Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[MX5#98]

It's a good idea but it would be tough to implement.  First, I don't think that the GPS data would be accurate enough to map a track.  Adding accelerometer data would help but you have to be careful how the phone was mounted in the car (it couldn't just be on the seat) and the program would have to know the mechanical characteristics of the car suspension.  Also just doing a lap wouldn't be sufficient to map the track.  At the very least you would need to do a lap on the outside border of the track, a lap on the inside of the track and a lap on the racing line.  I had the same idea but using PDs in car telemetry data thingy.  All this will still only give you a rough idea of the track.  You'd need a good track editor and some good modders to refine each track.  This would be much better than how modded tracks are done now but a far cry from the accuracy of laser scanned tracks.

I can see why you would think it might be hard to implement, or inaccurate. However, those of us that use Strava regularly have been amazed at the accuracy. Granted, you don't have as much variation in speed on a bicycle, but you still have variation. And there would be nothing to prevent you from driving the route at a semi-constant speed to increase the accuracy. And, if you drive the middle of the track the whole way, it could build the track around you by making it equidistant on either side. This would work. It would not be an exact copy, but it would be very very close.

[MasterGT]

Or data from multiple drivers could be merged into a composite.

[GoesTuna11]

It's a good idea but it would be tough to implement.  First, I don't think that the GPS data would be accurate enough to map a track.  Adding accelerometer data would help but you have to be careful how the phone was mounted in the car (it couldn't just be on the seat) and the program would have to know the mechanical characteristics of the car suspension.  Also just doing a lap wouldn't be sufficient to map the track.  At the very least you would need to do a lap on the outside border of the track, a lap on the inside of the track and a lap on the racing line.  I had the same idea but using PDs in car telemetry data thingy.  All this will still only give you a rough idea of the track.  You'd need a good track editor and some good modders to refine each track.  This would be much better than how modded tracks are done now but a far cry from the accuracy of laser scanned tracks.

I can see why you would think it might be hard to implement, or inaccurate. However, those of us that use Strava regularly have been amazed at the accuracy. Granted, you don't have as much variation in speed on a bicycle, but you still have variation. And there would be nothing to prevent you from driving the route at a semi-constant speed to increase the accuracy. And, if you drive the middle of the track the whole way, it could build the track around you by making it equidistant on either side. This would work. It would not be an exact copy, but it would be very very close.

As I said, I've had the same idea using PDs technology (and wrote about it either here, the old STVM site or the deathstar) so I've thought a fair bit about it.  The accuracy that you need for Strava is much less than what you would need to construct a track and model it in a game.  The problem with driving in the middle of the track is that track width is usually not constant.  Another problem is what do you do about curbs and run off?  Going off track on an asphalt runoff is much different than going over a curb which is much different than entering a sand trap.  Then you need to worry about the scenery.  That isn't just eye candy since you often use landmarks for braking points and such.  It also adds to the atmosphere of the track.  Would the Nordschleife be the same if it was in the desert?  Finally, small imperfections like dips and bumps are hugely important for how a track drives.  You can't get camber changes by running the centre of the track.  I doubt that GPS is accurate enough to give you camber changes.  In the end the gps data can give you a huge headstart for modelling the track but it would just be a headstart.

[DudeTuna]

Strava Death

[GoesTuna11]

Or data from multiple drivers could be merged into a composite.

If done properly this could help but could also hurt.  Lets say your phone accuracy is 2m (the best case scenario from what a quick google search brings up).  Now imagine a car on the left side of the track is measured 2m to far to the left and a car on the right is measured 2m too far to the right, now your track is 4m too wide.  With a 10m error your track is 20m too wide.

You'd have to be methodical like:
Step one: car stops at a designated point to calibrate gps offset
Step two: car drives outside border of track
Step three: car drives inside border of track
Step four: car drives the racing line

Step one could be used to reduce some systematic error.

If all drives from each step were averaged together you would get a good representation of the inside border, outside border and racing line.  It would be a bit difficult though to get people to go through these steps properly.

[MX5#98]

As I said, I've had the same idea using PDs technology (and wrote about it either here, the old STVM site or the deathstar) so I've thought a fair bit about it.  The accuracy that you need for Strava is much less than what you would need to construct a track and model it in a game.  The problem with driving in the middle of the track is that track width is usually not constant.  Another problem is what do you do about curbs and run off?  Going off track on an asphalt runoff is much different than going over a curb which is much different than entering a sand trap.  Then you need to worry about the scenery.  That isn't just eye candy since you often use landmarks for braking points and such.  It also adds to the atmosphere of the track.  Would the Nordschleife be the same if it was in the desert?  Finally, small imperfections like dips and bumps are hugely important for how a track drives.  You can't get camber changes by running the centre of the track.  I doubt that GPS is accurate enough to give you camber changes.  In the end the gps data can give you a huge headstart for modelling the track but it would just be a headstart.

Absolutely spot on with all the points listed. What you'd get would be a bland and inaccurate version of the real thing. Still. Think of it this way. How much better would this be than the GT track builder? I agree that this probably won't happen. Just dreamin'.

[TunahCroonah]

This is my first post here.....but prolly not the last, LOL. 

After the tirade that I put on MX5's the "Future of GT"  (sorry,bud, but ya hit a nerve, lol), I have to admit that after seeing all of the posts and examples of this title, Project CARS looks to be a GT habit breaker. 

Albeit, we DO need to see a full release of this title, and assess the pros and cons, it appears that  the gauntlet has been thrown down and it is up to PD to rise to the challenge.  for now, we can only hope that this will be the new kid on the block that is the next generation racing sim we all are hoping to try. 

Only time will tell.....but it looks REALLY great! Thanks MX5!

[Wiz]

[AgentWD40_FL]

Thats awesome Wiz .....   

[GoesTuna11]

It's amazing how much detail like this makes a difference.  I thought Spa in GT5 was a pretty great version until I saw Spa in iRacing.  The increased detail makes it feel more real.

[LooneyTuna]

Is that a mini golf green? 

[DudeTuna]

Is that a mini golf green? 

Yes.

[Wiz]

Project CARS - NEW Ford GT40 Mk IV, 1967 (Review + Test Lap)

[MasterGT]

That track looks like one in Ferrari Challenge.

I haven't driven on many of the tracks yet. Why are they not using real track names?

Any word on when on-line will be available?

[GoesTuna11]

That track looks like one in Ferrari Challenge.

I haven't driven on many of the tracks yet. Why are they not using real track names?

Any word on when on-line will be available?

Interesting that you should say that.  I was thinking that it sounded a lot like FC.