Hey guys,
Davide and Daniel, you don't really know me but I know your voices inside out, lol
Today I listened to the latest Podcast about apps and I really enjoyed it! Thanks for spending the time and energy for creating the podcasts
Very good input, great podcast! Nothing really to "criticize" so please don't take my long text in a wrong way!.
I couldn't stop myself from taking notes after some point though and I thought I'd share them with you. I'm no developer but I think I know quite some stuff about all this.
I will throw in a very unpopular opinion though:
FFB clipping as something that "shouldn't happen" and all the fuzz about the clipping meters is one of the most unnecessary things in simracing and it's causing loads of wrong assumption and bad energy in the community.
The most simple summary would be:
Don't go above the default ffb levels from the devs (100% in every games/sims I know) and if you have a strong wheel don't go above the motor strength you can handle without risking injuries, then lower the ffb until it feels good for you.
Now the extremely long text:
I had lots of discussions with Matteo Caruso (BhZ) and Andy Dixey (Rainhamiron) who both race in VEC and who raced a lot in the clubraces. Andy went from G25 to Simucube 1 and Matteo went from DFGT to T500 and is now racing with a simucube 2 since it came out.
I myself went from an old Thrustmaster 180° belt ffb wheel (2000-2007) to G27 (2014-2019) and to a CSW 2.5, just for some reference, maybe you know these chaps
I also had a lot of Motec analysis going on, spent lots of hours with Neil (neilski) talking about Assetto Corsa physics, algorithms for the grip levels etc.
If you look at my downloads you'll find the Logitech AC LUT guide and somewhere you'll find lots of flowcharts about ffb in different sims hehe.
I'm not writing all this to brag or anything! I just don't want you to think that some random stranger is telling you what's what although he doesn't know sh*t. Sadly that's a common thing happening in simracing or in general with PC stuff.
Anyway, let's start with my notes
Clipping:
Paul J you mentioned the audio/music comparison a few times. I'm mixing your podcasts as you know so I'm quite familiar with soundwaves and clipping there. For example everytime you guys have a laugh and your mics are clipping
Just kidding, the quality made some nice leaps forward over the years! I really enjoyed listening to the last podcast
Anyway, there's a HUGE difference between audio clipping and ffb clipping.
This is due to two things:
1. Audio = sine wave, FFB = strength + direction.
Audio is a sine wave. Or at least some sort of sine wave and if you're clipping, you're flattening the curve at the ends, resulting in distorted sounds. FFB however is a constant force, containing STRENGTH and DIRECTION. There are no alternating wave forms at all. At least no waves that are creating a nice sound out of overlaying frequencies. More like the rubber of the tyre scrubbing or a kerb creating rumble if at all.
2. "Clean calculation", then direction is kept, strength is "cut off above a limit".
So ffb clipping isn't distorting the "sound" like audio clipping is. Instead FFB is first calculated from the sim and as a result you get strength and direction.
The direction gets sent to the wheel motor, while the strength gets filtered by the clipping meter before getting sent to the wheel motor.
Different perspective: Audio is "read" by the ears as a wave containing amplitude and frequency. Clipping audio changes the wave form, resulting in a different sound.
FFB is a 2 entry list containing direction and strength and is "read" by the hands/arms as direction and strength. So if you cut off the nuances in strength, the direction is still read cleanly by humans and the strength only loses accuracy but not the general information about the category of force you're at.
What I wanna say: if you get FFB clipping, you still get 100% accuracy of the DIRECTION. But the STRENGTH gets cut off.
So while you won't feel the total amount of grip, since the ffb doesn't become stronger the further you get to the limit or the more grip the aerodynamics of the car create, you will definitely feel the grip loss, when the direction of the ffb changes.
You're right though about that you can feel the grip loss a little bit earlier without clipping, as you'll feel that the ffb becomes lighter before it changes the direction.
You will also feel little details in the ffb with any amounts of clipping. But you will feel them a lot less pronounced if they are just bumps in the strength, not in direction (going over a sausage kerb at perfect 90° won't really turn the front wheels but without clipping you will get like 1° wheel turn and a 5 Nm spike in strength on a DD wheel).
There's one thing though: You won't feel understeer with too much clipping. You also won't feel pronounced understeer with low Nm wheels like a G27 as the difference between max grip and understeering isn't big enough for the human body to perceive accurately.
Good thing: understeer is pretty easy to detect by the eyes when the car won't turn tighter when you steer more.
Overall the problem of clipping is way too over exaggerated in my opinion. The devs will code their ffb to be "fine" at default. Giving you enough difference in strength and direction. AC at 100% for example is definitely okay. So is rF2 at the default settings.
With my CSW 2.5 I use 60-100% motor strength, depending on how brutal I want to feel sausage kerbs etc and how pronounced I like the difference between mechanical grip and additional aerodynamic grip or how strongly I wanna feel the "load" on the front when braking or accelerating (awesomely coded in rF2 btw and a day/night difference between the G27 and CSW!).
For AC I'm using 50% gain and 100% motor strength. That's about 8.5 Nm maximum.
At the Nordschleife I could shave off a full second through the fast right corner before the Schwalbenschwanz.
I could feel the grip a lot better, feel the aero working through this 200 kp/h corner.
With my G27 I always had to get some clipping. If I'd lower the ffb so far that I won't get clipping, I wouldn't feel the difference anyway. My muscles aren't sensitive enough to tell the difference... But I won't get other problems, like a not quick enough self-correcting wheel when the rear would start to slip. The raw resistance of the G27 would be too much compared to the low FFB force.
Here's a Screenshot of rF2 with the LMP3 mod and a multiplier of 2.5 or something.. Don't really remember it. It was a LOT though!
This is a lap of Road Atlanta. You can see the peaks being clipped off through most corners. But you still see the dynamic going up and down, it's not a flat line. I definitely felt a difference between the corners, you still see different "shapes" for each corner.
You have to put the clipping into ridiculous spheres (300% gain or something) to get indistinguishable graphs between corners.
The other thing you can see in the screenshot though: The G27 has a deadzone of about 15%. You see on the "straight" I have forces between 20-30%. This is with a min force of 13 in rF2 and a car multi beyond 2.0. So the little bumps etc were there. If I'd lower the ffb to NOT clip, I would either won't feel details on straights and slight corners at all or I would have to use such high min force that there wouldn't be a difference between little bumps and normal cornering forces.
Also the drop off during understeer would drop into the "straight line min force".
Which would feel pretty weird...
Overall I think this graph looks rather nicely balanced for a consumer wheel.
Sadly I don't have the road atlanta graphs with my CSW but I did some comparison for Silverstone!
G27 Hangar Straight and Stowe, McLaren 720s GT3, 13% min force, multi of 2.5:
CSW 2.5 Hangar Straight and Stowe, 720s, 0% min force, multi at 0.45:
Of course there's a massive difference. But do you see massive clipping? Looks still like pretty clear information for me.
Only the peaks are cut off and the difference between the bumps on the straight and the cornering ffb is massively different between the screenshots. But I could still feel the grip level very well.
You would've seen the clipping meter going red about 6x though.
To get back to the music comparison:
Too low dynamic isn't great. When everything has the same loudness, you won't get the energy kick of a loud refrain/hook when everything kicks in. Apart from clipping you can use compression, limiters, maximizers to push everything to the same loudness without clipping/distortion (or very low amounts that still sound "clean").
But if you have too much dynamic, like every raw recording has, it's way worse. You won't understand the vocals when they are quiet and then you'll feel screamed in your face in the chorus/hook.
You need to find a good dynamic range for the song that sounds well and fits.
A too high max Nm output of a ffb wheel would be like too loud speakers on concerts. Scientifically the sound is still "clean" but your ears can't process it anymore.
Same for FFB. There's no advantage in beyond the needed dynamic range or exceeding what your body can process. I for example set my FFB gain so I won't get clipping in any corner while driving cleanly. But every sausage kerb will peak straight into the clipping range and will get cut off in strength. Same for wall hits etc. No need to hurt my hands during simracing.
Now there's often the argument of indy cars that would have 30 Nm during some corners due to no powersteering and massive downforce. I have to say I really doubt that this is correctly measured...
But even if that would be correct, I doubt that the current sims are replicating such massive force differences correctly. They are not programmed for this, it's outside of what the physics engine is designed for.
This has to do with the difference between an electric motor and the caster (pneumatic and mechanical trails) in reality.
Electric motors can do strength and direction. A real car has 2 spinning disks (tyres) self aligning themselves. So you'd need 4 electric motors, 2 being connected to each other, pushing against each other and then 2x2 connected to the steering gear with 2 electric motors keeping themselves in position by force.
You see that this is pretty obscure, lol.
All I wanna say is: For replicated extreme stuff on simracing hardware, there are simply limits that can't be overcome without special coding for exactly these scenarios.
To replicate a 30Nm indycar you'd need a lot of ffb around the center too but every wheel will start to shake violently then due to the difference between a real car and simracing hardware.
This now leads to the center feel, or the lack of it compared between low friction DD wheels or wheels like the CSW 2.5 and wheels like the TS-PC, T300 etc.
The TS-PC I had for 2 weeks felt way better when driving straight due to high initial resistance of the belt which went away once the wheel started turning.
What I wanna say: driving on a perfect road in a straight line will give you a constant, massive self aligning force from the spinning tyres and caster. The steering wheel in reality is clamped to the middle and you get a force the moment you start turning in any direction.
But there's no shaking at all so it's extremely tight in reality.
Low friction ffb wheels are awesome but they just can't replicate this (yet).
With ffb motors you have 0 force in this scenario. You get a force as soon as you turn but you'll have to have some very slight dead spot. The best solution I could think of is some friction/dampening going on at 0% ffb.
In my Logitech LUTs I created "white ffb noise" for this. Normally ffb or LUTs go like this:
0 | 0
0.1 | 0.1
etc.
My LUTs go like this:
0 | 0.01
0.1 | 0.1
This creates a rumble while you're standing still. It's pretty weird at first and if you put a too high value in this first row, the wheel will randomly shake itself!
It really is "nothing" = "some ffb". No idea what it is exactly in AC but I call it "white ffb noise".
But when you nail the value, this little "jittering" when there's no ffb will pretty much generate the friction you need to fill the gap completely.
Sadly with "almost no deadzone" wheel like CSW or DD wheels, the value has to be lower than the accuracy of the LUT function in AC/ACC and it would also only work for these sims, not for the rest.. So I didn't bother much with my csw and tried to enjoy the upside about low friction.
I'm waiting for the simucube 2 to get something similar with a firmware update though. "Center friction when FFB gets below XY percent".
Aaaaanyway, sorry for the MASSIVE talk here!
Summary:
FFB clipping isn't bad. You just shouldn't get too much of it for normal cornering but you should rather get some clipping instead of too low forces to feel what's going on at all (on Logitech wheels especially).
So Davide, leaving the ffb at default is pretty okay. It's not "that hard" to set it up and defaults on consumer wheels are definitely better than a controller for anything else than hotlapping alone.
I agree though that you can set faster hotlaps on a gamepad than on a wheel that shakes like crazy without giving usable informations.
I tried for quite some time to race with my xbox one elite controller. And I could do some nice hotlapping with it (AMG GT3 top 10 in the Nordschleife Leaderboard in Pcars 1) but every time I started to do an online race, I crashed. You just can't do a dodge mid corner while being on the limit with a controller. Heck, often it's almost impossible with a wheel!
Last point: I never saw someone using very low ffb apart from some top leaderboard hotlapping stuff. Matteo is in the "SNG, singularity" esports team. Dennis Lind (cousin of F1 driver Magnussen and real GT3 Blancpain driver) is the team leader and they know a lot of other people, very fast people etc etc.
They all use ffb wheels with rather high forces. No one they know is using these very low forces.
But they aren't doing leaderboard challenges, they are doing endurance racing so they probably sacrifice some "gaming speed" for actually racing well.
I definitely see why hotlap leaderboard challengers would use low ffb! It doesn't matter if you spin or not, endless tries and definitely less tiring etc. But that's not driving a car for me. That's nailing perfection in an algorithm, nothing more.
Now that was a long talk about FFB... let's get swiftly to the other points
Pro tyres:
It's not used as much as it might should. That's sadly down to how AC's Grip algorithm works. The most important parts in the equation are PSI and camber. Temperature barely changes with any setting... maximum toe, softest pressure, maximum camber
-->> barely any influence...
Since I know this I always just go for perfect psi and perfect camber.
For PSI I use Sidekick and the "show ideal pressure as delta" setting. It will show you directly and super straight forward what you need to do.
When it shows "-1.2" for the front left tyre, you just go into the garage and give the front left minus 1 click. It doesn't even bother showing you the psi. It tells you what you need to click. Perfect for idiots like me
You do this until these 4 delta values in sidekick stay as close to 0.0 as possible. You mostly get a range like -0.7 to +0.4 psi over a full lap.
If the tyre is too cold, you set the range to be more below the ideal pressure instead of more above. If the tyre is rather too hot, you keep the pressure rather above the ideal pressure.
And ofc you set some preference for the important corners and not for straights. Only need maximum grip when actually reaching the limit.
That's the display setup for qualy. When going into the race I flip the right section of sidekick to "degradation" instead of delta-psi so I can see what to expect over the duration of the race.
I also have the standard tyre app open all the time to see in real time when I get a red or dark blue tyre to know what to expect at the next corner.
The combination of sidekick and the standard tyre app does everything that's needed. I use sidekick either way so no need to "replace" sidekick with protyres.
On the other hand I also need the temperatures as a coloured-image in my view to see hot tyres at a glimpse so I could replace the standard tyre app with protyres but honestly I find the standard app, just for what I'm using it for, easier to see at a glance.
And it's prettier in my opinion.
So that's my input on protyres. Great app, sadly not needed to go fast and consistent in AC due to "bad" temperature coding.
For camber I'm using Camber Extravaganza. It simply puts the grip algorithm into a coloured graph so you can see how close you are to the "perfect" camber in real time. When I've found the perfect camber I disable the app to not clutter my view.
Last point from me:
Motec!
You might've seen my threads but anyway, here's a thread I created because I found a lot of setup talk-time would be better spent talking about the driving itself. Best setup in the world is still not giving the same amount of progression that "correct" driving will give you.
Matteo is an unbelievably fast driver! When he took part in the clubraces he won them all easily.. It was ridiculous...
Gladly I got him as a good friend over the last years so I can benefit from his natural skill
A story I love to tell about setups vs driving:
I tried to create a perfect setup for the MX-5 Cup in AC. I used Magione. I used sidekick, camber extravaganza etc and I gained 2.5 seconds over a week. Tested the different setups every day and I was convinced that it's really good.
I asked Matteo to do some laps for me and test it. And he really liked it. Balanced, fast, stable.
He was 0.2s quicker than me and couldn't get faster with it.
He said I really nailed my best laps. We looked into Motec and yep, that was pretty much identical.
Then he took the default setup and after 10 laps he was 1.5s faster. F*ck my life guys...
We took a look in Motec again, overlayed the laps and we figured out why the "imperfect" default setup was so much quicker:
The rear camber was too low, the car almost had positive camber mid corner when starting to go back on the throttle.
This resulted in a drift, the slip angles where all over the place... But the longitudinal acceleration was higher. Basically the setup sacrificed lateral grip for longitudinal grip.
The car over-rotated at the apex, going on a tighter and shorter line while getting more power on the tarmac for increased acceleration.
The mid corner speed was a bit lower than compared with my optimized setup but the exit with the higher acceleration gained enough to be faster overall.
This showed that there isn't really a perfect setup or a perfect line. At least not something that you can create in your head or on paper. You can for sure calculate it with a supercomputer and AI learning but overall it's what we call skill and talent.
Matteo is someone who just drives and is fast. He knows with his instincts when to take a "wrong" line to gain time or how to use a "bad setup" to be faster with it.
This was when Motec became my Nr. 1 tool for progress. I luckily can simply ask Matteo to do some laps for me, compare them to my own laps and see what to do.
His teammates are even faster than him so he gets the Motec files from them.
If you guys ever need some benchmark laps, just ask
There's no more efficient way to get faster than learning how to drive like a faster driver!
And also like you guys mentioned in the Podcast: You can also create a "super fast theoretical best lap" from multiple laps. I for example was faster through 3 corners around the Nordschleife than Matteo. But he was faster through the other over 9k corners and straights, lol.
He gained another 2 tenths by analyzing my Motec data.
Here's the thread I created and a second link where I start my analysis at Post 19. Sadly I never got to continue the work after some pages.. I'm still sorry for the guys who waited or might still be waiting for it. Anyway up to the end I think it's a cool thread!
Maybe you'll enjoy a little read there
- first analysis:
Additional stuff:
Raceroom ffb flowchart by me. Maybe of interest for you guys. Did a lot of testing for each setting until I figured it out. Not 100% guaranteed that it's correct but I'm very sure about it.
HI guys, im looking to get into R3E mainly for the SRS setup but really struggling to find a half decent setup for my G27. Ive spent the whole morning googling same but nothing as of yet has given me a relatively smooth feel ,all im getting is constant clipping and unrealistic effects. Note i...
www.racedepartment.com
Or just the picture:
