01-28-2022, 06:09 AM
(07-29-2021, 04:09 PM)checkpoint10 Wrote: I don't know if we ever got to the bottom of why we have that issue in qualifying. I was hoping it was because of some problem with an LP file and maybe it would be resolved once we spend more time with them.
If we have to, I suppose we can create a 2nd SG file that creates an extra track surface only for the .TRK file. We can then keep the first SG file which we use to generate the 3D/3DO.
I think that all of this stuff is teaching us so much and every time we have a problem it helps push us forward so we do even better next time. Look at how you've progressed with Las Vegas, then Rio, Nashville and now Detroit.... doesn't Las Vegas seem like childs-play to you now?
I figured out some new things with the LPs when I did Tamiami and Sanair... now just this week I made some LPs with edit45 and trafo20 for Pavel's Portland and I think they are the best LPs I have ever made. The racing I did at Portland in my 85 season the other day was really some of the best ICR2 fun I ever had. I spent the whole race struggling. I was stuck behind Scott Brayton's rear wing literally for 10-15 laps because I just couldn't get by.. the LPs came out so good it was like I was literally racing against myself. Finally I had to do a "geronimo" braking move into turn-1 and I was able to keep him behind me.
What I really need to do is to take the time to learn your new rpy->txt tool so I can make good LPs for Nashville so we can really understand if our theories about this crash are correct. I like to think it would really streamline my LP process. When I did portland the other day I took 32 laps of replay and of the 32 at least probably half of the data from Trafo was garbage. That left me with probably a dozen good laps to extract speed and position from. Once I got good speed and position- I applied the dennis formula to them and then took the average of the result of the dennis-corrected dozen laps. When I graphed this "average"- you could see it was not correct even with the dennis formula applied, for example exiting the festival chicane you go through a right hand kink completely wide open and you could see the velocity was skewed up while the car was traveling through the corner (it should look like smooth acceleration on a straight-away).
When I modify the speed data in edit45 I spam the hell out of it with the "fill with curve" command.
I loaded the "average" file as race.lp into edit45 and started smoothing things out and correcting speeds manually. After I knew the corning speeds were right I verified 1 CC drove correctly by turning opponents to "1" in the options. I then used my excel formula's to get good pass1 and pass2 files by a formula to get me positions and speeds.
I don't know how many of you guys are playing the game with my LPs but personally for me I am really excited about how difficult and challenging the racing is when the LPs are just perfect. I know Jake commented in his sanair video about how the CCs have trouble navigating eachother and I am doing my best with that. I think they do a better job on road circuits, but Sanair is a really tough track to pass on regardless I think.
This is my new way to do LPs. Even LPs I did in say, 2019 seem really crappy to me now with what I've learned lately. I need to do this for Nashville but I need to use your tool because we can't get good data with trafo. Does your tool allow the extraction of several laps of data from a long replay? (Can I drive for say 10 laps and get 10 text files from one command?)
