Blog

The power of data

The power of data

Early on in the development of Liftoff we were faced with a choice: should we create a simulator that models realistic drone behavior in every possible way, or a game that emulates the experience, regardless of the physics and mechanical constraints that give rise to drone flight in real life? A simulator that is correct or one that simply feels correct? We did a lot of testing and research early on and we made the decision to go down the hard path. During development we’ve always had close contact with ImmersionRC and with you, our player base. You are always quick to answer our questions and give feedback and we’re very grateful for that. Each time you guys were excited, we knew we were on the right track. However, up until now, we have had no hard data to prove that what we were doing was “physically correct”.

Then came data. Lots of it.

We got some nice data sheets from ImmersionRC. They contain loads of information about motor and propeller characteristics at any given ESC signal for several different setups. Think: lift, rpm, torque, power, … In other words, we can now objectively and in fine detail check just how physically accurate Liftoff is. Even though the results of all this will not be in the next update yet, we have by no means been sitting around idly. We’ve been testing, finetuning and comparing our simulator against the new data. And we’ve done a healthy dose of swearing in the process. However, we’re happy to see the data shows that we are close, but there is still some room for improvement.

All this data crunching is tedious work, but it has an enormously important payoff: it enables us to investigate the physical accuracy of discrete parts of our simulation. Being able to mark parts of the simulation as “correct”, means that we can isolate the possible causes of some known problems. This way, we can finetune the flight behaviour without fixing – or worse, breaking – what isn’t broken.

An example of this would be the “floaty” behavior that is sometimes reported. It’s often suggested that we “just change the gravity”. The problem with this is that once we touch the gravity constant, any correct modelling of reality goes right out of the window. We might get one particular type of drone that happens to feel right with unearthly gravity settings, but every other part of the behaviour would be wrong. It would simply be a case were a bunch of wrongs incidentally make one right.

We, however, want to create a simulation that is correct. We want the flight experience to be as natural as possible, but only because it is, in fact, natural. We want to be able to add new drones and have them act as they would in real life without adjusting arbitrary fudge factors.

You have high expectations of our game. So do we. We’re doing all we can to bring you the best experience.

15 Comments

  1. Si

    Hi Guy’s
    Any chance that in the future you could add a mini quad (160ish) ?.
    They are becoming very popular and with good reason.
    They are faster and more responsive than their larger cousins.
    Thanks
    Si

    1. Elvan

      I agree with request for smaller ships.
      Many are still wanting a timer.
      Thanks guys for you hard work.
      Mr. MIYAGI!

  2. Samuel

    This seems to be a great future for liftoff.
    The only big thing still missing for me and some friends is the multiplayer mode :))

    1. LuGusStudios

      Eventually multiplayer will come, it just takes a while to get done.

  3. Some Dude

    You’re going to have to also create physics for 4″prop, 5″prop etc. and also take AUW into consideration to have it be more realistic.

  4. Pascal de muziekman

    Tommetjuuuh 🙂

  5. Daniel Smith

    I just wanted to say that was very well written and EXACTLY what I wanted to hear. Accurate modelling is exactly what makes this sim so much more useful than just a fun game. Just hurry up would ya? 😉

  6. HyphPV

    Absolutely ecstatic to hear this kind of news from you guys. You’ve already lain down the groundwork for an amazing sim and I’m super hyped to see where it winds up in the future. Hearing about your intake and evaluation of data like this is such a great sign. As a fellow developer, I feel your pain and wish you all the luck!

    Keep up the great work, guys. If you need me, I’ll be over here throwing as many crazy new tracks as I can at you and the community!

  7. Ferrett

    This – with Multiplayer? Now you’re talking.
    Awesome sim…

  8. William Hagen

    I’m not able to get my Spektrum DX6 to work with Liftoff. I have a USB cable that I use with Phoenix flight simulator, but it won’t work with Liftoff. Any updates that I need to install?

    1. daerias

      @William Hagen

      take another option that works with all simulators

      http://www.rcgroups.com/forums/showpost.php?p=33583034&postcount=52

      works well with all flight simulators for me.

  9. Oz8

    This is fantastic news. I use Liftoff to train in order not to crash my real quad as much as I would if I didn’t have a sim to test out new tricks.

    Keep it real :-p.

  10. James

    >correct modelling of reality goes right out of the window.

    I totally appreciate what you are saying about doing it “right” – but ultimately, it’s how it feels to the player that is the most important thing.

    Whether you are running an accurate behind the scenes physics system OR it’s bodged together using tweaks and fixes – the end result is what counts, how it feels to fly as a user. I don’t think anyone is expecting this game to break new ground in terms of in-game physics so I wonder if it might be more fruitful to focus on getting the feeling right rather than getting the behind the scenes mechanics of the game right.

    I applaud your effort but I also know how difficult it is to get an accurate feeling in game. It’s impossible to accurately model physics in real time without a super computer, so in that way, everything is a bodge and a fix. They are just well tweaked bodges and fixes which emulate as best they can “real behaviour”

    1. LuGusStudios

      It’s not one or the other, if you get the physics right the feel should be right too, based on data and not on personal opinions.
      “Feel” is a difficult think to get right because it can be different for everyone especially since everyone getting used to flying one specific drone configuration that might feel different to another.

      Furthermore this allows us to do more than just what we currently do in Liftoff. In theory by simply adding the “art” and the right data we should be able to simulate any type of drone, now and in the future, with ease instead of having to tweak for days or weeks based on “feel”.

  11. Technical ramblings | Liftoff

    […] that out of the way, we had another look at the “floaty” gravity. To investigate issues like this, it’s good practice to assume all of one’s assumptions […]