Wednesday, May 21, 2014

ZeniMax fires the opening salvo.

It finally arrived.
A quick rundown of interesting things I see in this suit...

  • The assertion is not that Carmack's entire ability and experience belong to ZeniMax, or even that his game-related work all belongs to ZeniMax; the assertion is that ZeniMax is a VR hardware company.
    "4. For many years, ZeniMax invested tens of millions of dollars in research and development, including research into virtual reality and immersive technologies. In 2011 and 2012, John Carmack, a singularly experienced and highly proficient ZeniMax programmer who was at that time Technical Director for ZeniMax’s Texas-based subsidiary, id Software, conducted research to address technological issues associated with virtual reality."

  • They specifically cite Carmack's work on latency as a key area of interest and innovation:
  • "30. Carmack made breakthrough modifications to the Rift prototype based upon years of prior research at ZeniMax. Among other improvements, Carmack identified, applied, and developed proprietary solutions to address field of view, center of projection, and chromatic aberration issues; added specially-designed sensors and other hardware; and programmed software to reduce latency and to prevent distortions. Carmack noted some of that work in public remarks at the E3 Convention in 2012..."
    • This work was explicitly included in a whitepaper Carmack later released.  He thanks ZeniMax in the paper for allowing him to publish it openly--what are the terms attached to that "permission"? 
    • Carmack also notes in the E3 interviews that he incorporated elements of his Armadillo Aerospace motion-detection code in the VR testbed. He also points out how essential it was to have access to proprietary firmware from the manufacturers of the accelerometers he used (Hillcrest labs, also acknowledged in the latency Whitepaper mentioned above.) 

  • They assert ownership of every aspect of Carmack's work on the HMD testbed, including head-tracking technology: 
  • "46. Around the same time, ZeniMax also sent cables and customized sensors to Luckey and disclosed – pursuant to the Non-Disclosure Agreement – additional hardware design improvements regarding optics calibration and sensor mounting..."

What Oculus and Carmack will have to prove is that ZeniMax was not substantially interested in pursuing VR HMD development for its own sake, and that Oculus is not in competition (i.e. that Palmer Luckey was not acting in bad faith when he signed the NDA) with them.

The mind-boggling part of the lawsuit is this: what interpretation of the NDA does ZeniMax intend? In what world do you disclose secret information to a competitor and then sue them later for implementing non-patentable ideas that can be deemed to fall under that NDA? Why would Palmer Luckey have signed the NDA if he had any idea ZeniMax was interested in developing their own commercial product?

Palmer Luckey didn't need Doom or RAGE source code to make the Rift work--the games are not his job, there are plenty of developers out there working on that. Yes, he benefited from John Carmack's assistance, but that's not something proprietary to ZeniMax. Palmer Luckey knew he was doing something novel, and that it was an idea with legs given the right kind of publicity. What the Rift needed was software to show it off, and what Doom3 needed was something to make it interesting again. It worked. That's why id showed it at E3 and Quakecon.

In a hypothetical alternate reality where ZeniMax agreed with John Carmack that VR headsets needed their support, everyone with a Rift DK would also have Doom 3: BFG Edition. This year, we'd likely have seen a new version of RAGE with HMD support. In 2015 we'd likely have seen a version of Doom 4 that (while still the victim of interminable feature creep and development hell) could have been a truly innovative example of HMD support, likely becoming the "killer app" for a newly-released commercial version of the Oculus Rift. Assuming Valve or some other closely-associated developer could follow Doom 4 with a competing release, id's potential sales could double or even triple (Valve isn't a competitor to id in the way that Sony is a competitor for Microsoft; HMDs are a platform and the more titles that work with the platform you're on, the better your chances at a sale.)

Instead, ZeniMax chose to start a war with Palmer Luckey because he wouldn't respond to their investment offers. ZeniMax chose to start a war with John Carmack because they wanted him to fix Doom 4 as a 2014 2d-only release instead of a 2015 HMD release. ZeniMax might still win this lawsuit, but it seems hard to believe id Software will survive the next few years.

Monday, May 5, 2014

transcript: E3 2012 interview with Giant Bomb

1:44: Carmack started experimenting with off-the-shelf boutique hardware shortly after RAGE shipped (Oct 2011)

2:05: Carmack used code written for Armadillo Aerospace (gyro integration code) for early work interpreting data from head tracking sensors
2:27: Carmack references his custom firmware experience with Hillcrest Labs.
6:15: mentions that he didn't bother bringing any of his other prototypes to the show
8:16: mentions that he has ideas for future development of multiple-focal-plane displays
9:57: "I do an interesting thing in this demo..." (re: head tracking, discusses the potential of other kinds of sensors such as the Razer Hydra, and some discussion of demos he's iterated

11:18 "I spent some time trying to calibrate that out, but I didn't get that ready for prime time" re: issues with the use of Razer Hydra in head tracking

12:28: "what we've got here is, I have been pursuing... I've got couple of research devices that I've built, but... I ran across this Guy Palmer Luckey that had been building... he's got this enormous collection of head-mounts and he does a lot of work like this, and he had built this in his workshop that has... a 90° horizontal field of view, and a 110° vertical, so, it's enormous compared to... like the Sony device, it's five times the solid-angle ar... viewing area that you get on here. It puts you in there. Now, the resolution is low, because it's the resolution of one eye from here [points to Sony HMD] stretched over both eyes and that enormous field of view. So, you can resolve pixels in here, for some people that's more distracting than others... but, ah, the cool thing about this is, he sent me one of his prototype units to work with this an I integrated my sensors and a mounting strap, and the software for this, but, these are going to be made available as kits, it's called the Oculus Rift, and, the amazing thing about this is that, [picks up Sony headset] this is $799; a huge bargain compared to anything that came before it. Ah, the kits on this [holding Rift] are gonna be $500. Add on a tracker and a copy of Doom 3 with it, it'll be $600 plus some shipping on there. But it's still amazingly cheap, and this field of view you couldn't get in a $10,000 head-mount display, and--uh, you still can't get it today, it's, ah, it's that much higher. Now what ha--what makes this possible is that [talks about Oculus' design fundamentals: cheap optics, correct in software]

14:10 "so you can see what it's doing here where, I invert the projection that the lenses do. And this--you do all this in a pixel shader, it's all non-linear math, you can't do it in a projection matrix like you would only have years ago..."

15:32 [interviewer just asked about how close we are to a commercially-available headset experience]: "So what I think's gonna happen is... if we get these kits going out, and you build the first batch of 100 of them, they go to the whole hacker/maker crowd of people that are excited to build kits, and we'll see a lot of evolution there; we'll see all sorts of different 3D-printed mounts for these, ways to do the interocular adjustment or focussing adjustment, lots of different stuff can be experimented on, it's actually an amazing opportunity for people, that, there's lots of people that can do this level of work, and they can actually make a difference in how VR is going to evolve here, where, even with the Sony head-mount here [picks up Sony HMD] there's a little cottage industry of people... the ergonomics on this are not good, so people them in hats or welding goggles and all this different stuff; taking them apart, doing things... and that's... interesting, but [points to Oculus Rift] with something like this where you've got such a fundamentally stronger experience and people can make a bigger difference with it, that's going to be exciting. I think that you'll see people do novel, innovative things on this, so give that... you know, 12 months on there where, we have a commercial game out for this so that no matter what people hack up together they'll have something to show it off with; it'll have that existence [proof?] value of, 'here, let me show you something really awesomely cool I've been working on, and, given, you know, given a year, and we've got, another generation or two of display panels on there, ah, integrate the other sensors, do some of this other stuff on here, convince some other developers--game developers to enable support for something like this... and... I could easily see somebody... one of the majors producing a real polished piece of hardware, in a year or two's time on this. And this is... again it's more impressive when somebody's followed this for a decade on here, and you've tried the old ones that are really no good at all, because this is very much what people imagine the experience is: Virtual Reality, you block out the rest of the world, and you're around, looking around in the world, and we're finally at the cusp of being able to deliver that."


This 'blog is an attempt I'm making to gather information about what exactly is going on in the fervently-though-ambiguously-reported ZeniMax case against Oculus. From what I've sen, here's a timeline of early events:


10/04: RAGE ships, John Carmack looks for cool head-mounted display tech, starts tinkering with parts and prototypes.


John Carmack (employee of Id Software, under contractual obligations to ZeniMax) begins discussing head-mounted display tech with Palmer Luckey.

02/22: Carmack publishes a paper on mitigating latency; gives credit to Luckey for providing a Rift prototype

04/15: Palmer Luckey announces that "All of the hardest stuff (Optics, display panels, and interface hardware) is done", announces July 1 Kickstarter date.

5/24: Palmer Luckey signs NDA with ZeniMax (as owner of Id Software), which states that Id will provide "highly confidential and proprietary" information about "proprietary... software, including virtual reality (VR) testbed software and related assets..." (NDA obtained by re/code)

5/29: John Carmack demonstrates the Rift prototype for The Verge

June 2012: Carmack demonstrates early headset prototype at e3 ((here's a partial transcript from the famous Giant Bomb interview/demo.)) Mentions his work on previous prototypes, and that the Rift prototype is the only one working well enough to bring to E3. Mentions using code he developed for Armadillo Aerospace in work with head-tracking sensors, as well as custom sensor firmware he acquired from Hillcrest for his own prototype head-tracking unit (as a bolt-on to the early Rift prototype.) Rift does not appear to include any code; Carmack seems to be doing all output-side programming directly in Doom3 and outputting over VGA.


3/17: ZeniMax pulls the plug on Doom3:BFG edition pack-in

8/7 John Carmack announces role as CTO of Oculus

11/22 John Carmack officially leaves id/ZeniMax


02/3: Carmack tells USA Today that the real reason he left Id/ZeniMax is because of their refusal to participate in the Oculus Rift Kickstarter by releasing the Doom3:BFG edition.