Nashville Street Circuit (WIP thread)

[Tjerk]

The track looks great. I adore how you can see both parts of the track on the bridge sections.

The only minor criticism I would have is that I find the skid marks on the track too wide. I would prefer them to be only a single tyre in width.

[checkpoint10]

The only minor criticism I would have is that I find the skid marks on the track too wide. I would prefer them to be only a single tyre in width.

Not a problem, thanks for the comment. With my tool I can quickly adjust the number, length, width and placement of these random skid marks. How does this look?

[Tjerk]

I think that looks great!

[checkpoint10]

Just curious, have you tested this with a more-detailed carset like the ones I make?  I know some people have reported issues with tracks crashing only with specific carsets and I don't think that we've nailed that down yet.  The only reason I ask is I really think we are getting close to pushing the limits of what we can accomplish (with respect to details on tracks just flowing over the brim of a cup like this)

A few versions ago when I was doing my preliminary AI tests, I used nooRok's 2020 carset and I don't think it crashed except when I didn't have the correct number of LP records or something, or if some value in nash.txt was wrong. By the way, ultimately I don't remember how many records made it work, but I am attaching the SG file and full information output from TRK23D for your reference.

[samsepi0l]

I downloaded this today, so I can drive it and get smooth/fast enough to make LP files. It's an amazing job you did. It's frustrating to set up for, because the slow sections are so slow that I really have to tip-toe around. These make me want to soften up the car. Going across the first bridge at the end the slight right bend makes it want a stiff setup to keep the car tight on the line so you can brake late going into the left hander at the end of the first bridge. My best time is 75.6s. Good thing I committed the track mostly to memory from making the old lp files for the original version. I will keep driving this and get smooth pretty soon so I can do some good replays.

[samsepi0l]

https://www.youtube.com/watch?v=hRXn4S6bDu4

74.2 - I still have work to do. I need to be smoother in the tight sections. Car skids too much.

[checkpoint10]

Great stuff. 1:14.x is way faster than what I've ever achieved. Real world pole time was 1:16.3 so you are already faster (of course the ICR2 track is smoother than in real life and the back section might be slightly faster for the benefit of the AI).

[jaeger]

I'm consistently doing high 1:11s.
High 1:10s are possible.

[samsepi0l]

Just curious, have you tested this with a more-detailed carset like the ones I make?  I know some people have reported issues with tracks crashing only with specific carsets and I don't think that we've nailed that down yet.  The only reason I ask is I really think we are getting close to pushing the limits of what we can accomplish (with respect to details on tracks just flowing over the brim of a cup like this)

A few versions ago when I was doing my preliminary AI tests, I used nooRok's 2020 carset and I don't think it crashed except when I didn't have the correct number of LP records or something, or if some value in nash.txt was wrong. By the way, ultimately I don't remember how many records made it work, but I am attaching the SG file and full information output from TRK23D for your reference.

Once I believe you made a program that would give me the corner radii for each section in a csv file... do you have the ability to do this again? I'm not sure if you may have made a program to do this. It will save me a ton of time. I need to use the dennis formula to correct the speeds and I need the radii.

I think the track will need 1010 or 1009 records. I accomplished a 73.577 as my best time but I think I have a smoother line now that will be better for replays and LPs.

[Dennis]

Once I believe you made a program that would give me the corner radii for each section in a csv file... do you have the ability to do this again?  I'm not sure if you may have made a program to do this.  It will save me a ton of time.  I need to use the dennis formula to correct the speeds and I need the radii.

I think the track will need 1010 or 1009 records.  I accomplished a 73.577 as my best time but I think I have a smoother line now that will be better for replays and LPs.

You can get the correct record number quite easy. Just create LP files with a higher number of records, and if the game crashes, you know, it was too high, and try it with a lower number.

Probably the speed formula need to be adjusted, because the first chicane is in the SG file a straight (am I right?). At the moment, the speed values won't be corrected on the straights. If it's a problem, let me know, and I'll make an improved formula, because I expect too low speed values there (also if there is a corner in the SG file). But probably you'll need to make the speed values in the chicane by yourself, because if you don't get relevant values with trafo20, there won't be anything to correct with a formula.