Does rF2 implement chassis-flex or semi chassis-flex?

That image there is misleading (yet still lacks non-linear surface data) and false advertising on their part, since their Nordschleife physical mesh is only 33K verts (way too low for such a large 22.8km track), with like 5 verts on average across the width of the track.

 

I've opened AC Spa myself in an editor and the surface looks exactly like posted in the linked thread. They use a less dense mesh for the visuals (which may well be only 33K) and the denser one for physics, you probably confused the two. From this picture anyone can see that there are closer to 100 verts than 5 verts across the road width.

 

I'm not confused, since their Nordschleife physical mesh looks nothing like that. But as I said before, raw polygons alone aren't good enough.

 

a) Fundamentally speaking, in theory, the method of laser scan on track surface may result more accurate than other methods. But reality is another matter. Does any one know how and what techniques ISI was actually using? I knew they are not laser scanned.

b) Objectively speaking, I found ISI tracks actually very accurate. Some of them can be as accurate as ~95% in comparing with real track. Last year, DrRacing studied telemetry data and comparing the results between rFactor (rF1) and real data. He published his results on his blog. The track he studied is the Monza. He build his own F3 mod car to match the real one:
“…a Formula 3 in 2008 FIA configuration driven by a real driver on the same circuit during a free practice session of the British Formula 3 Championship race held in May 2008 in Monza. One of the reasons why I have chosen exactly this event to make this comparison, is that the car was used on track set in a way that was overall pretty close to the “Manufacturer delivery configuration”………”
The lap data you used to compare are very close in terms of drivers’ performance (RPM, gear Shifting, Steering turns, etc). Below are his results: the traces of spring/damper unit movement of real car on real track against that of his rF1 mod car on Monza. As I recalled that I read in somewhere, the standard deviation in difference between them is about 5%. This is very remarkable. Keeping mind this is resultant of combined both car and track modeling. It implies the track surface accuracy could be even higher than ~95%.




c) Subjectively speaking, we cannot judge by number. Feel of bump/drop/bounce on ISI track is better than other sims I tried. Here is video demonstrate this. My seat vibrating and bump (see the vibrating modes of strokes on the actuators) sync “exactly” with the chassis bounce. My seat level sync with track surface level perfectly. The inclination angle of track surface matches to that of my seat as can be seen.

 

Can you show an image of that or you're just basing yourself from the comment of a person on youtube who converted ac nordschleife to rf2? He said that ac has 33k verts for the entire nordschleife, and then he increased it to 300k for rf2 conversion. He just said numbers, didn't show any proof though, other than visual representation of the track in rf2 game while driving. And just for the example, assuming 300k for rf2 and 33k for ac, how is the 300k version as accurate as the real nordschleife if he only subdivided the polys and didn't model them from the laserscan cloud points? Again, just assuming those are the numbers as he said, couldn't 33k verts feel just like the real nordschleife when driven with a car? They developed nordschleife in ac in collaboration with rsr nurburg, so they tested both game and real life versions. Whatever the amount of verts is, it probably feels real as the track is.

Which cars and tracks you tested, and road surface conditions in both games?
Because for example I see videos of comparison with motion rigs between rf2 and ac, and they think they are using the same car, but lotus49 and brabham bt20 are different mechanically and that will reflect in the driving feel. So the comparison isn't that good from the start.
Then the tracks, this is the place where the comparison won't do that much, because despite being the same tracks on paper, the physical track road will be different between games, as they aren't copy paste between games, so the physical modelling of the road/bumps/elevations will be different.

 

And therein lies the problem, but it doesn't stop fans of whatever game spruiking its merits and putting down the ones they don't like. Who really knows what you end up feeling at the wheel, you can drive on a mesh with a 1mm density (note: I don't think any games have this!) but that doesn't mean you're feeling 27778 points through the wheel every second at 100kph. And, even if 2 games did actually process the same number of points in a given time, they might take very different approaches in how the tyres interact with that surface and then how everything else is put together to give you handling and feel. Most of these sims are so close to right - heck, they're better than what we had 10 years ago and even then people were talking about games/mods being 'perfect' - the differences are pretty hard to spot without introducing a lot of subjective stuff into the mix. But people still go for it

Most people seem to agree rF2 tyres behave better than rF1 tyres, and presumably that's got a lot to do with the new modelling. So that's a positive. How it compares to anything else, and which approach is better, and which is correct (none! You can't perfectly simulate this stuff!)... life's too short to argue about it.

 

He is exactly what we want around here.

His nod could bring a few Historics modders around.......think PnG, GTR/GTL/rF Historic tracks .................... Oh my.

As far as I was aware David's last try of rF2 was in early Corvette GT2 /Lime Rock demo, hope he gives full version a good tilt.


Honestly does anyone see PnG or similar mods ( DLC or otherwise) better placed in another sim ?

Sorry I just can't for the life of me, no offence meant.

 

That's the other side to it, whether any of our sim engines can actually give anything extra when it comes to all that data. For the purported iRacing millimeter accurate stuff, no way that any of your devices you interact with (monitor, wheel) can actually give you anything out of that.
Love to know what kind of frequencies the sim engines are doing for the different systems haha!

 

In iRacing, you don't drive directly on the point cloud, but on a parametric spline surface separate from the graphical mesh (as in every Papyrus sim) with additional bump maps where needed (introduced in Nascar 2003) and from what I understand they are similar to TDF sine wave definitions. See http://nrtracks.com/nrtracks_board/index.php?topic=145.0

Kaemmer stated in an interview http://issuu.com/autosimsportmediallc/docs/autosimsport_v6i2_march2012/31

Raw polygons without any smoothing and/or filtering, especially as course-crained as on the Assetto Corsa Nordschleife is not a good surface representation.

 

In terms of surface detail, I think it is limited by Sims' API output refresh rate (telemetry data update rate). No matter how detailed surface is and how fast you drive, this limits what and how you can feel on your simulator. For most of sim this rate is 100Hz. I think iRacing API physics update rate is 60Hz. If one talks about down to 1mm resolution, I do not see a point here.
Note: core eng physics update rate could be much higher up to 400 - 500Hz.

I did such analysis with a spectrum analyzer from SimVibe SW outputs on rF2, rF1, and GSC sims. I found they all truncated at exactly 100Hz:

@QUF
In terms of subject feels on directly comparison with exactly same cars between AC and rF2, I could not find exact same cars between two sims. the closet I could find are:
AC Corvette C7r gt vs rF2 Corvette C6r GT2
AC New NGTR GT3 vs rF2 old NGTR GT1
on same trucks: Monza and Silverstone.