Forums

Full Version: Nashville Street Circuit (WIP thread)
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
(10-21-2021, 10:23 AM)samsepi0l Wrote: [ -> ]This is how I fudged it:
PIT  252000 58800723 61411954 59578188 62306778 428250 2886755 14 59218188 59578188 2107806
PIT2  336000 58800723 61411954 64101421 2036460 523000 2886755 14 59218188 59578188 2107806
By the way, I am totally unsure of this solution to split up the pit lane - this is surely not Papyrus's intended way of using PIT and PIT2 so I don't know what kind of strangeness we may see for the AI (and even stuff like pit speed limit). Let me know if you think we may need to go back to normal pits.
I think the PIT/PIT2 should be OK because for a long time people have been using PIT2 to move the players pit so it isn't the first spot.

I think the most important thing for a first step would be to get a good set of panic files.  What I did to test them, was I used the "stupid high" value LPs I made and then copied the file I wanted to test (lets say maxpanic for sake of example) and I renamed your original maxpanic to race.lp so it would drive this without the effect of any bouncing.

Here are the "crazy high value LPs" if you want to do some testing..  obviously I only zipped the 4 you need.

[attachment=1930]

Now, when I want to view one of the LPs in the game, I like to use a really slow speed.  The lap will take forever but I change the speed of the entire lap to about 30mph.

Once you go into practice, you can fast forward with dosbox to super-hyper-drive mode by holding ALT and F12.  What I do is I hold ALT, then hit F12, I keep holding alt and I release F12 and hit F12 again, it will go even faster.  This way minutes go by in seconds.  I get about 10 minutes on the laptimer and then I watch the replay.

I do this with opponents set to 1- so there is no affect from pass1 or pass2.

For many parts your FSECTIONS trick that generated the MAXPANIC/MINPANIC worked good- but it got strange in your "funky areas".  It obviously will not work in the chicane since the game thinks there is all that extra track off on the right hand side which is just escape-road to us.

This is going to take a heck of a lot of manual editing, doing things the old-school slow way.  I think we need to focus on getting the 2 panic LPs right first- test them like I described above and then once they are right move to maxrace/minrace. 

When it comes to the monkey-business chicane the PANIC files should have the car drive as close to the left/right where we want it to. It may be better to just add a few feet for max (or subtract for min) from the values that you got to work when we made the car navigate correctly.

For Maxpanic- it should NOT be driving through the pitlane (since this will put you bypassing the chicane)

One other thing- I think I need to revise a comment I made when you were building ceasars:
https://www.icr2.net/forum/showthread.ph...88#pid9288

I think the 11th value of track.txt PIT has to do with where the DLAT barrier ends where the pit-wall is, not MAXPANIC(or MINPANIC).  This makes sense now.  Imagine if your pitlane wall extended past where it should- the cars would contact it, just like we saw in your replay.

Well, I'll work on this some more later.  I've got to go catch me a racoon.  One of the little buggers is actually moving my huge 10 lb mouse traps outside and turning them over so he can eat all of the mouse poison.  Mouse poison is expensive.  So I have to catch this racoon so I don't need to keep replacing it.

Good news is I can modify and test LPs at work, because I'm just doing trial and error... I don't need a joystick. I can do this in my "free time" when I'm not working on sanair.
I used your trick to see my MAXPANIC by replacing RACE with it and using your crazy high LP files. I am encouraged to see that for the most part, even in the twisty uphill section where the DLATs are constantly changing relative to the centerline, my AI car was clinging exactly to the wall. Now it does have a problem at the chicane because the DLATs jump so quickly that the car goes off road and seems to take a while to find its way back to the line. But I wonder if that's still okay - does the game actually use MAXPANIC/MINPANIC to calculate the physics of the AI car or just to establish a boundary? If it's just for the boundary, would it even matter that DLAT goes from 70000 to 600000 within one LP record in MAXPANIC/MINPANIC?
(10-21-2021, 01:14 PM)checkpoint10 Wrote: [ -> ]does the game actually use MAXPANIC/MINPANIC to calculate the physics of the AI car or just to establish a boundary

All my experience so far tells me that maxpanic/minpanic is just for creating a barrier where the AI cars will bounce off of.  The player car is bound by the TRK file, while the AI cars are bound by minpanic/maxpanic.  Think about it- the curve it took after the chicane was wide completely passing through the barrier, that would be outside of the TRK.

Maxrace/minpanic are more for when the car gets off-line for some reason (to avoid a collision with another car, or if for some reason it "falls off" of race.lp)

Nashville has no grass- so basically speaking, maxpanic/minpanic are just about the same as maxrace/minrace.  My plan is to get a good maxpanic/minpanic that basically follows the wall (if you make it say, 30mph everywhere) and then to use a spreadsheet to DERIVE maxrace and minrace. 

Eventually I will use maxrace and minrace to derive PASS1/PASS2.  It's like building blocks, the ones underneath need to be stable in order to build up.


(10-21-2021, 01:14 PM)checkpoint10 Wrote: [ -> ]If it's just for the boundary, would it even matter that DLAT goes from 70000 to 600000 within one LP record in MAXPANIC/MINPANIC?

For maxpanic/minpanic I think NO it wouldn't matter, but for maxrace/minrace I think Yes, it would matter.


(10-21-2021, 01:14 PM)checkpoint10 Wrote: [ -> ]Now it does have a problem at the chicane because the DLATs jump so quickly that the car goes off road and seems to take a while to find its way back to the line.

I have a strange feeling about this as if we have something going on with our calculation of position.... I wonder if I sent you replays from say, tamiami.... would it determine the same positions compared to trafo?  Would this teach us something?  This may be a good thing to discuss in our TRAFO thread....
(10-21-2021, 02:04 PM)samsepi0l Wrote: [ -> ]For maxpanic/minpanic I think NO it wouldn't matter, but for maxrace/minrace I think Yes, it would matter.

I have a strange feeling about this as if we have something going on with our calculation of position.... I wonder if I sent you replays from say, tamiami.... would it determine the same positions compared to trafo?  Would this teach us something?  This may be a good thing to discuss in our TRAFO thread....

Happy to check my RPY conversions on another track and I can even try to generate MAXPANIC/MINPANIC from your TRK file.

For Nashville, I decided to go with an assumption that my MAXPANIC and MINPANIC are okay, and started just playing around with RACE.LP.

I was actually able to fix all the bouncing issues by smoothing out the speeds or slowing down the car, especially in the twisty part where the turns are off centerline and the speeds were either too fast or negative.

Check out this relatively smooth lap (although perhaps not particularly fast).



Edit: I am attaching my latest LP files
Wow- great work! I totally forgot that the speeds were still way off in parts. Even more encouraging..... really I need to do an SG2TRK and run it through the Dennis formula.
(10-21-2021, 02:28 PM)samsepi0l Wrote: [ -> ]Wow- great work! I totally forgot that the speeds were still way off in parts. Even more encouraging..... really I need to do an SG2TRK and run it through the Dennis formula.

Thanks! We are definitely making good progress. Really appreciate all the discussion and help so far.

I was wondering, for MINRACE and MAXRACE, can we just have the LP take a relatively normal line through the chicane in order to keep things single file through there?
I feel like an idiot:

I turned MINPANIC down from 30mph, to 10mph.... it was DEAD SPOT ON the edge of the track the entire track except it tries to go through the escape road in the chicane, then everything goes to hell until it gets to the corner before the bridge. 

We just need to fix the MINPANIC and MAXPANIC in the chicane... can we go into OPE and get DLAT values and DLONG values and play "connect the dots" there?

Watch how awesome this is... I did it in fast forward because it is painfully slow


It will be single file (I'm sure) but it will need to have different positions in the proper LPs and we don't want cars dissapearing through walls when you are racing tightly with someone.

Not to mention- do you plan on having me make a two row grid in the text file for 2 wide on the pace lap?  I like to race single file starts, only because it's cleaner.  I was thinking of doing 2 different track.txt lines the user can select which one they want.

Also I think the track.txt has a wrong value the car stays on the PIT.LP for a long time (until after the first bridge is passed). I will look at that.
I'm still wondering if your video really shows that we have to fix MAXPANIC and MINPANIC, because even if we got the lines correct for the chicane, wouldn't we expect the AI to bounce around anyway because they cannot cope with those quick DLAT changes? 

Here is what I believe my MAXPANIC and MINPANIC lines currently are (green for MIN, blue for MAX). I don't think you want to smooth out and connect the green line, and I don't think we can smooth out the blue line either without cutting across the grass.

[Image: attachment.php?aid=1932]

As for the pace lap... I don't know what would be best there. Probably just need to experiment a bit.
I guess in my mind the main most important thing is that maxrace/minrace go smoothly through the chicane. If we can get that then pass1/pass2 should work out fine.

If 2 cars are side by side after the S/F line, remember the car on the outside will not wait for the first car to go into the chicane. The game will probably put them over-each other if we allow the LPs to do this. That's why we want them to be as separated as possible...

There will be cars driving on minrace/maxrace when you sneak in and try and outbrake them suddenly. I'm thinking we need to test the minrace/maxrace with using the "crazy high value" LPs that way we can see how much lateral movement we really can get away with....I wonder if 20ft per record would work? I am being greedy just thinking that?
Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29