Mapping "Zahara de la Sierra" (new GT6 track)

  • Thread starter Amac500
  • 2,570 comments
  • 273,450 views
Status
Not open for further replies.
What is REALLY needed for the GT6 Course Maker is ability to make point-to-point stages.
Through woods over hills, no less.

I keep getting the feeling that the news from E3 is going to be rather momentous. I have to wonder though if it isn't "GT7 Prologue is coming!! AndsomeofthepromisedfeaturesforGT6therestofwhichwillbebetterinGT7." ;)
 
Just wondering, why do they keep downgrading the size of the track creator map? I was planning on making a track of my commute and back (about 30 miles to and 30 miles back). Well it looks like my hopes are destroyed. It shouldn't be that hard to make a track map. The Codemasters Fuel game had a map nearly the same size of the state of Connecticut. If they could do it several years ago, why can't PD?
 
Just wondering, why do they keep downgrading the size of the track creator map? I was planning on making a track of my commute and back (about 30 miles to and 30 miles back). Well it looks like my hopes are destroyed. It shouldn't be that hard to make a track map. The Codemasters Fuel game had a map nearly the same size of the state of Connecticut. If they could do it several years ago, why can't PD?
Probably something to do with how intensive GT6 is on the system's resources. Having a static map that never changes, and also is never loaded all at once, is less intensive on the system.
 
It's possible that the Course Maker maps just aren't as efficient as premade maps. I recall that Imari lamented pop-in on these courses, though this is essentially happening on objects miles away, like treelines. At least that has been my experience in my Course Maker tracks, and I have a big five mile-plus one.

I have also noticed on this monster track of mine that there is frame stutter around a certain turn that doesn't have an inordinate amount of trackside clutter, and it happens every time whether there are other bot cars there or not. Maybe it's a loading "seam" point, just a guess. So I'm thinking that at least in Course Maker 1, the game is having to work a little more with these generated assets, and PD is working to bring Course Maker 2 to performance parity with the prebaked tracks.
 
If I could make my own track maker, it would be like this;

You could add corners to the track, either left or right corners and you could choose what type of corner it is. I.e. a hairpin, a sweeper, a switchback, a chicane and all the like. And within that corner type, there will be a subset menu for you to choose from like Chicane-A or Hairpin-M so on and so forth. To add to that you can choose the elevation and width of individual corners, the bank angle, the corner radius and the distance between the corner to the next corner (which would be a straight road). A long gentle bend would still count as a corner.

With this kind of tool, we can basically recreate any track or road in real life. So basically, the possibilities would be endless. Couple this sort of maker with community sharing features and you'll have the best community based racing game possible on console. A man can dream so I'll keep on dreaming :)
 
Probably something to do with how intensive GT6 is on the system's resources. Having a static map that never changes, and also is never loaded all at once, is less intensive on the system.
True to some extent, however the FUEL map had dynamic weather and day-night cycle. It was procedurally generated though, one of the first.

Just wondering, why do they keep downgrading the size of the track creator map? I was planning on making a track of my commute and back (about 30 miles to and 30 miles back). Well it looks like my hopes are destroyed. It shouldn't be that hard to make a track map. The Codemasters Fuel game had a map nearly the same size of the state of Connecticut. If they could do it several years ago, why can't PD?
Asobo made that game a long time ago, and like NOSwaster says, GT6 is juggling more than a few vehicles on a big map. They could have made it five times bigger actually, though they thought folks would be overwhelmed by that, so settled for the 15,000 Km² - sorry, I played that game to death back a few years ago. A great game but no way is it comparable to GT in what it's trying to do.

As for the reasons, I think they said something about not being sure they could not get an optimal version the first size running, so reduced that, which then happened again. Hopefully they have been working on solving those problems and we will get the second, or even original map sizes when it comes out.
 
What's significant is that point-to-points do feel different to circuits. I have no idea why, but they've always been my favourite scenarios in car games.

I grew up on point-to-point racers: Outrun, Rad Mobile, Super Hang-On, Rad Racer, Sega Super GT, Cruise'N USA etc. Definitely a different feel to circuits.
 
If I could make my own track maker, it would be like this;

You could add corners to the track, either left or right corners and you could choose what type of corner it is. I.e. a hairpin, a sweeper, a switchback, a chicane and all the like. And within that corner type, there will be a subset menu for you to choose from like Chicane-A or Hairpin-M so on and so forth. To add to that you can choose the elevation and width of individual corners, the bank angle, the corner radius and the distance between the corner to the next corner (which would be a straight road). A long gentle bend would still count as a corner.

With this kind of tool, we can basically recreate any track or road in real life. So basically, the possibilities would be endless. Couple this sort of maker with community sharing features and you'll have the best community based racing game possible on console. A man can dream so I'll keep on dreaming :)

I'd prefer a spline based track modeller like Bob's Track Builder. Might be a bit too advanced for a console but they already had the basics down in FZero X Expansion Kit for the N64 DD!


I was hoping to see some real life race tracks remade with that but people only seem to remake Rainbow Road layouts from all the Mario Kart games!
 
Last edited:
Splines are easy enough to code into the editor, it's just whether people can use them intuitively; which implies there is a requirement for some fancy controls and some way of translating that control to the language that determines the shape of the splines.

The difficulty with splines, as shown in GT5, is getting lots of variation / detail, because they tend to be very smooth - unless you use lots and lots of nodes. Lots of nodes is really a user-friendliness nightmare. Not for professionals / "hardcore" modders, but certainly for most people who play GT.

So what's needed is some way of abstracting the fine-control out for those who want to get good detail without the hardship of micromanaging, whilst still preserving full flexibility for those who can handle it. ModNation Racers was good, but it was a cartoon-style game, and probably still isn't detailed enough at any level.

The GPS import probably represents the best way to get the exact shape you want; expect people to make "synthetic" GPS traces for import into the game...
 
Never tried Modnation actually but that looks like a friendlier way of doing a spline based editor. (looks like you draw the path freeform, and the system simplifies it into a spline) I think I own the Vita version on PS Plus so I'll check it out. Gps + importing landscape from google Earth would be nice ;)

Have been trying to find some video of the Race Drivin' track editor, can't believe I used to spend so much time playing that in the early 90s, it runs at about 1fps.
 
Last edited:
So, i asked Shuhei Yoshida about the zahara track that was announced back on feb, this is what he replied.

@yosp: @Jose_Frias1 Please wait for the official source for the info. Thank you for your patience.

Im out of patience :banghead:
 
Never tried Modnation actually but that looks like a friendlier way of doing a spline based editor. (looks like you draw the path freeform, and the system simplifies it into a spline) I think I own the Vita version on PS Plus so I'll check it out. Gps + importing landscape from google Earth would be nice ;)

Have been trying to find some video of the Race Drivin' track editor, can't believe I used to spend so much time playing that in the early 90s, it runs at about 1fps.
Oh a touch screen could be perfect in so many ways (I expect someone's already said that). Rubbish for making a replica of Cadwell Park, but great for just penning out your idle musings.

That's your compromise there, surely. Let those who want to do the "real" replicas use external programs and hack them through the GPS import, and everyone else can use the "toy" interface. :)
 
Just hearing about the Course Maker brings me to tears that's it's not out yet GAAAHAhahaaaa!
t2301.gif


Damn you Kaz... :(

Some days I try to forget :(
 
It's possible that the Course Maker maps just aren't as efficient as premade maps. I recall that Imari lamented pop-in on these courses, though this is essentially happening on objects miles away, like treelines. At least that has been my experience in my Course Maker tracks, and I have a big five mile-plus one.

I have also noticed on this monster track of mine that there is frame stutter around a certain turn that doesn't have an inordinate amount of trackside clutter, and it happens every time whether there are other bot cars there or not. Maybe it's a loading "seam" point, just a guess. So I'm thinking that at least in Course Maker 1, the game is having to work a little more with these generated assets, and PD is working to bring Course Maker 2 to performance parity with the prebaked tracks.


I think the size of the track would be limited by what you could store in cache, or share amongst friends, or maybe just generate in a reasonable time-frame (assuming the terrain is generated in-game this time). Draw distance itself isn't an issue, as SSRX first demonstrated.

What is an issue is how to manage the scene detail dynamically when you don't know what's going to be in the scene. The level of detail we have in games is only possible by reducing the apparent detail of things as they become less important, be they off-centre from the "stage" being set, hidden (occluded) by other objects, or far off into the distance that their details can't actually be resolved by their pixel footprint. Reducing detail on things that can't be appreciated, or removing them outright if they can't even be seen, allows other things to be more detailed, given a finite rendering budget.

A lot of that, for static scenery, can be pre-computed for maximal efficiency - in racing games, the classical technique is to define a visibility set that is dependent on your location on the track. For tracks that are made in the game, either you need to do that intensive visibility testing before you can drive on it (could take forever) or you use something a little more flexible as you drive. There are plenty of realtime technologies for occlusion culling, distance LoD scaling and such, but they all imply a scene-dependent overhead and can be difficult to tune, and all have failure states and idiosyncrasies. Clearly, just rendering things that are a fixed distance away (as in GT5's course maker) isn't good enough.

I've thought for a long time that Ronda, for instance, is not open yet because that dynamic scene management isn't ready and the "classical" method is intractable with branching roads and everything visible from everywhere (a single loop typically has very limited locations it can view itself from, and you can usually build "tunnels"). Because of the overhead in a dynamic system, it would interfere with anything else that is squeezed into the scheduling on the SPUs (which is all done manually in code). For example, a new sound synthesis method, or a new virtual sound source selection and LoD method (more sound sources with the same memory footprint), both of which would be SPU-based and scheduling critical (buffer swapping in particular, with sounds), must be carefully integrated into any changes in the renderer / scene setup.

It's like you said: you can add these features in incrementally, but you've got to do that SPU balancing every single time, and every time you do that you have a risk of breaking something. Best to get it all working as one package: roll on Spec-II.
 
Oh a touch screen could be perfect in so many ways (I expect someone's already said that). Rubbish for making a replica of Cadwell Park, but great for just penning out your idle musings.

That's your compromise there, surely. Let those who want to do the "real" replicas use external programs and hack them through the GPS import, and everyone else can use the "toy" interface. :)

I just tried modnation Vita, the editor's not bad. I managed to draw a rough map of Suzuka with my finger and it came out pretty good, luckily it even got the bridge the right way round.
 
Through woods over hills, no less.

I keep getting the feeling that the news from E3 is going to be rather momentous. I have to wonder though if it isn't "GT7 Prologue is coming!! AndsomeofthepromisedfeaturesforGT6therestofwhichwillbebetterinGT7." ;)
If PD continues releasing updates about once a month and only uses the tenth decimal place (.1). 2.0 update will be here after the next update or in around 2 months!

My post was inaccurate sorry.
 
Last edited:
If PD continues releasing updates about once a month and only uses the tenth decimal place (.1). 2.0 update will be here after the next update or in around 2 months!

…Except it's going by the hundredths, as in 1.01, 1.02, 1.03, etc… so by your count there's actually 92 more months, almost 7 years...
 
…Except it's going by the hundredths, as in 1.01, 1.02, 1.03, etc… so by your count there's actually 92 more months, almost 7 years...
I made a big typo there :banghead: :). Mods please delete my 2.0 post rant :D.
 
Hey, it's understandable. It kind of sneaks up on you that each update for GT6 is a percent. So that means if we get a Spec 2, that's kind of the equivalent of about 90 some odd updates! :D
 
I think all will be revealed at E3, if GT7 is coming next year then they might as well keep most of the new content for that, if not than I bet a spec 2 update is in the near future.
 
Status
Not open for further replies.
Back