Really really cool idea and video!
Plus the calm music fits well with the absence of movement.
The only negative point would be the part in the middle where the cars are going slow motion. I think you should have kept the idea through the whole video and replaced this shot with one of the XRT jumping or FXO upside down. I think you wanted to add some contrast or something, but I think a still would have been better.

Otherwise, really good job! (going to watch it a second time )

Yeah I know it happened before too, but not that often I think.
It had never happened when hitting an incoming XRT before, and now I'm literally sinking into the ground

I just noticed the BF1 is more likely to go through the ground when hitting another car the wrong way. I know you're not supposed to come at another car at 300km/h but I had never gone through the ground until Z34 (did it twice in a row).

Quote from Scawen :

FIX : Endless auto car reset after falling into the underworld

Are you talking about the car rapidly resetting in the underworld? If so, I'm afraid it is not fixed yet, as it is the first thing I tried and it still does it (I also found that rapidly turning the wheel allows the car not to reset for about half a second ).

PS: While playing around in the underworld, I noticed that fast turning the wheels allows to gain complete control, replay attached for more info).
Replay in Z34 of course.

Quote from Squelch :

The replay is paused and I enter Shift+U and set follow.
Stepping forwards is no problem, but stepping back [index 0:27] the view is reset.
I return to the car, and follow it again but this time locking the view [index 0:44].
Stepping forwards is again without problems, and stepping backwards now stays with the car [index 0:50]
I then turn off the lock and the camera is once again reset when backstepping [index 1.04]

I have to say I don't understand why the camera does this on your video. I did the exact same thing and the camera stays on the car, whether I click "lock" or not...

EDIT: @ Scawen: Oh then that's the way you reduced it^^
I did a little test on AU, placed a few chalk objects here and there, randomly and also on precise spots (intersection with AU chalk, intersection between multiple chalk objects).
My conclusion: isolated objects (I mean those that do not intersect anything else) do not (cannot?) flicker anymore.
However, as I thought, the main remaining problem is intersection between chalk objects: they are placed at the same height, so their height difference being 0, they do flicker.
I'm not quite sure how you could fix that, though. (maybe a way of telling when these objects are intersecting, and when it is the case, move them separately?).
I also looked at how "high" the objects were away from the ground, and it is barely noticeable they're not actually touching it (only noticed it while intersecting them with the lower part of a ramp, a barrier, etc.).

Last edited by Bokujishin, Sat 7 May 2011, 19:15.

I tried too, and I can't reproduce either. I'm not locking the camera, only following, but it stays with the car while stepping backwards

@ Scawen: Did you consider my idea about the flickering issue?


I don't have Z28 anymore, but I will reinstall it to test one of my layouts, that makes you drive on top of barriers (I want to check if it was possible too in Z28). However, driving carefully is necessary, as the tyre will eventually glitch.

Thanks for Z32, Scawen
Gonna try this and see how my LFS changes.

Regarding the flickering issue (apparently not entirely solved?), do you think raising the chalk objects a tiny bit above the ground would help?

Quote from Inouva :

Dont think so, the mod have a rb4 with a GTR spoiler and a rollcage inside...

Having these parts on the car doesn't necessarily mean the collisions are changed. If your mod changes the appearance only, you can have your GTR spoiler, but it will not detect collisions.
Well, that's only what I think about it

By the way Scawen, are collision issues that are not directly caused by the new detection system considered off-topic here? Because I did a few tests in various situations to make the car bug (like driving on barriers, and a tyre deforms strangely before ejecting the car). However I don't really know whether they are caused by the new system or a previously present bug.

EDIT:

Quote from TehPaws3D :

Are you talking about borderless mode?, I don't think scawen would add it.

Borderless windows can be easily achieved via autohotkey scripts. I have one that can do that (you must install AutoHotKey to launch the script).

Last edited by Bokujishin, Fri 6 May 2011, 00:37.

Quote from Squelch :

Is that reasonable?

Putting the ramps at that angle are hardly by design, and I believe they are the height and angle for a reason.

Adding more collision surfaces will increase the number of computations to detect collisions. LFS is very good at cutting these computations down to the bare minimum, and gives us the tight network coding that allows many drivers at once. There are compromises that we must suffer to enjoy this. No very steep ramps are one of them.

PS. How do you do that anyhow?

Yeah I know, that was just some random fun to put them like this (you can do it by placing them on the chicane kerbs).
That's why I tried with a properly placed ramp. Conclusion: less chance of this happening, and only happens when trying (see replay).
And I'm well aware of the problem of improving the detection of collisions that would require more CPU, but I think that's the main reason the barriers still send you flying, and also causes all of the strange collisions.

Last edited by Bokujishin, Sun 1 May 2011, 11:53. Reason : adding 2nd replay

There are still some issues left with unmovable objects (see attached replay).
You can notice only the tyres are taken into account (or at least, the bottom of the car is not). Then the rear tyres can't bear the weight of the falling car and... we know what happens next.
I think you should add a collision detection to the bottom of the car, so the doesn't just fall through the ramp

Nothing more to add except that Z31 does its job properly^^

Quote from Scawen :

No, I think that would have been the same with the old version. Watching frame by frame with forces mode on, I see that that front left wheel that has penetrated the wall, is (between the first and second collisions) too deep (downwards) into the wall to detect the top surface. But at some point, the wheel is high enough to detect that top surface. It thinks it is very deep into the object at that point and must provide a great upward force. You can see the extreme tyre deformation at that moment that the "second collision" occurs.

That's not realistic, but it's what is happening, and the contact detection is actually working as designed in that replay.

I've been "studying" the new collision system on BL1: I've always enjoyed flying into the air in the straight by hitting the banners. I tried doing it in Z30 and Z31 too (20+ times) to compare the car's reaction.
Here's my conclusion: the car seems to fly less, but still does. Watching the replay at x0.25 speed, I can see the car going right through the banner before being ejected (mostly with the tyre being extremely deformed).
Which leads to a little question: could this issue be solved by increasing the frequency of collision checks?

And on a little offtopic: when you drive on multiple ramp1 objects, the tyre tends to "fall" in the small gaps between the ramps, and makes the car jump (even with Z31). Will this be fixed in the physics update? (as well as unrealistic tyres bounces after a jump)

I noticed that high speed collisions are sometimes strange with this test patch.
I was in BL1 with the BF1 and hit the wall in T1 at almost 300 km/h and I just went through the wall. Tested with other cars (mostly GTR) and any collision above 200-220 km/h here makes the car fly.

I also tried to push a UF1 with my XRT and the UF1 kind of "shakes" when I push it (don't remember if it was already the case in Z28 though).

Good work nonetheless, I hope you'll be able to fix all these bugs

Although I'm a huge fan of Portal, I think this is not the place to talk about it, nor about how portals could be implemented into the game

Back on topic: Err, nothing more to say here...

Quote from PMD9409 :

Yeah I would like to know what exactly causes this issue.

You can't see what the heck you are doing because the lines won't appear.

I believe this is caused by the objects intersecting the ground. The objects are placed at the exact same height the ground is, which is quite natural, but causes this glitch to happen (you can reproduce it in any 3D software by making objects intersect each other). The easiest way to solve this is to move the object up a tiny bit, and I don't think it would be difficult (just having to spawn chalk lines a bit higher, or making the objects higher).
However I don't know if that would cause compatibility issues...

Quote from Scawen :

Surface friction of objects is now CONCRETE (was previously ICE)
Added a new Ramp2 which is the same as Ramp1 but 2.8 metres wide
Ramp1 and Ramp2 are stored in the exe and loaded into all tracks

Woohoo!
Just wondering, why was objects friction even set to ICE?
Now, we can use those ramps as we please, and that's a really good thing

You can count on me, Scawen, I will try the improved collisions with the red barriers for at least an hour (non-stop of course)
Then I'll have a try at open configs.

Can't wait for this test patch, so many good things in it, good job!

PS: Are the ramps also considered as "unmovable objects" for which the collisions have been improved? Just asking because I always found it irritating to just slide on those.

Last edited by Bokujishin, Mon 11 Apr 2011, 17:59.

Hi Victor, I just downloaded the latest version to try out the new import option. The screen is now cleaner without all the keyframes here and there
However, instead of a blue square, do you think it would be possible to have a picture of the car (or even a generic car-looking thing) that can change scale so that we can evaluate the distance between the different cars when using multiple replays? (it would be even easier if we could have several replays playing at the same time, but I guess this is not possible, is it?).

Anyway thanks for your work!

Really good and well-edited video, I'm downloading it right now
I especially enjoyed the little crash, crunch, etc. effects

Quote from mrocht :

Look , its like if you paid for ms flight simulator . Is it updated ? no..

Except flight simulator is updated every now and then... but not for free
We all want the update, it's just some can't wait (even if it's true we've been waiting for a long time already).

Thanks for your answer.
I don't really think it's a necessary feature, but I had to record a bit more and then delete some screenshots to get what I wanted.
What I mean is (at least while recording fullscreen), since I can't see for how long it has been recording, I have to know where to stop manually.
Not that it's that troublesome, but having a way to tell "I want to record for xx seconds" could help a bit, IMO.

Anyway, I tried combining 2 replays and it works just great! (I just have some artifacts after using matte difference).

Oh, almost forgot a little bug, although it doesn't seem to be related to LFS Record: I recorded the same camera animation 3 times (once without any car, and then with 2 different cars), and I noticed the trees had no leafs at all, except in one of the recordings... And as I just started LFS, my trees still have no leafs.

Quote from tolas :

Will it be ever possible to import track and car models into max with all telemetry data for cars(not just car trajectory)?

I was wondering if it will be possible to create prerendered videos in 3ds max based on LFS models and telemetry one day?

I guess I can only say I want this to be possible, already been wondering if it was possible to get the rotation of the wheels via insim so we could have the whole car animated in 3ds max.

Quote from Flame CZE :

Yes, but isn't it because the wheel is spinning? But I am confused now :P

Quote from Amynue :

Here is another video in a bit better quality and with better lighting:
http://www.youtube.com/watch?v=J3KshKS_tvo

If I'm not mistaken, accelerating, with or without spinning, produces similar forces on the tyre, and Amynue's video shows an almost no spinning example.

In case some people still doubt the wheel is rotating clockwise, here's a proof: http://www.youtube.com/watch?v=3Ug6w4ZjwVo
Take a look at 0:47, in this case the wheel rotates CCW and the thicker part of the tyre is on the left, so in the other picture, thicker part on the right -> wheel rotating CW

This is a very nice tool Victor

By the way, is there any way to animate the "active camera" state or do we have to record multiple segments to get all the cameras?

Wow, thanks for the report, Scawen, some nice work we got there