One of the big use-cases is going to be kids maybe doing an extra, like instead of doing their homework in the normal way in the evening, they go on-line where they join a study group where they join a teacher..
At just over 89 minutes in length, this is a special show, exploring High Fidelity from the inside, so to speak, complete with conversations with Mr. Rosedale, Ryan Karpf (HiFi’s co-founder and ex-Linden), Chris Collins and Ozan Serim, while David Rowe (perhaps more familiarly known to SL users as Strachan Ofarrel creator of the Oculus Rift compatible CtrlAltStudio viewer), who has been working with the HiFi team, becoming a guest host for the segment.
Since its founding, High Fidelity has made remarkable strides in developing its next generation, open-source virtual world environment, both technically and financially. Since April 2013, the company has undergone three rounds of funding, attracting around US $16 million, most of which has come from True Ventures, Google Ventures and, most recently, Paul Allan’s Vulcan Capital (which also participated in the October 2014 US $542 million investment round for Magic Leap). In addition, HiFi has attracted a number of high-profile advisers, including VR veteran Tony Parisi and, most recently, professors Ken Perlin and Jeremy Bailenson.
The interviews themselves are quite wide-ranging. With Dave Rowe, (known in HiFi as CtrlAltDavid) the open-source nature of the platform is explored, from the ability to download and run your owner HiFi server (aka “Stack Manager“) and client (aka “Interface“), through to the concept of the worklist, which allows contributors to bid for work on offer and get paid based on results.In Dave’s case, this has led him to working on various aspects of the platform such as integrating Leap Motion capabilities to improving eye tracking within HiFi’s avatars, so they track the movements of other avatars, just as our own eyes track other people’s facial and other movements as they interact with us.
In terms of general looks, the avatars – which have in the past been critiqued for being “cartoony” (despite it is still very early days for HiFi) – are still very much under development. In particular, Ozan Serim has been working to raise – and no pun intended here – the overall fidelity of the avatars in terms of looks and capabilities. He’s well-placed to do so, being an ex-Pixar animator.
One of the problems here is that the more real in appearance and capabilities they get, the closer the avatars come to the Uncanny Valley, which has led HiFi and Ozan to look at a number of avatar styles, from those which are very human in appearance through to those that are more “cartoonish” in looks.
A 2014 video showing Ozan’s work in improving the rigging around a more “realistic” HiFi avatar to more actually reflect mouth forms and facial movement when singing. High Fidelity now use Faceshift for real-time facial expression capture, rigging and animation, using either 3D or standard webcams
In discussing the Uncanny Valley, and particularly people’s reactions to avatars that are somewhat less-than-real (and we can include SL avatars in this, given their inability to naturally reflect facial expressions), Ozan raises the interesting question of whether people who critique the look of such avatars actually want to have a “realistic” looking avatar, or whether it is more a case of people wanting an avatar look that is appealing to their aesthetics which they can they identify with.
This is and interesting train of thought, as it is certainly true that – limitations of the avatar skeleton aside – most of us in Second Life are probably more driven to develop our avatars to a point where they have a personal aesthetic appeal, rather than in wanted them to be specifically “more realistic”.
Currently, HiFi is leaning towards a somewhat stylised avatar as seen in Team Fortress 2, which is allowing them to develop a natural-looking avatar look that doesn’t come too close to the Uncanny Valley. They use Adobe Maximo as their avatar creation tool, which Ozan views as a capable workflow package, but which may have some creative limitations. However, as an open-source environment, HiFi does offer the potential for someone to script in “in-world” character modelling tools, or at least to offer upload capabilities for avatar model generated in tools such as Blender. Avatars can also, if wanted, by uploaded as a complete package with all required / defined animations, such as walks, etc, included.
Chris Collins has very much become the voice of High Fidelity on You Tube, producing a wide range of videos demonstrating features of the platform, together with short tutorial pieces. The video above is one of his, demonstrating how to code interactive 3D content, using the Planky game as an example
While Ozan and his team work on avatar animations and rigging using real-time capture, Ryan Karpf reveals that by default, an avatar’s facial expressions are driven via the audio more than by direct capture: the mouth movement, for example, comprises 3 positions based on the audio, while a rising of voice or tone can result in the avatar’s eyebrows rising and falling. Ryan also touches on the Uncanny Valley issue of people’s increasingly discomfiture the closer avatars become to looking “photo-realistic”.
In talking to Chris Collins, an ex-Linden Lab alumni who headed the former SL Enterprise division, who now wears a number of hats at HiFi, Drax discusses how HiFi deals with the ever-changing face of the emerging VR hardware market, where headsets, input, tracking, and so on, is in something of a state of flux. Chris points out that while open-source, HiFi does have a set of strict coding standards and licensing, and offer external libraries to help support third-party SDK integration.
One of the powerful elements of High Fidelity is the ability you to have full agency over your environment, if you so wish; using the Stack Manager, you can create your own server / world / space, and control who might access it. The scripting tools similarly allow users to download and tweak elements – such as walking animations, a basic avatar appearance, etc., quickly and easily.
The interview with Ozan gives me an excuse to embed the High Fidelity cover of “Easy”, featuring Ozan on guitar, Emily Donald and another ex-Lab alumni, Andrew Meadow (Andrew Linden), collectively called “AKA”
Elsewhere in the interviews, Philip Rosedale demonstrates, HiFi unique approach to distributed simulations physics (using the Planky game, seen in one of the videos above). He also discusses more of the way High Fidelity is put together and why the company is approaching their virtual world in the way they are and how should a distributed approach can be beneficial to the company and users alike, with those offering computing time and resources on their systems to help with the rendering, etc., of distributed elements of a “world” able to earn revenue by doing so. He also offers up some compelling use-cases for VR / HMDs and the High Fidelity environment, particularly within the education sector, which certainly get one thinking (since the interviews in the show were recorded, HiFi have now launched their STEM VR Challenge – see also my report here).
This is perhaps one of the finest segments of The Drax Files Radio Hour yet produced, and the above only scratches the surface of things. The show offers an in-depth tour of High Fidelity, and Drax, already an Alpha user, is an excellent guide in opening doors within the interviews to help bright the entire concept and approach of High Fidelity to life – and the enthusiasm the team have for what they are striving to achieve.
There’s also a huge range of questions raised as a result of what is discussed here, particularly around some of the assumptions / parallels drawn when discussing some of the potentials raised, and also the advantage of continuous experiences as opposed to “canned” experiences. While not necessarily denying the sheer power of the former, the fact still remains that we live in a world where most people expect their entertainment, or the experience they wish to enjoy as a canned, packaged event they can partake of and then move on to whatever else interests them.
I also feel that perhaps Tony Parisi’s comments, as mentioned by Drax at the end of the show should be held a little more firmly in mind: the web is there, the coding is getting simpler, and the likes of Google and Mozilla are integrating VR capabilities into their offerings – all of which could go some way toward offsetting people’s interest in massive, open-ended environments in favour of things that deliver what they want, when they want. although, as Dave Rowe points out, HiFi could actually be both (and I suspect Sansar will have the same approach).
If you’ve not listened to the entire show, I urge you to do so – as noted above, it is one of the best in the series.