[WIP] Monterey Laguna LIDAR

Discussion in 'Locations' started by Mario Morais, Sep 4, 2012.

  1. Tuttle

    Tuttle Technical Art Director - Env Lead

    Joined:
    Feb 14, 2012
    Messages:
    2,480
    Likes Received:
    775
    There is a bit of disinformation about this kind of data...:)

    Mario is using a Lidar point cloud to build the track. So, yes, technically he's using discrete return laser informations but they're aerial scanning;

    [​IMG]

    Aerial scanning are a great resource but you can't compare it with a laser scanned track, where you have to use ground based instruments (tripod fixed or mobile) to grab a gamma of details you can't get from aerials;

    [​IMG]

    [​IMG]

    Aerials working great for terrains, thanks to the vertical accuracy but you've a lot of limitations for other aspects, like obstruction (angles), weather (cloud obstruction), building structure at the terrain level, vegetation, points per minute, density etc...

    With mobile/fixed instruments (ground based) you can get a density of 0.005/0.015 fps (foot point spacing), a huge amount of data you can't get from airborne.

    But when you read about laser scanned tracks, you don't have to think for a superhighpoly track mesh. This is a legend...as even a laser scanned track needs to follow the game standard and the same GPU limitation as other tracks. So the main asset is to get a lot of informations about everything in the circuit...but then you've to build your own track using the point cloud as a reference. The result is a very correct elevation profile, correct and realistic cambers, correct main bumps and a huge amount of data for buildings, structures, objects etc..

    At the end of the story...the work is in the 3D Artist hands even with Lidars...like other tracks.
     
  2. GTClub_wajdi

    GTClub_wajdi Registered

    Joined:
    Feb 28, 2012
    Messages:
    3,243
    Likes Received:
    572
    So even iRacing laser-scanned circuits are not accurate 100%?
     
  3. Tuttle

    Tuttle Technical Art Director - Env Lead

    Joined:
    Feb 14, 2012
    Messages:
    2,480
    Likes Received:
    775
    Well, depends what you mean with the term "accuracy". :)

    The accuracy is still present in the RAW set of points cloud, but THEN, you've to create the track using the dataset as a main collection of reference points...but they CAN'T put all those points into the track because it never works.

    So, when you're driving above an iRacing track, or in Ferrari Virtual Accademy...you're not driving above a 1 millimeter resolution surface...but above a standard mesh which is an average output (even created with a standard modeling process) of the RAW dataset.

    100% accurate? Of course no! It can't be and can't work...but the final product can offer more accuracy thanks to the dataset used to build the track. Better slopes, better cambers, some good main bump, tons of data for buildings, obstacles, TSO etc.

    It's like a lidar car scan. You grab a point cloud of the car...then you go home and you have to create the game mesh using those data as a reference...but you can't create a mesh using the same amount of data, thinking at those points as mesh vertices...:) They aren't vertices...

    You can understand what I mean thanks to some FVA public shots;

    This is a F1 nurbu points cloud;

    [​IMG]

    Then...you can see how the mesh looks in the final product;

    [​IMG]

    As you can see the mesh it's just a standard. Pretty the same you see in other non-laser scanned tracks...but with more accuracy. That's all..
     
    Last edited by a moderator: Oct 8, 2012
    1 person likes this.
  4. GTClub_wajdi

    GTClub_wajdi Registered

    Joined:
    Feb 28, 2012
    Messages:
    3,243
    Likes Received:
    572
    Thanks mate:)
     
  5. MaXyM

    MaXyM Registered

    Joined:
    Oct 5, 2010
    Messages:
    1,774
    Likes Received:
    29
    What Wajdi and Tuttle said is only partially true: in the case of graphical representation.
    But LS data may be used to create invisible layer just for physics. And it is how iRacing makes use of it. Then, even if polygons density gives resolution for example about 1m, physics engine may be driven by 1cm precision data.

    BTW: I'm curious if rF2 provides possibility to define high resolution surface data for physics
     
    Last edited by a moderator: Oct 8, 2012
  6. Tuttle

    Tuttle Technical Art Director - Env Lead

    Joined:
    Feb 14, 2012
    Messages:
    2,480
    Likes Received:
    775
    Invisible roadnoise layers can be used by everyone. It's a common trick to get more feedback from the road..but you need proper frequencies to make this stuff working correctly. Even if you're using 1mm resolution the physics can't reach such density...so it can be pretty useless. BTW...honestly I can't feel anything special in the iRacing road feedback, maybe because there is some lack in the tire/suspension physics field, so I miss the real asset for using this method...and, thanks to God, I'm not affected by the placebo effect caused by the illusion to drive above the real deal. If it feels linear something is wrong...or it is linear. ;)

    About the 1cm, honestly, I've some doubts...maybe they're interpolating with linear connections every X reference points...that is FAR AWAY from a millimeter non-linear resolution...but the most important thing is...it's working? :)

    PS: We're hijacking the WIP thread...sorry Mario. :)
     
  7. GTClub_wajdi

    GTClub_wajdi Registered

    Joined:
    Feb 28, 2012
    Messages:
    3,243
    Likes Received:
    572
    you are right ! I played iRacing for 3-4months and I was expecting an elite-accurate ffb for the laser-scanned circuits but it was a big delusion I found iRacing ffb a bit' flat!!
     
  8. Tuttle

    Tuttle Technical Art Director - Env Lead

    Joined:
    Feb 14, 2012
    Messages:
    2,480
    Likes Received:
    775
    You're not alone...;)
     
  9. Mario Morais

    Mario Morais Registered

    Joined:
    Oct 26, 2010
    Messages:
    1,465
    Likes Received:
    169
    Nice info Tuttle.!

    LIDAR (Light Detection And Ranging) :)
     
  10. Mario Morais

    Mario Morais Registered

    Joined:
    Oct 26, 2010
    Messages:
    1,465
    Likes Received:
    169
    no prob.

    iRacing is Nascar2003 !!
     
  11. MaXyM

    MaXyM Registered

    Joined:
    Oct 5, 2010
    Messages:
    1,774
    Likes Received:
    29
    in iR you may feel less, because of powersteering simulation.
    Also microbumps are not about frequencies it self. it's about noise. but it must be not really 'micro'. for example 20 square centimeters bump, with height of 5mm may be modeled or represented by hidden layer. of course it may be modded but who will model it with bevel for example?
    Im sure it would affect physics, so it would be good technology for us.
     
  12. Tuttle

    Tuttle Technical Art Director - Env Lead

    Joined:
    Feb 14, 2012
    Messages:
    2,480
    Likes Received:
    775
    Well it's the opposite. :)

    A powersteering get you more feedback from your front wheels, especially for speed sensitive systems, not less and we're talking about vertical elements (bump maps into physics).


    PS: real bumps needs proper frequencies or you get just a missed effect or a flying car...:)
     
  13. Gearjammer

    Gearjammer Registered

    Joined:
    Jun 11, 2012
    Messages:
    1,823
    Likes Received:
    24
    Very interesting information on how laser is used for track building. I would imagine the primary reason that real drivers like to use laser scanned tracks is because they can learn exactly how far it is from beginning of straight to end as well as what degree a turn makes and what elevation changes will occur on real track. I don't think they are trying to figure out car handling or setups :)

    BTW, I have worked along side road crews and know how real asphalt is laid down. I seriously doubt that anyone could recreate the hidden grooves on the track that are left behind from the paving crew.
     
  14. Adrianstealth

    Adrianstealth Registered

    Joined:
    May 28, 2012
    Messages:
    4,578
    Likes Received:
    1,072
    I know every bit of the iracing version ( bumps dips cambers etc) so it will be interesting to try this one out

    Laguna is one if those sort of tracks you really don't like at first but then it slowly grows on you untill it's amongst your favourites

    Fingers crossed for this rf2 version
     
  15. feels3

    feels3 Member Staff Member

    Joined:
    Sep 4, 2011
    Messages:
    1,201
    Likes Received:
    142
    Hey Mario, any news, progress? :)
     
  16. Mario Morais

    Mario Morais Registered

    Joined:
    Oct 26, 2010
    Messages:
    1,465
    Likes Received:
    169
    Hi, the terrain and track modeling is ready.

    now i'm adding fences, walls and tires. this track have lots of fences an tires :)


    "Haste is the enemy of perfection". :)
     
  17. Kknorpp001

    Kknorpp001 Banned

    Joined:
    May 5, 2012
    Messages:
    1,928
    Likes Received:
    21
    Exciting! Keep it up, Mario!
     
  18. feels3

    feels3 Member Staff Member

    Joined:
    Sep 4, 2011
    Messages:
    1,201
    Likes Received:
    142
    There is no need to rush :)

    Of course it would be nice to see some onboards when it's ready. :)
     
  19. Reprobator

    Reprobator Registered

    Joined:
    Jul 30, 2012
    Messages:
    202
    Likes Received:
    6
    i strongly agree!
     
  20. jimcarrel

    jimcarrel Registered

    Joined:
    Jan 15, 2012
    Messages:
    1,720
    Likes Received:
    24
    hey, lighten up a little, I happen to still like Nascar 2003, but I dumped iRacing (lol)
     

Share This Page