Xtreme Gaming Network

ACC News

Author Topic: ACC News  (Read 4560 times)

Online Wally

  • Administrator
  • Hero Member
  • *****
  • Posts: 8480
  • Karma: 130
  • AC Admin
    • View Profile
ACC News
« on: August 01, 2018, 07:55:08 AM »
A new thread for ACC...

Not strictly ACC related (great start to the thread, I know), but here are some interesting comments from KS's Aristotelis at the Spa 24H event where ACC was set up for real drivers and sim racers to try (from https://www.facebook.com/aristotelis.vasilakos/posts/10156639993558069?notif_id=1533033161405304&notif_t=mention)

Quote
The Spa24Hours event was particularly important because it allowed to watch real drivers and alien simracers really going for it, especially alien simracers as they also had some big prizes at stake. Here are some of my thoughts.

So what did we learned? Some pretty important things.
1) First of all everybody was driving in the same way. If you could put the real drivers and aliens side by side and watch their movements carefully, you wouldn't notice any actual differences. Same movements, same lines, same attack style, same braking points... the lot. At the end of the lap there would be 1 sec more or less difference but it would be down to the aliens having more practice and being able to gain half a tenth at every turn. Being a 7km long circuit that was the difference. We could even watch the in turn speeds be like 1km differences lol. Tiny tiny stuff.

2) pro drivers in real life have nowadays extremely low practice time. Most of them would get 5 to 15 laps between practice and qualifying. When the race engineer see in the telemetry that they are "on target" then the call them in the radio and..." good job, pit now and let the car to the amateur gentleman driver". So their job is to be able to push the car close to the limit with that amount of tiny practice. It was impressive to see them do the same in our rigs. 4 laps and they would be instantly on their real time laps. 10 laps and they would start to go down towards alien laptimes.

3) Alien simracers also did the same initial laptimes as real drivers. We had nobody coming in, sitting down and posting 18s. Everybody would do 1st lap in the 20s, some in the very high 19s. then practice made perfect. They are fast to adapt, but not really "magical". Their strength is that when we stop at low 18s, they would continue to improve down to 17s. No "unrealistic" techniques were seen.

4). When real drivers decide to "practice" they will do 100% valid laps. Eau Rouge taken always in the same way, never putting a wheel outside.
When Simracers decide to push, they have 50% valid laps through eau rouge. I'm sure they can do also 100% but that's pressure for you and pushing to get to the 17s to win the competition. Also explains the difference with reality.

5) Our hotlap competition was set at 25°C ambient 35° asphalt which we thought realistic for Spa... During the practice and qualifying sessions, thanks to global warming, we had 35°C ambient and "catering services cooking eggs on the asphalt"°C in real life. Also I saw fish flying in the air, it was that humid. Significantly enough, when we got some rain, the temperature would stay almost identical. Race had a bit more normal temperatures. That could be another indicator on the slight difference of laptimes between aliens/real life. Still VERY small gap considering the long track and all the other factors.

6) Modern alien simracers adapt. FAST! They are proper pilots in terms of technique. They don't use "tricks" so to say. At the end of the competition, with no more pressure we gave them a surprise with a "hotlap race for a sparco gloves prize" competition on condition totally invented by me and Kevin Stuck just on the fly. It wasn't easy. Wet track, light rain, sun going down, slicks. We let them decide their own TC and ABS levels but with no practice or testing before. They handle it like pros. A single spin was all the drama that happened, then all valid laps and great car control and attention, even though you could see they were struggling on the conditions. Watch the video to get an idea.

7) Real drivers are hunting animals. I don't mean this in a small way. I mean it in a BIG way. When a simracers misses a corner, especially on a chicane, then he will sacrifice it and try to take the second corner properly to retain the car rythm and improve the line.
When a real driver misses a corner he will ATTACK anyway whatever he has infront of him. He will not even blink before going full attack on the next corner on whatever situation he is in. The mentality is "someone is behind me and he will probably try to take advantage of my error, I MUST NOT LET HIM PASS!" Even though they are in a hotlap mode, their mentality is always this. ATTACK ATTACK ATTACK!
I know, but still impressive to watch them.

8) Setup was in the safe side. The cars are a bit of a short blanket so you can't always make them perfectly balanced and neutral. Especially the Lamborghini has lot's of rear biased weight so it suffers some power understeer on its DNA. Still you can work it out with setup, but then you can get brake and turn in oversteer and nervous behaviour on high speed and kerbs. As you know we need to showcase the new game to thousands of people so the main setup was quite "safe" to avoid seeing people constantly spinning at every braking point.Obviously drivers hate it at first, but interestingly enough, after they could arrive on their times they would say that "yes there is understeer, but you can work around it". That was positive. Some of them even went so far to say that "in the end I could modify a bit my line and it wasn't an issue at all for doing what I wanted to get my laptime". I couldn't change the setup in place as the UI is still WIP. All I could change for them was TC, ABS and some brake bias, but not while on the hotlap competition.
Still, a trend feedback from all real drivers was that the car was perfect at slow and mid speed corners but lot's of push understeer at high speed flat out turns, while in reality they get a more nervous car at high speed. Part of it is in the setup as I said, but there's also something I'm working on the aero as we speak. We should be good by early access.

9) Everybody from the real drivers as well as the simracers that tried it, where highly impressed by the rain physics. Reactions where in the range from "oh wow that's amazing" up to "what have you guys done!?". Stefano Casillo should be very proud for the core physics code, I'm so lucky to be able to work with him on it. We still need some fine tuning on the whole system, but I'm confident we will deliver.

10) Believe me when I tell you, the hardware got some through extreme abuse from Wednesday to Sunday. Non stop. Rseat rigs and Fanatec wheels and pedals are an amazing combination. Everybody was extremely satisfied. I think we had a single screw to tight a bit, during the whole event, or something like that. Ak Informatica PCs and monitors, rock solid and stable.
“You can please some of the people all of the time, you can please all of the people some of the time, but you can’t please all of the people all of the time”

Online Wally

  • Administrator
  • Hero Member
  • *****
  • Posts: 8480
  • Karma: 130
  • AC Admin
    • View Profile
Re: ACC News
« Reply #1 on: August 02, 2018, 10:06:17 PM »
More information from Aristotelis about what sounds like extremely complex changes to the physics in ACC:


Quote
The tyre model has been heavily reworked. 3 dimensional flex of footprint and sidewalls with specific rates for each axis that are also influenced by the pressure and temperature. Temperatures now have even more layers and even get affected by brakes heating. Pressure heat and wear do not only change a simple grip variation but also the actual behaviour of the tyre, affecting values of slipangle, slipratio, and the grip curve as a whole. You're not only feeling a different grip but a different response too. Tyre also has a different response depending on the different surfaces it touches.

The weather simulation is not a simple lower grip, but an actual water film layer with different depth. The tyres have algorithms that simulate the drainage of the water from the tyre, in order to keep the contact patch in..."contact" with the surface. If it succeeds, you get good grip, if not, the tyre "raises" over the water film and you get aquaplaning. That is what it makes wet driving so unpredictable. You don't have to deal just with a lower grip and that's it. You might think you got grip, push a bit more and instantly get a different behaviour on the same turn a lap before.
Many more big and smaller details got implemented and this obviously makes the fine tuning of the whole system, extremely delicate, considering that the rules predict just one compound for all cars/tracks combinations.

On top of that we have a brand new aero model coordinated by Stefano Casillo and created by Fernando Barbarossa our new physics developer. Much more complex, doesn't take into account different wings as single entities anymore, but a single aerodynamic object that gets affected by flow, and becomes very sensitive in pitch and yaw situations. 1-2mm ride height difference can be felt by good drivers and makes the setup a very complicated matter.

Because of that, we had to improve our suspension model. It now includes variable rate bumpstops that I can set with extreme precision to keep the car from pitching too much. Also new dampers with proper graph curves and not just 4 values. The bumpstops are quite complex so we're trying to figure out a way to make the setup screens better designed for the users.

We also have brand new TC and ABS algorithms, again more complicated than the original AC ones which by the way were an industry first at the time. Now we control slip and slide differently and even how much the engine cuts instead of on/off situations. Those maps alone can change quite a lot the behaviour of the car.

On top of that, add all the different dynamic grip and weather simulation and you can imagine that the possible combinations and situations where something can go wrong are quite a bit. We are also working in a condition that everything can change from a week to the other because of new features added and this means we might need to re do the physics of the cars.
Even BoP is different for each car and each track, and this is getting updated by SRO as we move from one track to the other...

So, I don't know where the 1.5 months per car comes from, as we keep working on all cars at the same time, so I can't really give an accurate estimation of how much time we spent on each car, surely Marco has this under control. But I can tell you that it has become extremely complex...
“You can please some of the people all of the time, you can please all of the people some of the time, but you can’t please all of the people all of the time”

Online GzeroD

  • RestrictedRacing
  • Full Member
  • *
  • Posts: 832
  • Karma: 46
  • ∞+1 Seanus
    • View Profile
Re: ACC News
« Reply #2 on: August 02, 2018, 10:28:04 PM »
I don't understand the bit about 'bumpstops'. Does it mean something else in Italian?
Racecar is racecar spelt backwards. So is ffok cuF.

Offline Guybrush Threepwood

  • Hero Member
  • *****
  • Posts: 2028
  • Karma: 67
    • View Profile
Re: ACC News
« Reply #3 on: August 03, 2018, 04:41:32 AM »
I don't understand the bit about 'bumpstops'. Does it mean something else in Italian?

They are as they sound.  Think of them as an extra set of firm mini springs right at the end of travel of the regular springs to stop a car bottoming out.  This way you could have softer springs or lower ride height without bottoming out and spinning off across the curbs.

I'm pretty sure AC has these in the more complex cars setup wise.

I think the variable rate part refers to how stiff they are based on compression rate of the spring.  Same of the absorbers.
« Last Edit: August 03, 2018, 04:46:10 AM by Guybrush Threepwood »

Online Wally

  • Administrator
  • Hero Member
  • *****
  • Posts: 8480
  • Karma: 130
  • AC Admin
    • View Profile
Re: ACC News
« Reply #4 on: August 03, 2018, 07:35:53 AM »
That's right - some AC cars have bumpstops, but I understand it's a fairly simplistic implementation. From memory, you can set a bumpstop thickness in some cars' set ups.
“You can please some of the people all of the time, you can please all of the people some of the time, but you can’t please all of the people all of the time”

Online GzeroD

  • RestrictedRacing
  • Full Member
  • *
  • Posts: 832
  • Karma: 46
  • ∞+1 Seanus
    • View Profile
Re: ACC News
« Reply #5 on: August 03, 2018, 10:04:15 AM »
I don't understand the bit about 'bumpstops'. Does it mean something else in Italian?
Sorry I was too brief. And sorry guys it's one of those pseudo philosophical 'don't understand's :o
I understand what a bump stop is, what I mean is why on earth do race cars get anywhere near their bumpstops. >:(
Surely you design sufficient rising rate into the system so as you approach the end of travel you also approach infinite stiffness. Or at least stiff enough to prevent ever reaching end of travel.

 >:( FFS why doesn't every car have active suspension?
Racecar is racecar spelt backwards. So is ffok cuF.

Online Bacchulum

  • RestrictedRacing
  • Hero Member
  • *
  • Posts: 2460
  • Karma: 72
    • View Profile
Re: ACC News
« Reply #6 on: August 03, 2018, 10:29:29 AM »
Cost.
Rising rate springs are far more expensive than a standard spring + bump stop.
As far as designing sufficient rising rate into the system, these are cars designed for rich people not racing, then converted. :-X

Online Wally

  • Administrator
  • Hero Member
  • *****
  • Posts: 8480
  • Karma: 130
  • AC Admin
    • View Profile
Re: ACC News
« Reply #7 on: August 21, 2018, 08:11:45 AM »
“You can please some of the people all of the time, you can please all of the people some of the time, but you can’t please all of the people all of the time”

Offline Guybrush Threepwood

  • Hero Member
  • *****
  • Posts: 2028
  • Karma: 67
    • View Profile
Re: ACC News
« Reply #8 on: August 21, 2018, 08:44:40 PM »
https://m.facebook.com/story.php?story_fbid=467103350433405&id=378582002534902&refsrc=http%3A%2F%2Fwww.assettocorsa.net%2Fforum%2Findex.php&_rdr


Check out the night time reflections and drying line at 28:40.

Looks good.  Although there didn't seem to be much of a difference in grip between the dry line and the wet line - but only so much you can tell from that kind of footage.

Online Gratulin

  • Hero Member
  • *****
  • Posts: 2119
  • Karma: 44
    • View Profile

Online Joe

  • Hero Member
  • *****
  • Posts: 1763
  • Karma: 28
    • View Profile
Re: ACC News
« Reply #10 on: August 22, 2018, 08:33:54 AM »
Here’s the ACC graphics card  8)

https://www.pcauthority.com.au/feature/nvidias-new-geforce-rtx-2080-everything-you-need-to-know-500621?utm_source=feed&utm_medium=rss&utm_campaign=PC+%26+Tech+Authority+

Waiting on benchmarks but does look promising. Rumors are a fair bit more powerful than the 1080TI and at 1200 bucks.

Online Wally

  • Administrator
  • Hero Member
  • *****
  • Posts: 8480
  • Karma: 130
  • AC Admin
    • View Profile
Re: ACC News
« Reply #11 on: August 22, 2018, 09:10:44 AM »
That's too expensive for me. I'll wait to see how ACC performs on a 1080. But triple screens with eye candy might need more grunt. We'll have to wait and see.
“You can please some of the people all of the time, you can please all of the people some of the time, but you can’t please all of the people all of the time”

Online buellersdayoff

  • RestrictedRacing
  • Full Member
  • *
  • Posts: 500
  • Karma: 17
    • View Profile
Re: ACC News
« Reply #12 on: August 22, 2018, 03:23:11 PM »
2080ti $1900 in Australia, ridiculous. No way...

Online buellersdayoff

  • RestrictedRacing
  • Full Member
  • *
  • Posts: 500
  • Karma: 17
    • View Profile
Re: ACC News
« Reply #13 on: August 23, 2018, 06:36:42 AM »

Online buellersdayoff

  • RestrictedRacing
  • Full Member
  • *
  • Posts: 500
  • Karma: 17
    • View Profile
Re: ACC News
« Reply #14 on: August 23, 2018, 07:25:02 AM »

with a 1080Ti currently, I'm not interested in these new ones, too expensive for not a huge bump in performance.