I can understand how a new user might not be able to tell the difference after making these changes, but as someone who has spent a long time with rF2, I can definitely feel the difference, and it's quite the improvement imo. I'm trying to curb my enthusiasm a bit, because sometimes things can be percieved as feeling better simply because there is a change, however trying to be objective here, this is such an improvement and why oh why is it just not the default setting? Just one thing though, I had to change G920_workaround:false back to true, and then change my FFB from negative to positive. What my feelings are initially: -The deadzone with the G920 seems greatly reduced - this alone is a huge plus - something I was prepared to go pay big $$$ just to improve. It still could be better, but any improvement is very welcomed. -The steering just feels more direct and more linear. I've always felt that the rF2 tyre was VERY sensitive on initial turn in - cars like the BMW Class 1 would exaggerate this feeling. It made it harder to be smooth and I had to be very gentle on turn in otherwise I would be consistently apexing early. It could just be in my head but it seems fixed now. The tyres also feel less spongy and floaty and just more solidly connected to the wheel. -Less exaggerated bumpy feeling to the wheel, wheel less prone to julting from side to side making the car feel unstable. Does everybody know about this and I'm just late to the party? I feel like every new user should do these changes straight away. Or am I overstating how much better this is and it's a bit of the placebo effect going on? @doddynco if I can ask you, is this the most raw things can get or were these changes a bit of trial and error to suit personal preference? Thanks again for sharing the info.
AFAIK most of those JSON FFB settings are legacy parameters carried over from rF1 and do nothing (they were doing stuff in rF1 as rF1 used more of an effects-based FFB rather than a pure forces calculation thing). Or to be precise, I know at least some of those are legacy settings. Things like throttle effects on steer axis used to add effects to your wheel when pressing throttle, etc. Obviously the negative/positive factor will toggle the direction and then you have the steering torque filter and steering torque minimum, which are important, but those settings are already in the GUI. Also the rotation settings have an obvious impact if you now have a software lock set to a different value than before. The dead zone in Logitech wheels is a problem of the dual motor tech they use, so it cannot be fully overcome by the sim. The minimum force is the main thing that helps to make it feel a tad less annoying in my experience.
Wihout knowing what your previous settings were, it's hard to know exactly what fixed your steering. However, from what you've described, it sounds like you had your rotation set wrong i.e not matching the virtual car steering rotation. Green Seepent - 'Does everybody know about this and I'm just late to the party?' No they don't know, because the controller profiles are a mess - the chances of clicking 'load' on a default controller profile for your wheel and it being correct (proper rotation and no post fx) are so slim that you may as well not bother, and instead make your own profile from scratch (including editing the json). Green Serpent - 'is this the most raw things can get or were these changes a bit of trial and error to suit personal preference?' The only subjective input in these settings is "Steering torque filter":2, A small amount of filter is recommended because the unfiltered signal is slightly noisy - something that's probably not noticable on the g920. The maximum here I suggest is 4, any higher has a diminishing effect, removes ffb sharpness and adds input latency. "Steering Wheel Bump Stop Harshness":20 - kind of arbritrary and generally doesn't matter unless you're drifting. Every other setting is removing 'fog' from the steering shaft forces and is not subjective. It's quite annoying because if there's one thing RF2 does right, it's the raw ffb. Litterally all the profile needs to do is get the wheel rotation and direction to match but somehow these profiles make a meal of this and also add in post fx. As a consequence, new users generally don't actaully feel the proper physics until they come on the forum asking for help. Just last week I have a friend who just got into rf2 and had a DD Mige. He was already blown away by the physics but I asked him to send me his controller.json and then I sent him back the raw version and he couldn't believe that he thought what he had was ok and proper. He said he thought he was just a bad driver and that was why he was crashing all the time. This has happend loads of times. For example, imagine having the center spring set to 10% on a 20nm wheel by default (which is the case for some presets). This is going to be up to 2nm of fake force constantly pushing towards the steering center. It's hard to explain how wrong this feels but it just makes it feel like a worse simulator than it is. All these Fanatec DD users complaining about a bug with the steering resistance setting which shouldn't be turned on in the first place. Other than mapping buttons, there's no need for a different profile for each wheel base (the t500 is a notable exeption due the high friction). All that needs to be asked to the player is: Wheel base driver wheel rotation setting? Wheel base torque in NM? Positive or negative signal?
Still doubting this. First question, aren't most of those settings already at zero by default? I had a look at my profile and they were mostly 0, although I may have edited the defaults years ago. Second question, have you confirmed any of those settings change anything? Here is one post from years ago regarding the throttle effects on steering. As I suspected, at least those settings do nothing and are not implemented in rF2 code.
On the few profiles I've just checked, steering resistance is active in all cases and steering spring is active in some cases. Note that you need to have both saturation and the coefficient to be non-zero for there to be an effect. Throttle effects, brake effects, ect, as far as I know are inactive, but turning to zero won't hurt anything. Interestingly, the BMW M4 had additive effects for brake locking, though I don't know how this was implemented. If I remember correctly, 'brake effects on steering axis' didn't control the amplitude. Here's a list of effects/modulators that I know are active: Jolt magnitude Off-road multiplier Steering effects strength Steering resistance coefficient Steering resistance saturation Steering resistance type Steering spring coefficient Steering spring saturation Steering torque capability Steering torque extrap blend Steering torque extrap time Steering torque filter Steering torque minimum Steering torque per-vehicle mult Steering torque sensitivity Steering torque zero-speed mult I'm unsure about the next ones, but can anyone confirm that these effects are active, and perhaps an argument for why they're a good idea to use or not? Rumble strip magnitude Rumble strip pull factor Rumble strip update thresh Rumble strip wave type
IIRC, those parameters are used for fake curbs. I's possible to make a curb in a circuit without modelling it in 3d and assigning to that zone a curb material. In this case there are no real physics in the curb because it's not modeled.
When I bought my simplicity DD wheel, it also installed a rf2 controller profile. My guess is, it's the makers "suggested best" values to this. Don't know if anything is off here, but I like the feeling. Code: "Rumble strip magnitude":0, "Rumble strip magnitude#":"How strong the canned rumble strip rumble is. Range 0.0 to 1.0, 0.0 disables effect.", "Rumble strip pull factor":0, "Rumble strip pull factor#":"How strongly wheel pulls right\/left when running over a rumble strip. Suggested range: -1.5 to 1.5.", "Rumble strip update thresh":0.05, "Rumble strip update thresh#":"Amount of change required to update rumble strip effect (0.0 - 1.0)", "Rumble strip wave type":0, "Rumble strip wave type#":"Type of wave to use for vibe: 0=Sine, 1=Square, 2=Triangle, 3=Sawtooth up, 4=Sawtooth down.", "Steering effects strength":-10000, "Steering effects strength#":"-10000 to +10000, applies to all steering effects (torque, resistance, static spring, jolt, etc.)", "Steering resistance coefficient":0.1, "Steering resistance coefficient#":"Coefficient to use for steering resistance. Range: -1.0 to 1.0", "Steering resistance saturation":0.1, "Steering resistance saturation#":"Saturation value to use for steering resistance. Range: 0 - 1.0", "Steering resistance type":0, "Steering resistance type#":"0=use damping, 1=use friction", "Steering spring coefficient":0, "Steering spring coefficient#":"Static spring effect rate (-1.0 to 1.0)", "Steering spring saturation":0, "Steering spring saturation#":"Static spring effect peak force (0.0 to 1.0)", "Steering torque capability":20, "Steering torque capability#":"The maximum torque capability of the wheel (in Nm, obviously)", "Steering torque extrap blend":0, "Steering torque extrap blend#":"Higher blends of extrapolated value allows driver to feel torque changes even when actual torque exceeds 'input max' (0.0=disables, 1.0=max)", "Steering torque extrap time":0.001, "Steering torque extrap time#":"Time in seconds to extrapolate steering torque based on current change (Range: 0.001 to 0.050. To disable, set 'blend' to 0.0)", "Steering torque filter":1, "Steering torque filter#":"Number of old samples to use to filter torque from vehicle's steering column (0-32, note that higher values increase effective latency)", "Steering torque minimum":0.015, "Steering torque minimum#":"Minimum torque to apply in either direction to overcome steering wheel's 'FFB deadzone' caused by friction", "Steering torque per-vehicle mult":1, "Steering torque per-vehicle mult#":"Per-vehicle steering column torque multiplier (this is a copy of the .CCH value)", "Steering torque sensitivity":1, "Steering torque sensitivity#":"Sensitivity curve applied to representable torques: 0.0=low 1.0=linear 2.0=high", "Steering torque zero-speed mult":0.05, "Steering torque zero-speed mult#":"Multiplier at zero speed to reduce unwanted oscillation from strong static aligning torque",
With respect @stonec, I have been mucking around with min force and smoothing of various levels for like 7 years... I really feel like whatever changes occured as a result of the controller profile changes made an improvment. For example I'm now running min torque on 0, and yet I have less deadzone than ever.
Looks ok however I would turn Steering resistance saturation to 0. If you want to add damping/friction/inertia you should do it in the simplicity drivers instead. My personal experience with a SW30 simplicity wheel is that is already has more than enough in-built, non-adjustable damping than the equivalent Simucube (SC1) wheel.
I can't be totally sure of the situation because this stuff is like a different language to me, but from converstaions I've had over the last couple of weeks and now this, I'm now pretty convinced that there are post effects acting on the ffb (at least in my case), not just the curbs but during normal driving. I've got no issue at all with people wanting effects to make the ffb more communicative, but all this begs the question, why would the devs not put an option in the UI to disable all effects and just have the most raw FFB?
Other than the BMW M4 antilock brakes experiment, I've never experienced or seen any evidence to suggest that there are hidden ffb enhancement fx being used. What are you feeling and in what car? There are postprocessing fx being used, however they're not making the FFb more communicative. It's the opposite really - they're just slightly damping the signal, which is probably a cheap way to temper oscillations. In addition, on some profiles they add a centre-spring, but I have absolutely no idea why they'd want to do this.
Well you're probably right, however from what I have just subjectively experienced, the FFB before had more going on, more "noise", more vibration, more josteling about, yet somehow a bit more vague (more deadzone) and less direct. The steering now feels a bit more calmed down, a bit less going on, but more direct... more "solid" feeling - much of this is probably just from the reduced deadzone. Edit: I'd have to do a "reality check" and use the old profile again to have a more objective view of the changes.
I think this is where so many options can obfuscate things. On top of that, people don't always test things well before making statements - like the kerb parameters, which currently I'm struggling to find any evidence of being active at all (despite some being convinced they are - but being short on details when queried about it). I suspect most people are running around with standard raw FFB, with just some damping as above. And maybe @green serpent your profile file doesn't indicate any different (make a copy!) and some other quirk of the game was at play instead. Just imagine if custom FFB were possible like rF1... I think they should clean all this old stuff out, put what's left in the UI, and a simple screenshot or two will make it clear for everyone how things are configured.
Well, after switching back and forth between original profile and changed profile, I noticed something that I didn't pick up on the first time. The new profile changed the car specific multiplier of the car I was using form 100% to 130%. This is likely what made the ffb feel more solid but slightly duller probably due to clipping. After running both profiles at the same %, I can't say I can notice any significant difference now. Not saying there is no change, but I can't really tell. Sorry if I led anyone astray with my first impressions. At the very least I'm pretty confident now I want a higher Nm wheel, I've never really ran it that high before and it actually felt quite good even if there was some clipping.
@green serpent get something with 20Nm, and send us pics from hospital Even a T500 feels much better than the Logitech, one of the 6-8Nm options should be fantastic. Plus then you have to play heaps to justify the cost.
I plan on going the direct drive route. I might wait until we find out the price point of the Thrustmaster unit, or maybe see if Logitech announces something. Either way I will have to play HEAPS to justify the expense!
Those parameters are not easy to test, because only are going to work in certain kerbs. Only work with the kerbs that are not modeled in 3d. These kerbs are only painted and are referenced in the TDF The modder makes in the 3D program this But he wants that the FFB is similar to this Almost all current circuits have the kerbs made in 3d. I found some of those "non 3d kerbs" in old circuits imported fron rF1. One way to check these rumble parameters would be to unpack circuits in devmode and check the GMTs in 3DSimEd or similar to find a "non 3d kerb". And then drive on these kerbs changing the parameters.
Yeah. To be honest I don't think I'll bother, the people referencing these as canned rF2 effects do much less work than that before shooting their mouth off. In future I'll just say they're wrong
Yes, @Lazza i agree with you. Too much work for obsolete effects, because current circuits have the kerbs designed in 3D
Ok, quick update post just on the rumble strip effects, it's sort of off-topic here but when various json entries get discussed I suppose it's still relevant. The controller.json has this section: Code: "Rumble strip magnitude":0, "Rumble strip magnitude#":"How strong the canned rumble strip rumble is. Range 0.0 to 1.0, 0.0 disables effect.", "Rumble strip pull factor":1.5, "Rumble strip pull factor#":"How strongly wheel pulls right\/left when running over a rumble strip. Suggested range: -1.5 to 1.5.", "Rumble strip update thresh":0.05, "Rumble strip update thresh#":"Amount of change required to update rumble strip effect (0.0 - 1.0)", "Rumble strip wave type":0, "Rumble strip wave type#":"Type of wave to use for vibe: 0=Sine, 1=Square, 2=Triangle, 3=Sawtooth up, 4=Sawtooth down.", These settings were used since rF1 to produce a rumble strip rumble effect (vibration varying with speed) on flat painted kerbs. These settings also work in rF2, including on proper modelled 3D painted kerbs. But - and this is the important part - that first value (magnitude) applies to all the effects. 0 means none of them are used - and every single controller profile in rF2 has it on 0. So the canned rumble effects are available, but won't be experienced by anyone unless they go and change that value in the json. This happily reconciles my past experience of confirming these did something in rF2, those saying the canned effects are there, and the widespread experience of these not seeming to do anything in the game. Cheers!