around? It’s a fair experiment, one that’s been executed with varying levels of “success,” but the results are always a bit off-putting. There’s always something very unsettling about it, especially when watched in VR.

This technique attempts to confer a spatial experience like walking via an inherently time-based medium like video. I know they intend it to feel like a slow theme-park ride. Instead, it's more like having someone grab your head and shove it slowly down a hallway. No Bueno.

Giving your potential homebuyers the heebie-jeebies seems to be the opposite of good marketing material. Give me a few tightly-edited slider shots with a mirrorless any day.

 


      A buck a square foot? Startups are adorable...

 

Technical Issues abound

First generation 360 cams were laughably low-resolution, but newer devices have gotten better, with the minimum floor now set at 4K. If you’ve shot video with a drone, you know that 4K is a lot of K: when it fills up your drive, when you have to upload it, and when the viewer is streaming it.

Normally, I’d say higher quality is worth the investment, except in this case, the user never sees it! That beautifully expensive 4K video gets spread out around the viewer’s head, and they only see a small slice of it (about a 1/6th). All those excess megabytes are like a tree falling in the woods, never heard.

And that’s the best case scenario where the potential homebuyer can actually stream 4K. Most can’t. So those that aren’t so lucky to have a fiber connection end up with pretty bad looking video; not 480p bad, more like 120p bad. Nobody is buying that blurry house. And if they’re watching in a headset, I hope they have a bucket with them.

The biggest bummer in this collection of bummers: almost no one actually turns around! This is a sad truth of 360 video. We shoot the whole sphere and very few people actually look around. If you have 360 video posted to YouTube, check out the statistics with your own 360 videos. If you do, let us know in the comments below. 

 

Wait for stereo

Even if the technical problems I have with shooting, editing, and viewing 360 video were solved tomorrow, I still feel that monoscopic spherical video fails at its core mission of being an immersive medium. Removing the boundaries of the frame isn’t enough, it needs depth to be a whole solution.

Viewing a 360 right now in a headset is like seeing the world projected onto the inside of a 6ft beachball. It sort-of works, if you place the camera in the middle of a room where everything is at the same distance. But that’s never the case in the real world. (It also makes for a boring shot)

There are too many things working against immersive video, and if we don’t have a good solution for that depth problem, then consumers will simply move on, and rightly so. Camera manufacturers know this, which is why there are a number of high-end stereo 360 cameras emerging – for example, the Yi Halo in partnership with Google. These types of cameras typically feature sixteen or more lenses to ensure proper coverage. This also makes them expensive (about a grand a lens, on average).  

 


      A 3D model of the Yi Halo. 17 action cams all synced together.


We’re still waiting for prosumer stereo 360 capture. There’s the Vuze camera, but I’ll let you be the judge of whether it meets your quality requirements. When you have 8 times (or more) the number of lenses, sensors, computation, power management, etc… Well, those tough engineering decisions usually manifest as noisy, blown out video. Not good enough for polished real estate marketing. There are just no easy wins here. It’s a hard problem, especially for the sub-$1K market.

 

What we should shoot instead

It’s my hope that other kinds of immersive video will supersede 360 video for showing real estate.

For example, I’m very excited about the potential of Google’s new immersive video standard, VR180, which arranges two lenses simulating human eyes. The resulting stereo video should give a much better sense of depth and will be more immersive. It could succeed where 3D TV failed before it, because it is not bound by the edge of the TV frame. Instead, when viewed in a headset, it genuinely feels as if you’re peering into a memory. It’s an oddly emotional experience.

Check out Google’s intro video. I strongly suggest watching it in a headset and some headphones (sound is an essential component of immersion). Or even watch it in a cardboard viewer. The video on a 2D screen just doesn’t communicate the feeling.
   


     VR180 is another way to shoot for VR that may be more immersive than 360 video

 

For matters of practicality:

  • it will be easier to shoot with than 360. Since it only sees 180°, you can actually operate the camera and won’t have to leave the room.
  • It’ll be easier to consume, since the viewer won’t have to worry about missing stuff behind their head. This makes casual “couch VR” easier.
  • It also has the benefit of producing a regular wide-angle video for non-VR viewers. This makes it easier to justify, since you get multiple use-cases.

 

Google has partnered with two different camera makers. The Yi Horizon and Lenovo Mirage should be coming out later this spring and will be around $400. The Yi Horizon is the more likely pick, since it has a screen for framing.

Lenovo-Mirage-Solo-VR-Mirage-Camera-and-YI-Horizon-VR180-Camera.jpg
     Maybe the next generation will come with eyebrows


We’ll all have to experiment to find how to best use VR180, just like we tried with 360 video. There are open questions:

  • Can we move the camera around without making headset wearers sick? What is the best visual language to use?
  • Does it need humans to empathetically bring the viewer into the moment? Can an empty room feel immersive? Or does it feel like an empty stage after the play has ended?
  • How does editing work? Does it make the viewer feel disembodied? Or does it feel like teleporting in VR?

Lots of best practices to be figured out.

Try. Fail. Try. Fail. Try. Fail. It’s the only way to really know what works.

 

Let’s start a conversation!

A lot of the stuff I’ve said above is just from having done it a lot. Too much, maybe.
But I know that many of you have also shot a metric-ton of 360 video as well and may have different thoughts, so please share them! Leave a comment!

This close to the edge of the possible, there are no real answers yet, only intuitions. And that’s why we follow our nose.