Modded tracks, which are worth going for?

no offense taken, but given that you think they look like visual meshes would illustrate the fact that you also think they arent high enough resolution surely?
yes i agree spa, mugello, ring etc do look right, just like lilski's, or the luccaring example i posted. my examples show that in the case of paul ricard and others, the mesh is lower resolution than kunos non scanned tracks and it doesnt need to be, as it has no impact on fps and can have negative side effects as mentioned.

i dont see how they can be the visual mesh, in the case of paul ricard, that file is called Paul_Ricard_colidable.kn5, and in the case of a zandvoort its definitely the physics mesh as 2.kn5 is the detailed track mesh only and the main kn5 has what must be the visual.

however i am happy to be wrong as i know full well you are more qualified to talk on the subject than me, maybe i am viewing the files incorrectly, and apologies to the author of any of the aforementioned tracks if i have done so and made false claims. but so far i havent found anything to suggest i am viewing them wrong, as i can drive a track, e.g. road america mod, feel its crap and low res, go and view the wireframe and sure enough its got whole tris stretching across the width of the track (no wonder it feels so bad), so it seems the software is viewing the correct mesh (in the case of this track i think there is no physics mesh and i am just driving on what is the visual mesh...)

personally paul ricard feels great to me, so maybe i am viewing the visual mesh, either way in my list i did say they are still good to drive, so no offense is intended anywhere, just my findings from digging around

 

I'm not saying they can't be physical meshes, they just don't look like the ones I know. I have no access to any of the files so I can't tell, but I'll believe you on your word.
Obviously as Zandvoort does not have a laser scanned surface, it might just not need the triangulated mesh we all know. Simply because there is no real data for bumps, so it's best not to try and fake any. In your Zandvoort example it does however look like the more pronounced bumps are moddeled (a few lateral rows that might be a bump), but that's probably from watching onboards, etc. Though I have to confess I did not notice the lower resolution while racing at Zandvoort, so it's probably fine for most average simracers. At least it's 9000% better than tracks like Solitude..
Oh well, derailing this thread, sorry

Edit: yes PR does feel great. But perhaps the mesh densisty is just enough to pass that on to your wheel

Edit2: Oh btw, there has been a huge discussion on the Coste thread, started by @Lunae about proper physics meshes. Worth a read

 

hmm i dont think we are derailing the thread as its all relevant to what the OP asked.

agree with you on zandvoort, i screenshotted that section because i know there are bumps there when driving it, whereas a lot of the track looks less bumpy in the wireframe. of course hard to tell just looking at the wireframe but yeah it seems that this is the physics mesh, which imo is a decent example of what modders should consider a benchmark (in terms of resolution, and having a separate mesh for physics and visual of course), as i assume this is similar to what you would get out of race track builder, with the bumps observed from onboard or whatever add either after in max or in RTB if thats possible. as for not faking bumps across the whole track, i think thats up for debate, luccaring, bridgehampton, etc all convey a realistic feeling road imo (obviously bridge and gentrack are intentionally larger bumps compared to a typical racetrack) and their meshes look just like the kunos ones, there are even conversions from other games which the converter has followed a similar mentalitiy and been able to create a "laserscan style mesh" (i.e. not what you see on zandvoort, monza 66, trento, but much more like any of the kunos scanned tracks in terms of randomness) but they arent authorised ones so cant be discussed here.

point is that those tracks in my "generally lower resolution" mesh list, are not a whole lot more detailed than the solitude one, so they are possibly close to being affected by the physics issues, so just my 2c and putting it out there that they could consider increasing the resolution with no drawbacks

 

Yeah I guess you're right on that. A dense (but maybe not as dense as the Bridge') RTB mesh would be fine with bumps added later on.
I've got a similar mesh as Bridgehampton with only very very small randomness (random 2 mm vertical offset) to liven up the road a bit, eventhough Assen is generally very smooth. It feels a ton better than when it's flat. With signature bumps taken from onboards added, any mod track can get that same feel. And they should, indeed

 

I went a little over the top with the latest build of the bridge with 1 million triangles on the physical mesh. I haven't seen any issues with it being that dense. The soon to be released 0.9.5 will have it as an experiment.

As for other tracks just about any scratch made track is worth your time. Even some conversions are great. But there are a large amount of conversions that are improperly done that you should avoid. They have all the bumps in the world but not the kind you want.

It is also always good to show your support if you enjoy someone's track. Just a simple thanks in their thread or a small donation goes a long way.

 

Hmmm now we need a small RC car to drive and test that dense mesh

 

The physical mesh, does it impact cpu performance the denser or more complex it is?

 

Nope, more is better

Sent from my Nexus 5 using Tapatalk

 

It should impact mainly on loading time but also slightly in run-time calculations (and then CPU occupancy) because the algorithm, that find on which triangle or quadrilateral each of the 4 tyres is, has to process more elements.
Once the element is found the interpolation functions give the current elevation value.

 

I almost agree with all your track categorizations. The only ones that I feel different is Modena (I feel no bumps or randomness) and Tor Poznan (here I feel perfect bumps and ramdomness like in Kunos tracks)...

 

I'd give a go also to old NK tracks: Newbury, Prato and Aosta

 

I think it's not far from the thruth to say that there are no such things like good quality conversion from last gen title. The surface is always very low poly if not heavily re-worked. Hate to drive those tracks with a powerful wheel because you get a lot of violent jolts through the ffb. Not even starting how these spikes affects the behaviour of the car..

 

Just soften the suspension up a bit you'll be fine

 

No you're not. Your wrists are still aching on the next day..

 

ah yes i made a bit of an assumption with modena, as i couldnt view the wireframe for it i just guessed it would be the same style as mill metropark but it does feel quite flat, however certainly doesnt suffer from too low res mesh so its a keeper. poznan you are probably right, i was just going by the wireframe which is similar to zandvoort, my categorisation just suggests it could be modified to have the more random style mesh but i guess that doesnt matter now its getting scanned

 

To Op....
Go to Grobnik with loose fillings and you'll need virtual ones afterwards.
It's one nice and pure 'fun' track though.

 

80% of the mod tracks lack Majorly in this dept.....15% of mod tracks have Far OVER the moon type of mesh with Very Unrealistic surface mesh (which tend to crash the motion software or my try and break my neck)....the other 5% of mod tracks are actually making an effort. The Ripped tracks are the worst of all with them feeling Literally flat as a pancake. Kunos laser scans are BY FAR the best feeling surface/driving mesh i have EVER had the pleasure of driving on.

 

Lucca Ring is most likely THE most detailed mod track to date, ir Sindaco is well aware of this and I will be helping him with the Final Final mesh when he's all done with the fancy 3d scans he's currently doing. the actual mesh you posted basically backs up my feeling i get through my back side.


I Will make it a point to test all the tracks you have in the first category. Lake Louise on the "old" section of the track is in the "over the moon detail dept" basically shook me so bad in a 235I that i have to report it as a No go (235i suspension isn't THAT Stiff, lol).......Aosta is in my download folder already, just never got installed. Gentrack if i remember was rather flat, Granted that could have been updated since i last tried it. Transfargarasan also felt rather lack luster....but ill give them all a whirl again and report back.

PS the Zanvoort mesh pic. Is that the Big Bump just before the left hand up hill section ? if So I CURSE that freekin bump, Its FAR bigger than it needs to be imo and gives me a HECK of a Shunt at speed.

 

appreciate your input hopefully you and sindaco can work out a good baseline to share with other track modders as a good practice in the random mesh area.

agree the lake lousie one is OTT especially in that section, that is clear even through ffb alone

i think you may find aosta a bit bumpy unless you are in a road car, but as i mentioned its not a defect of low resolution at least.

and yep thats the section of zandvoort, my fav part of the track

 

I will most likely test out the questionable tracks with something that wont try and kill me if i find something fishy.

PS any chance i can see La Source mesh before and after the corner ?