Watch Andy Tudor Demonstrate Project CARS 2 At PC Gamer Weekender

Slightly Mad Studios’ Andy Tudor gave a big presentation at PC Gamer Weekender, introducing the crowd to the many new features of the upcoming WMD-powered simulation.

A full video of his presentation is now available on Youtube as the Creative Director behind the upcoming detail goes in-depth on the title’s new features as well as answering questions from the crowd afterwards.

Project CARS 2 will bring 170 cars and 60 tracks to the PC, Playstation 4 & Xbox One.

Powered by the studios’ Live Track 3.0 technology that allows dynamic surface conditions and the title’s comprehensive dynamic weather engine, Project CARS 2 is able to simulate racing in any weather & season, including heavy winter weather & snow.

Aside from graphics & sound improvements, Project CARS 2 comes with refined physics featuring the title’s cutting-edge tire model and improved AI as well as brand-new gamepad controls & FFB improvements. The new version comes with robust eSports & online capabilities including Online Championship mode, driver rating as well as race directing & broadcast controls.

Project CARS 2 is due to be released in late 2017, more info on what’s new in the second version of the WMD-backed title can be found here.

GTOmegaRacing.com

  • svenvangent

    Landmines invisibly walls grinding of the steering wheel and crash to the main screen never been solved on ps4 , hopefully I don’t encounter them on Project cars 2 for playstation 4 .

    • Steven Shears

      Yeah that was always an issue – NFS had it as well. The Miami track used to have an awful inviso’ wall blocker on the start/finish straight but the conditions had to be right for it to work (or not!). IIRC as the graphics engine was part ported over, the problem never went away although i didn’t notice it much in pCARS.

  • Markus Ott

    I don’t know if they really fixed all the problems of pCARS1, but I can tell I like them pushing even further of what should be normal for a modern sim. Especially stream integegration and that stuff. SMS knows what sim racers want and the sim genre needs to grow. I only hope the most important part of pCARS2, of any racing game, follows that level commitment: the handling of the cars and FFB.

  • melanieuk1

    That low speed spin out that plague’s every single racing sim that uses the ISI pmotor engine, they have fixed, thank god for that. I’ve watched many videos and spectate many people play Race Room and Project Cars, and have watched them spin out when taking corners at low speeds or just spin out for no reason.
    Very very annoying bug, and leaves players wondering what the hell happened, great Slightly Mad Studios has burnt that bug with fighah.

    • brrupsz

      It’s not a bug, merely a correlation of things. Firstly, data put into the tire model, which is too unforgiving for most of the compounds (I’m talking about ISI engine based games however, it’s not a bug of the model itself, just a different sets of data). Then there are lags, coming from slow, 60hz LCD panels, which you don’t really see or feel in sim but affect your reactions. Lastly FFB wheels, where majority of simracers have either too slow motors (G2x, T150) or wrong settings overall. Not to mention consoles, where FPS cap and hardware lag is another big problem.

      Plus there is something like development process. Game dev rarely has enough time to completely test various tire data inputs and pick the correct ones, because models are very complex. Mostly companies will just build their stuff based strictly on the documentation supplied by module / engine coder.

    • Marc Collins

      Where in the video is this issue discussed? Or is this your usual beating of dead horses? As brrupsz mentions below, it’s not a bug in the model. Plenty of cars in AMS and rF2 behave as they should. Over the years, plenty have not, so blame the devs of those cars. PCars cars are not all consistent quality, either, nor are iRacing cars. Same issue.

      • melanieuk1

        Jog on, it’s mentioned in the video, couldn’t careless what brrupsz says, fact IS all games that use the ISI pmotor physics engine, suffer this low speed spin out and loose of force feedback for no reason, even the developer themselves when they first showed rain back 2009 in rfactor 2, he struggled to keep the car on the road, no matter what you say it’s not going to change my views until I see a change and try it myself.

      • Marc Collins

        “Fact is,” you are wrong. It’s car-based as we have said. Aside from all your typos, let’s just agree to disagree about the programming, but agree that your comment is gratuitous here. Why don’t you tell us all the facts about precisely what PCars2 will be like, since you have likely made up your mind about it. It will still be your largely fact-free rantings, but at least on topic.

  • HardRock

    11:57 That is just pure, unadulterated sexyness. Should be age restricted content really.

  • TripleJay_X

    the LiveTrack 3.0 and Season Changes sounds great.

    so could i drive from spring to winter in one race-session? 🙂

    • HardRock

      365 days of Le Mans? 🙂

      • TripleJay_X

        …in an actual one-hour race ^^

  • Marc Collins

    Very exciting. Console racer getting closer and closer to real sim and of course surpassing them in many ways related to graphics and game play.

  • Leper Messiah

    I’m really glad about the evolution of the weather system. Pcars was visually amazing but the lack of drying line and on off wetness of track is not so good. Addressing that in PC2 as Andy says will truly make it more authentic and interesting, especially if cooling down wets/inters on a drying track by seeking out water off line is a thing (please let it be in the game!). I’m praying for properly random and not scripted weather in career as well.

  • 5hitm4k3r666

    So as Rene Rast is helping to develop the game, I hope he does it without the lag that I have seen in his last video playing the game. Alot of the stuff is the usual marketing talk. Live Track 3.0 is the only thing that I find standing out, the rest is very vague and in some cases just not true.

    • HardRock

      Which part was not true?

      • 5hitm4k3r666

        To get perfect replication of tracks using photogrammetry and drone laserscanning. That’s not possible with current technology and especialy not with drones. It’s a technique that is pretty much at it’s very beginning of it’s livespan and especialy areal land survey has a very high chance of height meassuring errors. Next to the fact that differential GPS offer an accuracy of 1-3m at best you get an idea how far off meassurments can be using that technique. Height accuracy gets more and more of an issue the more height variations you have at the track. It’s a sum of different aspects that don’t sound plausible at all. Terrestrial laser scanning is still the way to go if you ask me. Not sure if they did that for all tracks, as it is very expensive.

      • HardRock

        I don’t know the exact details of the technique, but I don’t see how GPS has any bearing on it or how the 1-3m inaccuracy could be real. Surely, if that would be the case it would totally invalidate it as a surface scanning technique.

        As I understand it the point of it is that it is cheaper than laser scanning and more accurate than making circuits by photos and by eye, so while it’s not as good the best techniques out there, it can be applied to all the tracks of the game where laser scanning was not feasible, raising the general quality of the tracks. Seems like a win to me.

      • 5hitm4k3r666

        The problem is that you need the GPS to use the drones for the georeferencing of the images. I have stitched together satelite photos myself during a photogrammetry course at university last year, so I have an idea how accurate those methods are. They are good for bigger areas, but not for high detail applications like track modeling. A satelite is not prone to stuff like wind, so you get an idea how accurate a drone influenced by wind and dependend on GPS tracking is. Is it the future? Sure, but terrestrial laser scanning or CAD data is still more accurate. The big issue I have is, that they give zero information of how they survey the tracks or the advantages or disadvantages, yet they have no problem to call it cutting edge, even if it isn’t the case. Marketing at it’s best …

      • HardRock

        So tell if I’m really off here, because I’m only going by what little I know of this technique from videos and a bit of reading.

        As far as I know, by taking a lot of pictures the accuracy of the final point cloud can be improved a lot. I mean when people can create life-like recreations of 3D objects just from a couple of pictures taken with their iPhone’s I can only imagine what a drone could do with many hours spent in the air, at various heights, with a professional camera. The small example shown at 11:30 looks pretty convincing to me. Of course, as with laser scanning the final data will have to be tweaked in places (using the other references you mentioned when they are available), as errors do creep in, neither techniques are perfect.

        As for the stitching part, that should be done automatically these days and for sure, the drone will be affected by the environment, but with enough pictures as reference it’s still quite possible to algorithmically figure out how the terrain or object in question looked originally.

        Regarding the GPS coordinates, I’m still not getting how they affect the modeling of the track at all. SMS is building a localized model here, trying to get the relative positions of the track and surrounding objects right. They are not trying to place everything on an Earth sized globe in its exact real position. The GPS data is only used for fetching the local weather when needed I think, and for that the main limiting factor is the amount of weather stations nearby and the accuracy and the availability of satellite data.

      • 5hitm4k3r666

        Don’t confuse modelling a single object like a rock via photogrammetric tools with building a whole terrain model with drones, planes or satellites. For the rock for example it doesn’t matter if the object is georeferenced. We did that aswell using a DSLR camera. We basicly modelled a wooden crate just by taking 6 photos of it and it was pretty amazing how fast it works and how accurate it is even for a noob. For single assets and especialy for game design it’s a great method – I think they used it for BF1 extensively. But don’t make the mistake to expect better results with more photos. That’s not how it works. Best is to use the minimum possible amount of photos to model an object.

        About your question of the drone technique: the reason why you need the GPS tracking of the drone is to make sure that all of the pictures that you stitch together have the same scale. In theory you could do this by allways flying at the same height with the drone or adjusting the objective automaticly but that’s difficult with stuff like wind or over terrain with big variations. I dare to say that it gets allmost impossible for tracks like Laguna and the result will be much more an approximation than a perfect replication unless your track is flat as a pancake. What they try to create is an orthophoto, wich indeed is done automaticly, but it is not very accurate and leads to some funny results as you can see with Google Earth. Best example is the Golden Gate Bridge 😀

      • HardRock

        Gotcha, thanks for the detailed explanation. Given all that you said, I’m really curious how the final tracks will look, although again, I’m sure they used multiple methods to try to correct for the various shortcomings of each one.

        The one area where accuracy will most suffer without laser scanning will probably be the track surface. This is the one key thing I’m really looking forward to comparing to the laser scanned tracks, because the feel of the track surface was noticeably more detailed in pCARS 1 on tracks that were laser scanned compared to those that weren’t.

      • 5hitm4k3r666

        I am convinced that the finnished tracks will look superb as they did in PCars 1. Got to say though, that laser scanning is a nice help, but it doesn’t mean that your tracks will be more accurate at the end nor that you will feel more. If you consider how dense the physical meshes are in rF2 for example, you need to remember that we aren’t driving on mm accurate pointclouds. Alot of it is placebo and we aren’t that far with current hardware to get perfect circuit replications that let us feel every bump. Tbh, I don’t feel a big difference between the laser scanned version of AC or rF2. At the end your physics engine needs to handle it and your FFB needs to transmit the signal and that’s something where I had big problems in PCars. Laser scanning is just one method of doing it right and not the only true method. 🙂

      • HardRock

        Your last sentence is very true. I forgot the possibility that the laser scanned tracks I perceived to be more accurate in minute track detail could simple have been better modeled anyway, or maybe it’s just that the track itself simply has more detail. The latter is a very real possibility, because the track I’m talking about is Oulton Park, it’s just a pleasure to drive on it in pCARS in terms of FFB (and every other way as well).

  • Patrik Marek

    it’s always easy to laugh at others when you never created anything yourself

    pcars2 will definitely be nicely improved over pcars1, just like other developers take experience from their past projects into new ones

    I’m sure people would agree that rFactor and other isi based games are vastly different to first games released in first generation of gmotor etc.

    SMS is quite ballsy and I really wish them all the best

Back to top