AMS 2 | November Development Roadmap

ermo

100RPM
One purpose of dev updates is to tell people about something we think is noteworthy acheivement - important for motorsport simulation and interesting to motorsport simulation fan.
We didn't explain how ffb works (as in ffb in games), but what happens in real vehicles in regards to what will make it's way over to steering wheel (could say FFB real life counterpart).
Just to clarify another potential misunderstanding of our words, we didn't try to make ffb nice. We made it accurate, applying ordinary methods used irl. Fact that this approach works, just goes to prove how underlying physics engine is good. "Nice" ffb (aka realistic steering wheel feedback) is the end result and proof of that.
Nice to hear.

@Renato Simioni has mentioned on several occasions that the technique you use to design tracks also has an influence on FFB?

Would you be interested in shedding some light on where you have encountered "misfeatures" or "calibration issues" with the MADNESS physics engine and specifically the SETA tyre model + the forces it feeds into the FFB?

Are you at liberty to share which adjustments you've made to resolve these issues (if any)?

So far I've only seen it mentioned that you've upped the physics tick from 600 Hz to 720 Hz and that you're using as few filters as possible in the MADNESS FFB pipeline? I've also seen mention of the unsprung inertia calculation fix (an old rF1 bug)?
 
I still wait on the news about the new content, but i kind like the way the tings are going, with the new generation of Racing Sim not allowing Mods, the vast varity of content thats AMS2 brings will be a necessary sim, for those who tired of GT3 dominance (ps: i hope you guys get a GT-Open licencse).

I realy hope the AMS2 bring other BR racing series like, Turismo Nacional, F-Inter, and some others International Racing Series that dont have represantative to be in the major games.

AMS2 needs to keep bring to the public minor race tracks and series (for ex, Argentina have a lot of good tracks) , together the premium content.
 

Cote Dazur

1000RPM
One purpose of dev updates is to tell people about something we think is noteworthy acheivement - important for motorsport simulation and interesting to motorsport simulation fan.
We didn't explain how ffb works (as in ffb in games), but what happens in real vehicles in regards to what will make it's way over to steering wheel (could say FFB real life counterpart).
Just to clarify another potential misunderstanding of our words, we didn't try to make ffb nice. We made it accurate, applying ordinary methods used irl. Fact that this approach works, just goes to prove how underlying physics engine is good. "Nice" ffb (aka realistic steering wheel feedback) is the end result and proof of that.
Thank you for taking the time to answer my post, what people expect to feel from their FFB wheel is a very interesting and vast subject. What FFB can actually deliver based on what it is actually able to read from the game is also fascinating. I would have assume that Reiza studio, as well as the other driving simulators developers, know that subject in depth. The same should apply on how a real car, including its wheel rack, suspension and tyres is behaving when driven on track.
the IRL is very complicated and the FFB can only give feedback on a few forces in a certain manner. Then of course, again, it all depends on how much we want to feel in our FFB wheel in order to compensate, or not, for all the missing feelings we do not get sitting in a fix chair watching a screen. What we are going to feel is left to the developers, even though it is based on the underlying physic, it remains an interpretation.
So, in your case, as per your words, "we didn't try to make ffb nice. We made it accurate, applying ordinary methods used irl", are you saying the FFB wheel will feel like the real counter part wheel, nothing more, nothing less?

When looking at your drawing

It looks to me a lot more like a simplistic schematic that shows what FFB can actually do than a sophisticated view of what a real life car suspension and tyres behaves. We actually had a great video recently, explaining what FFB can and cannot do, showing fairly similar drawings.
 
Last edited:
A somewhat related consideration is that the driver head physics have some considerable dampening which absorbs most of the vibrations - this will be adjustable to some degree as its highly subjective to taste.
Please make the default the same as AMS--it is by far the best implementation of the "headphysics" options I have experienced across all the pMotor implementations!
 
The audio engine (FMOD) is just fine, it is I believe the same as ACC. We are still in process of adding all the various sound effects, filters and finding the right mix. There will be more going on there.

In terms of engine sounds, they sound like a 3L V12 really - pretty much as it did in AMS1.

Of course if we get access to that particular engine for some recording we may find more room improvement, and thats always possible - again as those who are familiar with our sims can attest, we are rarely “done” with trying to make our content better irrespective of how old it is. There are in fact several cars scheduled for new recordings and fresh sound sets. But we try not to let perfection be enemy of good enough - we feel that if it sounds as good as AMS1 on initial release that will be a good foundation to build on over time.
I think this is the issue: AMS sounds pretty darned good and well above average; the YouTube videos sound universally thin and weak. So, there may be an encoding or quality setting issue in the capture rather then AMS 2 itself being bad? Can @Domagoj Lovrić cook-up a super high fidelity sound demo "video" for us? I would even take a pure audio recording that could be downloaded and listened-to. Most YouTube music videos sound like crap relative to the real recordings, too, although that doesn't seem to affect their popularity.

I see consistent comments about poor sound quality/fidelity and it getting blamed on Madness roots. Probably time to set the record straight with that beautiful screaming 3L V12. The AMS one sounds very impressive in-game.
 
Last edited:

BrunoB

500RPM
Premium
After reading all these complaints about the sound in the video I re-visited it.
Normally I use Raceroom which is more or less acknowledged as top of sim sound - and have to say that the sound in this video isnt bad.
The directionality of the tire squeal is pretty low and the engine noise is a bit irritating - but my guess is that the complaints is either because people dont own a competent sound card and/or use a pair of cheap headsets.
Listening through a competent 7.1 X-Fi sound card and a pair of Sennheiser headphones the sound is reasonably clean.
Sterile and rather puristic yes but not bad.

ByTheWay: If just Reiza would use the real surround format Ambisonic instead of the lousy FMod I think a lot of people would invest in both a competent soundcard and real headphones. :)
 

Spinelli

1000RPM
There's nothing at all natural from the way that car slides. It still has all tendancies of Project Cars 2 and ISI physics engine.
- slides happening in strange & lazy conditions almost as if the physics are running in slow motion or the vehicle has suddenly become very sloppy and unresponsive
- when you apply steering lock to correct the slide, the slide then suddenly out of nowhere ends rather than the slide being in a neutral state where the slip angle is neither decreasing nor increasing and you are in control
- as the slide ends when it re-grips, the slip-angle suddenly ends rather than decreasing via driver control
- during slip-angle, vehicle either handles lazy/unresposnive & sloppily or it abrupt & snappy. There's no in-between where the wheelspin and vehicle's inertia and slip-angle are behaving naturally in control by the driver
- there's still that affect at times where you only return the steering to centre (or a few degrees past it) and just keep it there while waiting for the slide to sort of die off and end on it's own rather than you actually applying proper opposite lock and having control and influence of the slide


I personally think that video does much, much more to show the weirdness and un-naturalness of the PCars2 / ISI physics engine during slip-angle than it does to show some sort of evolutionary differences/improvements in core vehicle dynamics/kinematics behaviour. Sure, it looks like there's more forgiveness baked in which means lower chances of crashing/spinning but the overall strange, unnatural behaviour is still there; it just happens slower now with more leeway/forgiveness.
 
Last edited:
Thank you for taking the time to answer my post, what people expect to feel from their FFB wheel is a very interesting and vast subject. What FFB can actually deliver based on what it is actually able to read from the game is also fascinating. I would have assume that Reiza studio, as well as the other driving simulators developers, know that subject in depth. The same should apply on how a real car, including its wheel rack, suspension and tyres is behaving when driven on track.
the IRL is very complicated and the FFB can only give feedback on a few forces in a certain manner. Then of course, again, it all depends on how much we want to feel in our FFB wheel in order to compensate, or not, for all the missing feelings we do not get sitting in a fix chair watching a screen. What we are going to feel is left to the developers, even though it is based on the underlying physic, it remains an interpretation.
So, in your case, as per your words, "we didn't try to make ffb nice. We made it accurate, applying ordinary methods used irl", are you saying the FFB wheel will feel like the real counter part wheel, nothing more, nothing less?

When looking at your drawing

It looks to me a lot more like a simplistic schematic that shows what FFB can actually do than a sophisticated view of what a real life car suspension and tyres behaves. We actually had a great video recently, explaining what FFB can and cannot do, showing fairly similar drawings.

Let's hope we don't need to re use those Jackspade FFB files for PCar's 2
 

Renato Simioni

Reiza Studios
Regarding content, PC2 has animated pit crews. Will AMS2 be including them? And, if so, will they be period-appropriate for the vintage cars?

Because PC2 is kind of silly having a pit crew with full face helmets on when running the Lotus 25...
Yes AMS2 will have animated pit crews, and they will eventually be period appropriate to the track (rather than the car). 3D animation is the one front we´re still doing slower progress than we´d like so it may not be 100% upon release, but this is not the sort of inconsistency we would let slide in the longer run.

I think this is the issue: AMS sounds pretty darned good and well above average; the YouTube videos sound universally thin and weak. So, there may be an encoding or quality setting issue in the capture rather then AMS 2 itself being bad? Can @Domagoj Lovrić cook-up a super high fidelity sound demo "video" for us? I would even take a pure audio recording that could be downloaded and listened-to. Most YouTube music videos sound like crap relative to the real recordings, too, although that doesn't seem to affect their popularity.

I see consistent comments about poor sound quality/fidelity and it getting blamed on Madness roots. Probably time to set the record straight with that beautiful screaming 3L V12. The AMS one sounds very impressive in-game.
Talking this over with Dom yesterday, it seems to be a combination of Youtube compression screwing up treble a bit and having over treated the samples for Miles (AMS1) in a way that´s not optimal for FMOD. In sum if you liked what you heard in AMS1 you should have little to worry about AMS2 matching it on releae it will only get better over time.

There's nothing at all natural from the way that car slides. It still has all tendancies of Project Cars 2 and ISI physics engine.
- slides happening in strange & lazy conditions almost as if the physics are running in slow motion or the vehicle has suddenly become very sloppy and unresponsive
- when you apply steering lock to correct the slide, the slide then suddenly out of nowhere ends rather than the slide being in a neutral state where the slip angle is neither decreasing nor increasing and you are in control
- as the slide ends when it re-grips, the slip-angle suddenly ends rather than decreasing via driver control
- during slip-angle, vehicle either handles lazy/unresposnive & sloppily or it abrupt & snappy. There's no in-between where the wheelspin and vehicle's inertia and slip-angle are behaving naturally in control by the driver
- there's still that affect at times where you only return the steering to centre (or a few degrees past it) and just keep it there while waiting for the slide to sort of die off and end on it's own rather than you actually applying proper opposite lock and having control and influence of the slide


I personally think that video does much, much more to show the weirdness and un-naturalness of the PCars2 / ISI physics engine during slip-angle than it does to show some sort of evolutionary differences/improvements in core vehicle dynamics/kinematics behaviour. Sure, it looks like there's more forgiveness baked in which means lower chances of crashing/spinning but the overall strange, unnatural behaviour is still there; it just happens slower now with more leeway/forgiveness.
First off the parallels between ISI and Madness as far as tyre physics are concerned don´t follow - the tyre model is the one component of the physics engine that is completely different, whatever you perceive their respective flaws to be will have zero relation to one another.

Secondly, unless the car is under neutral acceleration with absolutely no other forces at play there´s no such thing as a slide "dieing off" on its own - if for instance the rear tyres are sliding (whether laterally or longitudinally) but not so much that the car is still able to keep building forward momentum, the sustaining rearwards load and rising downforce could lead sliding tyres to regrip with minimal to no counter steering at all. At least three such examples from 4:38 to 4:56.

There are instances, specially on the outlap where the tyres are colder and under pressure where the tyres produce a lot of lateral movement and somewhat unresponsive steering - you can actually see the sidewall deforming through the outlap and gaining stiffness as the pressure builds up. By the time the hotlap starts however the tyres are closer to optimal operation and fully responsive to the steering inputs and the car is going exactly where it´s expected to go from the input - I know that because I´m the one the driving :p

Beyond that its veering a bit too much into pure conjecture to have a worthwhile discussion, certainly tyre development is always ongoing and theres always something that can be trimmed but Id generally suggest a bit of firsthand experience may be needed to push that point further.

I actually meant to include an extra lap after the ones in the video featuring a couple of induced spins, one from too much steering input at the wrong time and another following a burst of wheelspin exiting the chicane - both with plenty of fast and larges amount of opposite lock but not enough to recover. Unfortunately we ended up cutting it as it had a few art issues and too much time spent getting out of gravel traps to be of value, but it´s available for Reiza Backers in this topic along with further development info for those who may be interested.
 
This may have been mentioned previous - again, I like what I see but the sounds are really light, clinical, clean. It would be amazing if they had a bit more dirty, visceral bass to them. Something that sounds like it's ripping your eardrums apart. Something that RRRE when turned up high absolutely nails without a doubt. All the splutters and crackles and just the overall depth of the samples is class leading. BUT the video does look good, I'm looking forward to this title.
 

Spinelli

1000RPM
First off the parallels between ISI and Madness as far as tyre physics are concerned don´t follow - the tyre model is the one component of the physics engine that is completely different, whatever you perceive their respective flaws to be will have zero relation to one another.
Yes, I'm aware that apparently the Project Cars tyre model was "written from scratch." In fact, that's further evidence that, as I've been saying for probably 10+ years, that some of the issues are either fully or partly related to the car physics engine itself (tyre model aside) - the modelling & simulating of vehicle kinematics & dynamics. When Project Cars 1 was released, I had no idea it was using the ISI physics engine (tyre model aside); well, within the 3rd or 4th corner on my first ever lap of my first ever time playing (not exaggerating) I experienced some of the oversteer abnormalities I've been talking about for a while and I instantly knew it just had to be based on the ISI engine - it's that obvious & easy to detect for me. 3 corners in a game I had no idea was based on the ISI engine and I detected it. The fact these physics peculiarities have been happening on ISI-based physics engine vehicles for 15+ years even when the tyre model is changed or possibly 100% completely different/separate is further evidence that it's not to do (or not solely to do) with the tyre model itself but rather other aspects of the core physics engine EXE/coding.

Secondly, unless the car is under neutral acceleration with absolutely no other forces at play there´s no such thing as a slide "dieing off" on its own - if for instance the rear tyres are sliding (whether laterally or longitudinally) but not so much that the car is still able to keep building forward momentum, the sustaining rearwards load and rising downforce could lead sliding tyres to regrip with minimal to no counter steering at all. At least three such examples from 4:38 to 4:56.
I understand a slide doesn't just "magically" die off, of course there are certain physics (whether correct or not) that would lead to that since everything is (or should) be dynamically based on physics. I say "dies on it's own" as a way to try and convey what it's as if is happening. I should have said it's "as if" the slide dies on it's own.

The slide behaving as if it dies on it's own is another ISI engine thing that can be seen in something like F1 Ch 99-02 all the way to Project Cars with a different tyre model. I think a lot of us ISI physics drivers may not notice it because we adapt to the game/sim we are playing but it can easily be noticed (while driving and just visually). If you look at the video at around the 0:57 or 0:58 mark, you'll see the slide enters what I refer to as the 3rd stage of oversteer (slip-angle reducing, ie. oversteer is reducing) yet you almost immediately brought the steering back to centre (just a little past) and just held it there while waiting for the slide to "die off". This is in contrast to real life where reducing the wheel accordingly with the slide is generally required - but doing so in-game would have almost certainly resulted in a snap-over-correction. This quickly-return-steering-to-centre-even-though-there's-still-a-decent-oversteer-angle-present technique is applied time & time again in RF, PC, AMS, etc. I do it all the time too (as I adapted to the physics). If this technique is not done (let's say a driver hasn't adapted to the physics engine) and the car is controlled in a more intuitive & realistic fashion then a snap-over-correction will happen (ever notice that snap-overcorrections are so much more common in ISI engine based sims compared to others, especially for new players?). I've seen this millions of times in my almost-20 years of playing and watching ISI-based sims. This is not how real vehicles are controlled, whether racecar or roadcar, wet or dry, new tyres or old, under-pressure tyres or not. This is a common theme/occurrence in ISI-based sims - returning the steering to almost-centre in a seemingly premature fashion in order to save a slide otherwise snap-over-correcting will occur if applying a more realistic & natural technique.

Then there's also when the slide does get a little bigger, the oversteer and overall vehicle seems to be unresponsive to driver input unless almost all power is cut (which then leads to another problem of the slide almost instantly, over-abruptly ending). A decent (but not the best) example of this is at around the 1:25 mark - the exit of the hairpin. The oversteer looks like it wants to keep increasing even though it sounds like you were reducing/modulating throttle - not to mention applying a nice amount of opposite lock. The car seemed to not respond at all to your opposite lock, and then not even to any of your throttle modulation/reduction until you lifted the throttle enough for the wheelspin to fully end (resulting in an abrupt, unnatural end to the slide). There was no middle ground; it was either "keep oversteering with no response to driver's steering correction & throttle correction" or "car abruptly re-grips, slide abruptly ends". Even though the oversteer happens in a fairly slow, gentle manner - helping the driver not spin/crash - the slide/vehicle still displays little-to-no actual control. The oversteer angle should be able to be held in a steady state as well as get reduced in a controlled manner. There doesn't seem to be any/enough control of the slip-angle, wheelspin, and vehicle in general while it's in a state of oversteer - it's "as if" the state of oversteer isn't truly dynamic. As it stands, it's like an on & off game. The modulation of wheelspin and oversteer angle in relation to throttle and steering modulation does not seem dynamic and "in tune" with each-other.

That is also another recurring theme in ISI engine even when the game uses a different tyre model (eg, Project Cars).

I'm a huge fan of Reiza and I'll be buying AMS2 for sure, 100%. I just see recurring themes in all ISI engine based sims no matter how old or new (eg. F1 Ch 99-02) no matter the tyre model (eg. Project Cars) which make the vehicle behavior at/over the limit often strange & unnatural.
 
Last edited:

Denis Betty

The older I get, the better I was.
Staff
Premium
@Spinelli I've seen your description of these alleged phenomena in several threads here at RD. I have to admit, some of your descriptions do strike a chord. I've never driven a RL race car or even driven a RL car at high speed (motorcycles were my thing in my youth). I get what you mean about snap over correction in racing sims, and about wheel spin seeming to be less controllable than one might expect. But like I say, I don't have the RL experience to judge for myself.

So, I have a couple of questions, and I'm not knocking you, I'd honestly like to know the answers. Firstly (there is no real point to this one, I'm honestly just curious), did you formulate these ideas completely on your own, or was it something that somebody suggested to you? Would you say that there's a body of like minded people that you've become aware of, before or since formulating these beliefs? Are you aware of any pro drivers (racers/test drivers/pro drift drivers) who agree with you? Could you give examples please and references. :thumbsup:

Secondly, I mainly race rF2 and AC. I actually much prefer rF2 from a pure driving perspective, but as I've always said, this could easily be because I owned/drove more ISI games (GTL, rF1, AMS, rF2) immediately before trying AC. Perhaps I got used to ISI physics and didn't like to change. I actually don't mind AC physics these days, so would that be a good place to look for a difference in the over correction and wheelspin behaviours that you mention?

Are you one of those who feel that ISI cars feel like they rotate EDIT: (pivot) around a central point?

Thanks in advance. :)
 
Last edited:
Those who have kept track of our record though will know that these early investments tend to pay off :)

With regards to the AMS2 / ACC comparison, its worth pointing out AMS2 will pack about 5x more cars and tracks on release than ACC currently does, covering many different series, and that number will more than double over its dev cycle. Thats not a dig at ACC, which features only premium content produced with premium quality and is IMO excellent bang for the buck specially for GT3 fans, but I think people will eventually find AMS2 has a different scope altogether. Its a matter of what you are looking for in a racing sim.
You will be pleased to hear then that over 30% of the content planned for AMS2 is historic cars and tracks :)
 
Last edited:
Top Bottom