Archive for the 'media' Category
I have been fooling around with ThreeJS and virtual reality boilerplates for desktop and mobile browsers using Oculus and Cardboard for a while, but this just takes things to a whole new level.
A-frame is described by its creators as
an open source framework for easily creating WebVR experiences with HTML. It is designed and maintained by MozVR (Mozilla’s virtual reality team research team). A-Frame wraps WebGL in HTML custom elements, enabling web developers to create 3D VR scenes that leverage WebGL’s power, without having to learn its complex low-level API. Because WebGL is ubiquitous in modern browsers on desktop and mobile, A-Frame experiences work across desktop, iPhone (Android support coming soon), and Oculus Rift headsets.
It is not the first time we see something like this —remember VRML and more recently GLAM— but this is the first time I sense a strong design and content oriented vision behind a toolset of this kind. It has been clearly built taking into consideration the full spectrum of creative people that currently fuel the web as well as the mobile space, and this I hope will help it stick around. To see what I mean just launch http://aframe.io/ from the broswer in your iPhone if you have one (sorry androids), browse through the examples, and hit that cardboard icon.
Ever since I started working in cinematic Virtual Reality I have fantasized about the time when cameras will evolve from optics based mechanical contraptions to sensor based computational machines. Instead of projecting light into a flat image using lenses, computational photography collects data from the environment and uses it to reconstruct the scene after the fact. I find this subject matter fascinating. In fact, I almost attended Frédo Durand’s Computational Photography class at MIT, but I got too busy fooling around with symbolic programming and pattern recognition instead. I was not surprised to find out that Frédo is an advisor for the upcoming Light L16 digital camera. It looks insane and I definitely want one.
Before we had a Light 16 we had Lytro, a company famous for their shoot-first, focus-later consumer level funny looking cameras. To my knowledge this was the first time ever a data driven photography device has ever hit the consumer market. I didn’t get one, and I didn’t get their next generation DSLR model, but I always believed the Lytro guys were up to something interesting. It made total sense to me when they announced a few months ago they had begun development of a light field camera for Virtual Reality, and I even thought they might actually be the ones to pull that off.
Later I learned Wevr had been selected as a development partner to try the first working prototypes of Lytro’s VR capture system, called Immerge, and I might get to play with it before the end of this year. It will be a great relief after a couple of years dealing with custom rigs made with GoPro cameras and the limitations and difficulties inherited from having to stitch a bunch of deformed images at the very beginning of the postproduction pipeline. And since capturing light fields delivers data instead of pictures, you can move inside the scene almost like you were actually there, instead of being limited to just look around it.
Lytro CEO Jason Rosenthal sums it up in a recent press release: “To get true live-action presence in VR, existing systems were never going to get you there. To really do this, you need to re-think it from the ground up.” I can’t agree more.
Casey Reas just opened a show at the Charlie James Gallery in Chinatown LA last night. It is interesting to see how his generative work has recently shifted from the purely algorithmic —using rules and numbers as a base to create form from scratch— to a deconstructive commentary on media that utilizes content units —like digital photographs and video streams— as a source of [not quite] raw data that generates his quasi abstract forms over an extended period of time. One of his pieces, the one I photographed for this article, retrieves the main photograph from the cover page of the New York Times every day, and uses it as is as a topological stripe that stretches across the digital frame over and over again, weaving a familiar, yet unrecognizable tapestry across the big television screen that Casey chose as his canvas. Well done.
The New Context Conference, an annual conference hosted by Digital Garage and Joi Ito, co-founder of Digital Garage and current director of the MIT Media Lab, took place this year at Toranomon Hills in Tokyo, and focused on The Future of Digital Currency and Virtual Reality.
I was invited by Digital Garage to represent Wevr and talk about our virtual reality cinematic work. I also participated in a panel moderated by Joi with Daito Manabe and Rei Inamoto. I was proud to be part of such an impressive line-up.
Some time in the nineteen fifties a serious attempt was made to bring stereoscopic photography to the masses. Stereoscopic photography faces similar adoption challenges to 3D movies and Virtual Reality because up to this point there is no easy way to experience any of them without attaching a contraption to your face. An interesting note is that while 3D movies and Virtual Reality are fairly recent, stereoscopic photography has been around since the eighteen fifties, and even commercial viewers were mass produced back then.
In my quest to learn how to make my own Virtual Reality work, I got interested in the display of stereoscopic photography in VR. Inspired by the idea that properly placed in VR space, a stereoscopic photo can feel a lot like a VR sculpture of VR hologram, a moment frozen in time with a large potential to make the viewer feel “there with it” as opposed to just looking at a flat projection of it in a two dimensional picture.
While I was looking for cheap and easy stereoscopic camera systems I came across the historical mid twentieth century consumer cameras, and it was easier for me —or at least more reliable/fun/interesting— to get my hands on a couple of film cameras and start taking pictures than to find a digital solution. Eventually I came across a smartphone solution in Kickstarter that worked fairly well, but not before I was already shooting stereoscopic 35mm film all over the place. Here is an inventory of my stereoscopic gadgetry:
- Revere Stereo 33 35mm Camera Released around 1953. In perfect condition. Works great.
- Stereo Realist by the David White Company, available from 1947 to 1971. In perfect condition. Works great.
- Poppy 3D for iPhone via Kickstarter. This works great too. Except it’s a little painful to manage the files. Overall easier than shooting film, developing it, scanning it and putting it together for digital viewing, but it could be easier on the viewing/playback aspect of the product.
Once I had my pictures, the next step was to build a program that could let me look at them through a VR headset. Since I have no time or interest to learn how to use a game engine, I was left with only one option: the web browser and WebGl/ThreeJS. I already knew there are experimental builds of Firefox and Chrome that are compatible with the Oculus Rift DK2 headset, and I also wanted something that could run on a mobile browser for Google Cardboard viewing. I knew where to get sample boilerplates from the VR Chrome team and Mozilla, so all I had left to do was to find a way to feed two textures onto the same piece of geometry while rendering for each eye. Luckily I found a code example that did just that, and it didn’t take me long to adjust it to my needs. You can find the code here (Stereovision).
My Stereo Realist after a photo session
Stereovision for Cardboard on my iPhone. Featured photo taken with the Stereo Revere
Stereovision screenshot. Left and right images are rendered for the corresponding eye