Dr. Kazunori Yamauchi Gives Lecture on Gran Turismo's Driving Physics & Production

Discussion in 'Gran Turismo Sport' started by GTPNewsWire, Oct 31, 2017.

  1. MrWaflz55

    MrWaflz55

    Messages:
    4,428
    Location:
    Canada
    :lol:
     
    MikeV27 likes this.
  2. rono_thomas

    rono_thomas Premium

    Messages:
    2,093
    Welcome to GTPlanet :gtpflag:
     
    PirovacBoy and Deadheadbill like this.
  3. DGXR

    DGXR

    Messages:
    1,068
    Location:
    United States
    Maybe so... keep in mind that the dictionary definition of the word "literally" has been changed to include "figuratively." No, this is not a joke. Clearly these words are complete opposites of each other, but that is the direction our language is headed. So it could be a simulator, or a knee slapper, or a fart analyzer... hard to say anymore :D
     
    phl0w and Johnnypenso like this.
  4. DGXR

    DGXR

    Messages:
    1,068
    Location:
    United States
    At 5:45
    upload_2017-11-8_12-45-8.png

    Who knows... for the needs of his presentation, this shot could be from GT5 or GT6. But watch from 6:23 through the end of the video and it shows detailed car modeling for the LP400. Pretty sure it will appear in GT Sport in the future. There will be hundreds more cars to come to this game, but hardly anyone has patience anymore :(
     
    ProjectF likes this.
  5. ShiftingGears

    ShiftingGears

    Messages:
    2,868
    Location:
    Australia
    I like GT Sport. It is fun, and usable for the jagged inputs of a controller user. But it isn't a simulator. How can you possibly claim it is when your tyre model doesn't include tyre pressures, or when all cars float over the most aggressive of kerbs, and maintain cornering speed and stability when dropping two or three or four wheels on the grass?
     
    super_gt and freedom1104 like this.
  6. Nando deBem

    Nando deBem

    Messages:
    788
    Location:
    Canada
    I don’t experience that smoothness at all.
    I went roof first into the wall after attacking that high kerb on the Dragon Trail track.
     
  7. Stotty

    Stotty Premium

    Messages:
    8,300
    Location:
    United Kingdom
    Got bored of grinding Nurb 24, so took the GT3RS to Nurb for a play.

    On SS, with toe at +0.10, LSD at 5 15 10, and front camber at -2.0, it's really quite nice, and handles broadly as you'd expect it to for the drivetrain layout. Doesn't get too lairy off the gas, but tuck the nose in nicely.

    I'll have a proper go at tuning it in the months to come, but the overall balance feels right :)
     
  8. GT6mebe

    GT6mebe

    Messages:
    916
    and there are also many that are pleased with its modelling. I wasn't really talking about feel though as thats subjective. More on the actual engineering principles and what is captured.

    An unfounded statement. You can make claims but none of them are conclusive.
     
  9. Johnnypenso

    Johnnypenso Premium

    Messages:
    28,424
    Location:
    Canada
    This makes no sense given the post you replied to. I think you have your lines crossed somewhere.
     
  10. andrewhake

    andrewhake

    Messages:
    373
    Location:
    United States
    Are you guys able to upload the entire presentation? Would love to see the whole thing!
     
  11. 7HO

    7HO

    Messages:
    789
    Location:
    Australia
    @Scaff I've been unable to come online until now and the conversation happened to long ago to bring back up. I just wanted to say yes in response to some questions you asked, I have done the things you asked if I have done in real life and yes I do have plenty of driving experience that allows me to know how real cars handle at the limit. It is part of why I have talked about the grip issues so many times because it is really clear to me what the biggest things GT Sport does wrong. But just to mention something you might not be aware of from something you said. It is very normal for a sim to have different low speed and high speed physics and for there to be a strange transition or for a sim to be odd. I'm not an expert in why but I've been told it has something to do with being unable to divide by zero and that basically normal physics calculations go to hell when a car is standing still or moving very slowly. Even iRacing has had issues in the past with their slow speed code except it went much higher than 20 and the good thing is we don't often race so slow. I'm not going to read through the whole thread now and I don't want to respond directly to posts to bring up an old argument, I wouldn't even bother replying. I just noticed one post where you replied to me where you talked to me as if you imagine I lack experience or understanding in these areas and that isn't the case.

    As to why I disappeared if you are interested, we had some storms and some damage at the inlaws and power out here and a big clean up and at the same time I had 3 trips to the hospital in one day, one for each of my kids. Then I had my own commitments on top of that so if I got a free moment I've been racing instead of wanting to come back on here since then. I still don't have so much time at the moment to come on and talk about games but I just came on here to find out some info if I can find it and saw all the notifications which I don't really have time to go through.
     
  12. Johnnypenso

    Johnnypenso Premium

    Messages:
    28,424
    Location:
    Canada
    I don't think that there are different physics at low and high speed. It's more accurate to say that what works in the physics engine at higher speeds breaks down at a lower speeds, in some games more than others.
     
  13. Scaff

    Scaff Staff Emeritus

    Messages:
    23,175
    Location:
    United Kingdom
    Far enough, but that doesn't gel with a number of statements you have made, in particular that Camber isn't used to balance tyre temps.


    Its not normal at all for sims in this day and age to do so at all, quite the opposite.

    This illustrates the point well:



    Only one of the three has an issue with coming close to how reality is, and its GTS.


    Its not to do with divide by zero errors, but rather a loss of fidelity when you get to the lower end of Pacejka tyre models curves, its also an issue that just about every tyre model apart from GTS has resolved, mainly by developing much more details tyre models.

    The iRacing issue is more a problem with how grip relates to tyre temperature and the absurd situation in which you can get more grip at lower tyre temps than you can with high tyre temps (the rather well known slow out lap exploit in qualifying is an example of this).


    No problem, the real world always trumps this place.
     
    freedom1104 and JulesDennis like this.
  14. 7HO

    7HO

    Messages:
    789
    Location:
    Australia
    I don't think that is what I said. I'm pretty sure I would have said that is what it is used for and that it is not used to add heat or take away heat from tyres primarily, only to balance heat across the tyres and then not even primarily for that but that is a method of evaluating the contact area and if the tyre is being used effectively.

    A video comparing low speed code doesn't mean what i said is incorrect. iRacing have talked about their own low speed code and have discussed why developers use different code for low speed. I'm no longer a member of iRacing so I can't search for or link to the discussions.


    Wow now you are a game developer expert too and an expert on what all developers are doing. I can only go off what developers have said themselves because I am not a game developer and what I have read from the developers is different to what you are telling me. I'll trust them.

    And not always in a positive way but Cheers.
     
  15. Scaff

    Scaff Staff Emeritus

    Messages:
    23,175
    Location:
    United Kingdom
    Please go back and take a look at what I have actually said.

    I've said (and provided supporting links to a wide range of independently verifiable sources) that tyre pressure is used as a principal tool to manage the overall temperature of tyres, that camber is used to balance the temp across the tyre, and that toe isn't used to manage tyre temp.

    The reason I said this (and provided the links to clearly show this is also how reality works) is that members in this thread have stated that pressure isn't needed, that camber isn't used to manage tyre temps and that toe is. All I note claims that have not been supported by any real world evidence of any sort.

    It doesn't mean yiou are right either, but what it does show is that for the three titles in question only one of they has an issue with it, which does go against the claim that its a common issue among sims.


    Now aside from you logical fallacy immediate meaning we can immediately discount what you have also said (an appeal to authority cuts both ways), you seem to not know that Pacejka tyre models curves are not just used by sim race title developers.

    So yes I do have a reasonable degree of experience working with them, certainly enough to know what the limitations are and how they manifest themselves.

    However as you like your appeals to authority....

    http://www.edy.es/dev/2011/12/facts-and-myths-on-the-pacejka-curves/
    http://www.mate.tue.nl/mate/pdfs/8726.pdf

    .....no divide by zero issues.
     
  16. Griffith500

    Griffith500

    Messages:
    8,532
    I did wonder why no-one had re-formulated the Magic Formula to avoid certain issues, but I guess plenty of people did "in secret". Perhaps it is old-fashioned of GT to still be switching models at different speeds - whatever those models actually are.

    Still, the fact that the Magic Formula only works for the data measured makes it very unsuitable for a sim with varying conditions, geometry, speeds etc. without having all of that data that you can somehow interpolate furiously between (which itself is fraught without a physical basis). Playing with the coefficients willy nilly when you can't obtain real coefficients clearly isn't going to help, either.

    I personally would rather sacrifice some of the precision of the Magic Formula, in respect of its steady state behaviour (and then only in the same conditions as measured for), for something a little more adaptable and naturally expressive in what is supposed to be an interactive experience, not an engineering evaluation tool.

    Most of the issues with the dynamics of tyre modeling has to do with the contact patch shape, its pressure distribution and hence friction. This is why things like tyre pressures, camber and the like are so tricky to pin down - it just isn't taken into account, except "empirically". Again, what do you do with that data if it doesn't fit what's going on in the game?

    E.g. you can have situations where the dynamic camber might be so many degrees, with combined slip numbers, fine, but the contact patch shape might not match that at steady state for those measured inputs, e.g. because of suspension dynamics or track surface irregularities, or because you have only just started turning in or braking etc. etc. etc.
     
  17. Scaff

    Scaff Staff Emeritus

    Messages:
    23,175
    Location:
    United Kingdom
    To be honest the vast majority of Sims have moved beyond it and developed tyre models that account for far more values and areas of tyre performance.

    At the end of the day most are likely to still have the formula behind the magic curves as part of it (the curve is just a visual representation of how the formula can be used for a set data set), with a lot more feeding into it.

    Keep in mind that most who have done so didn't get it right first time either, but have progressed to the point that they do now have robust tyre models that are accurate enough in a very wide range of situations and surface types.
     
    Johnnypenso and Griffith500 like this.
  18. 7HO

    7HO

    Messages:
    789
    Location:
    Australia
    @Scaff I swear if you were not a mod I'd have you on ignore because you are such a frustrating person to talk to. I'm not sure if you just scan or you just lack comprehension in general and I don't pay attention to you to see if you have the same issue with others but I have noticed you are one of the most difficult people I have ever talked to and I'm over it.
     
    SKAFISKAFNJAK likes this.
  19. NICKname

    NICKname

    Messages:
    780
    Location:
    Netherlands
    Gentlemen, a short view back to the past...
     
  20. Scaff

    Scaff Staff Emeritus

    Messages:
    23,175
    Location:
    United Kingdom
    OK

    Not an acknowledgement that you are right, simply that you don't seem to want to discuss in a reasonable manner.
     
    Last edited: Nov 29, 2017
    ImaRobot likes this.
  21. Johnnypenso

    Johnnypenso Premium

    Messages:
    28,424
    Location:
    Canada
    I don't have any trouble understanding Scaff at all. In fact, I find him one of the most logical and easy to understand posters on GTP. Your post on the other hand...:confused:.
     
  22. ImaRobot

    ImaRobot Premium

    Messages:
    12,387
    Location:
    United States
    Can't say I've had any trouble whatsoever interacting with him, and I've seen him respond thoroughly to each and every point brought up to him. So I really have no idea where you're coming from.
     
    Johnnypenso likes this.
  23. ALB123

    ALB123

    Messages:
    5,162
    Location:
    United States
    Personally, I've found Scaff to be very good at dumbing down certain complex subjects so plebs like me can understand the subject better. To me, that's a good communicator. Plus, you have to really, really know what you're talking about before you can translate highly technical subjects into things common folk can understand and recognize as true in the real world. Neil DeGrasse Tyson, anyone?