John Carmack's Armadillo Aerospace "wins" Lunar Lander Challenge [video] | Hacker News Reader

John Carmack's Armadillo Aerospace "wins" Lunar Lander Challenge [video]

thelaunchpad.xprize.org

74 pointsJabavuAdams28 comments

Open articleView on HN

Until the last few seconds, I thought I was looking at a ~2' high model.

Same here! It's a lot bigger than I thought.

How about the nerves of steel on someone running up to seven foot high flames right next to a fuel tank?

Flames like those do not hold much warmth or danger.

Congratulations to John and the Armadillo team!

Commander Keen FTW!

Is it my imagination, or are there very short samples from Commander Keen playing in the second video? (E.g., about 5 seconds in.)

Those are the standard sounds made by some brand of walkie-talkie (Motorola I think) to mark begin/end transmission. Now that you mention it, they do sound an awful lot like the sounds Commander Keen made with the PC speaker.

What exactly was the challenge? The X-Prize page that should do so (http://space.xprize.org/lunar-lander-challenge) gives no idea of what the competition is or what the 2 levels of competitions are.

If I recall correctly, and am seeing what I think I see in video #2, the purpose of the challenge was to lift off from one pad, move a certain distance and land safely on a different pad.

http://www.xprize.org/blogs/wpomerantz/ng-llc-rules-explaine...

Anyone know why it doesn't tip over? Gyroscopes? Model rockets have fins to keep them going straightish, but this lander doesn't and it's not going fast enough for them to matter anyway.

It uses thrust vectoring. The engine is on a gimbal and is constantly moving to keep the vehicle upright and on course. The fact that you can hardly see the movement means the sensors and software are doing a really good job keeping it stable.

Visit http://www.armadilloaerospace.com for the project blog - goes back a few years, and covers a huge amount of hands on practical experience with materials, propellant, components, control systems etc.

It's been a bit quiet lately - I assume due to this event.

That footage looks really fake doesn't it?

I mean, I have complete faith in John Carmack but if that had come from Iran I think we would all be calling it fake.

Yes. In fact, the entire private space race is so fantastical that it all seems fake. A bunch of hackers going to space?

But its real. I love it!

People associate anything involving space flight with fake. Usually because they can't comprehend it.

It has nothing to do with comprehension it has to do with the footage looking like it was created with Maya.

It looks fine. You're just skeptical because you can't imagine it being real.

Is the oscillation at the end of the first video due to back-blast? It looked like it was increasing, as if in positive feedback.

One small flight for robot.

I definitely find it hard to doubt the veracity of the project, being that the team and Xprize foundation are both well known. However, I had to watch the videos a couple time and there's just something about them that seems... fake. Its really hard to put my finger on, but very subtle details like the shine off the metal and the structure of the flame seem like high-grade CGI. Another thing that bothers me is the shaking of the camera itself as the vehicle is in the air. I've done some work in the past few months with After Effects and one of the nice features is called "wiggle()". You can attach wiggle to cameras, null objects, and even lights and simulate a very realistic camera shake. The movement of the camera in both videos seems like it was done like this.

I'm probably dead wrong, but I feel its important to at least state such thoughts publicly. I guess that we're finally getting to a meeting point between high-grade camera systems and cgi effects that they're starting to blur enough to fool the eye, or last confuse it. Guess its similar to the uncanny valley effect.

You're doing it wrong. Look at what's hard to simulate, not at what's easy to simulate.

Metal shading is easy to simulate, so look elsewhere. Look for the engineering details that a non-rocket-engineer animator would get wrong or not know about. I guess, though that you don't know about these.

Instead, you could look at the density of information and secondary effects in the shot. Typical CGI shots, as engineered intentional artifacts, lack the level of detail and chaos / serendipity of reality.

E.g. the spalling of the concrete, the smoke/steam dissipation, the fully volumetric flame with mach-diamonds.

CGI smoke is rarely done right. If you've seen enough particle systems, you'll start to recognize their signature. CGI flame (your comments aside) doesn't usually correctly capture the features of a rocket exhaust plume. Note the colour changes and mach diamonds, as well as the subtle thrust gimballing.

Also, look for non-rendering details. E.g. the rock in one video that's just canted on its side, being supported by the back-blast of the plume. If this were CGI, then someone would have had to decide to do that. But, who would think of that?

Every CGI shot has a budget, and usually only enough work is done to "sell" the shot, within budget. If you look frame-by-frame at most CGI animations, you'll see that it's just a veneer of reality. Usually when you compare to photo or video reference, you notice all kinds of missing effects.

What may be confusing you re: the metal is that the oxidizer is LOX, which is super-cold, and causes frost to form, altering the metal's reflectivity.

Thanks for debunking me. Everything you pointed out is absolutely correct. Also, I didn't know about the oxidizer, which is very cool to learn.

It's almost like playing "Lunar Lander" in real life!

Given that John Carmack is a massive Republican jerk, part of me wants to say "socialism (and Jet A subsidization) rules!" The rest of me wants to steal the rocket and fire myself into lower earth orbit.

I don't understand the point of your Jet A comment, considering that this thing is burning methane and liquid oxygen.

You'd have to stage those. I don't think there are enough units, and there are probably significant difficulties in making many additional copies. It would be evolution in action.

Need somebody to push the button for you?

Reply on news.ycombinator.com