What?? race_1 sucks, default sucks more. Default setups should be reworked for the final version.
I agree with the rest
Congratulations for the award!!!! 
EDIT: Ooops, seems everyone noticed the setup thing LOL
What?? race_1 sucks, default sucks more. Default setups should be reworked for the final version.
I agree with the rest
Congratulations for the award!!!! EDIT: Ooops, seems everyone noticed the setup thing LOL
Wasn't trying to suggest that he wasn't a very good programmer, I just don't believe the current programming has reached the limits of the internets capability...
IMO, the problem is that there will never actually be an impact, because the cars are either not touching, yet, or are hanging inside each other.
So it would make sense to say: The more they hang inside each other, the higher the impact speed must have been. This works only if the impact is elastic, which (however) means, that the objects don't change their shape (= the cars don`t get damaged).
The problems are:
1. the energy that is used for the deformation of the cars is not calculated, thus the cars behave like Billard balls.
2. if you have lag and your car moves into another, the engine thinks you crashed into it with like 700kmh :]
3. When you are very fast, you can drive through small objects and your car will be accellerated extremely, because the engine thinks at some point, that you crashed into the wall from the opposite direction, with very high speed.
Yes, it does sound weird, but that's my theory
So it would make sense to say: The more they hang inside each other, the higher the impact speed must have been. This works only if the impact is elastic, which (however) means, that the objects don't change their shape (= the cars don`t get damaged).
The problems are:
1. the energy that is used for the deformation of the cars is not calculated, thus the cars behave like Billard balls.
2. if you have lag and your car moves into another, the engine thinks you crashed into it with like 700kmh :]
3. When you are very fast, you can drive through small objects and your car will be accellerated extremely, because the engine thinks at some point, that you crashed into the wall from the opposite direction, with very high speed.
Yes, it does sound weird, but that's my theory
I think your theory is right At least makes sense with the strange things LFS does. If you are fast enough, you can even go through some thigs (the starting lights in the drag field) without even noticing or getting damage.
Do you think this would ve solved by adding more Hz to the physics engine? AFAIK LFS works at 100Hz, right?
At least makes sense with the strange things LFS does. If you are fast enough, you can even go through some thigs (the starting lights in the drag field) without even noticing or getting damage.Do you think this would ve solved by adding more Hz to the physics engine? AFAIK LFS works at 100Hz, right?
Yea maybe, but for online play you will also need more packets/second, and this will require a high bandwidth. Sadly, only a few people have like 10mbit or more...
Im happy with my 1024 dsl
Im happy with my 1024 dsl
And another well deserved award for the devs.... this is fantastic! I sincerely hope this will be a boost to our community as well!
Anyway, big up for the devs!! Awesome!
Anyway, big up for the devs!! Awesome!
That's my understanding of how seemingly 'standard' collision detection systems work... it seems in this case it needs to be a little more involved than 'standard' though. Don't see why some basic vector calculation using a bunch of previous packets, acting at least as a sort of 'check value' is such a laughable concept, but like I said, I'm not a programmer
i believe its a deeper more generals problem ... the proof for this is that is also occurs in solo play and in practically the same way and strenght
its hard to tell at which rate the collision detection works atm but most probably its neither the 2 khz nor the 100 hz the inner and outer physics loop run at (see for yourself by drinving through the fence at the bl carpark from the outside slowly) and the problem certainly gets worse if you add lag to the equation ... but either way i assume the basic problem is lacking synchronisation between the different levels of calculation lfs does
its hard to tell at which rate the collision detection works atm but most probably its neither the 2 khz nor the 100 hz the inner and outer physics loop run at (see for yourself by drinving through the fence at the bl carpark from the outside slowly) and the problem certainly gets worse if you add lag to the equation ... but either way i assume the basic problem is lacking synchronisation between the different levels of calculation lfs does
I guess the only ways to fix this are:
1. increase the operations / second (-> slow)
2. verify somehow, if the collision is "valid" - if the relative speed of the objects to each other is high enough for the distance they are hanging into each other. This would at least kill all the "lag-explosions", where the cars are flying through the air and are all over the place.
A higher amount of packets could help either, but anyhow S2 limits the online packets to 3. In S1 it was possible to use 4-6 not only in LAN, but for people with slower connections this was really bad.
I would suggest to allow those higher values again, and add a small function that warns you before you join a server, if there are too many people online, which will result in lag because of your bandwidth.
Maybe it's even possible to use only every second packet from slow connection members, and the other racers put this packet twice into their mp replay file so they wont get out of sync.
Another (probably good) idea would be removing the collision from a car if it has much lag, so it will have no clipping and will no longer be able to crash other cars...
[EDIT]
about the singleplayer "bugs":
when you are driving at 100kmh and have 100 operations/s, you already move 27.8 cm each calculation, this is quite a lot when you look at those barriers which have approx. this width.
with 1000 ops/s, we would have only 8.3 cm at 300kmh - much better
1. increase the operations / second (-> slow)
2. verify somehow, if the collision is "valid" - if the relative speed of the objects to each other is high enough for the distance they are hanging into each other. This would at least kill all the "lag-explosions", where the cars are flying through the air and are all over the place.
A higher amount of packets could help either, but anyhow S2 limits the online packets to 3. In S1 it was possible to use 4-6 not only in LAN, but for people with slower connections this was really bad.
I would suggest to allow those higher values again, and add a small function that warns you before you join a server, if there are too many people online, which will result in lag because of your bandwidth.
Maybe it's even possible to use only every second packet from slow connection members, and the other racers put this packet twice into their mp replay file so they wont get out of sync.
Another (probably good) idea would be removing the collision from a car if it has much lag, so it will have no clipping and will no longer be able to crash other cars...
[EDIT]
about the singleplayer "bugs":
when you are driving at 100kmh and have 100 operations/s, you already move 27.8 cm each calculation, this is quite a lot when you look at those barriers which have approx. this width.
with 1000 ops/s, we would have only 8.3 cm at 300kmh - much better
2 of the best awards to get
goodjob devs and everyone involved
goodjob devs and everyone involved
The problem is, that collision detection itself is awfully slow already, so letting it run at 10x the frequency would kill performance.
What we basically need is a "sanity check" that looks at the relative speed a car had a few frames before the collision. Now if we collide and we calculate a collision force that would send you flying to the other side of the track, even if we just clipped into each other due to lag, we could look at the previous speeds and see, that both cars were driving at each other with 2km/h and can thus use a much lower force.
PS: Tyre physics run at 2000Hz already, just FYI
Congrats to LFS for this spectacular achievement! Both awards are weell-deserved and I'm sure that over the coming year, LFS will consolidate its position as the simulation with best physics and best multiplayer as well as reaching new frontiers in some of the other sections Where is Deleure now so we can shove this in his face?
Where is Deleure now so we can shove this in his face?Another win for the Scavier corporation. Nice one.
This two awards let the other sims like a university degree final project,
congratulations devs, i sense lfs have the best physics engine since i drive first time S1, and S2 change is impresive..
uff i cant wait for aerodynamic patch
congratulations devs, i sense lfs have the best physics engine since i drive first time S1, and S2 change is impresive..
uff i cant wait for aerodynamic patch
It's great that LFS won in those 2 (most important) categories. But let's not forget that it was vote based system and in the end these are usually won by the title that has most active users.
Still, great achievement by scavier (and LFS forums)
Still, great achievement by scavier (and LFS forums
)As I said on another thread, LFS has a very realistic physics IF you don't use bug exploiting sets. BUT to be competitive these days, you almost NEED to use buggy sets.
Probably the guys at BHM don't know that (well, they say that the default sets are great, so I'm sure that they don't know that), or didn't tried to be on par with WR laps.
OK, Simbin physics system is not marvelous, but depending on the way that you look, today it is better than LFS physics. At least you don't get extra speed if you put the nose of the car all the way up.
I'm afraid that those awards will delay the release of a new patch, since Scawen will probably feel on a more confortable situation, but what I see is that the non hardcore fans from LFS are starting to leave LFS, and as I said before, LFS won't survive only from the hardcore fans (Scawen, Eric and Victor won't be able to afford the developing and living costs with only the money from the hardcore fans, they will have to find a job, what would make the LFS development even slower, or leave the LFS development)
The hi-nose bug, tire grip abnormalities and locked diff issues... all heard million times already. They need to be fixed, when they will be fixed is up to scavier and posting here changes nothing.
The bhms didn't decide to who to give the awards. The whole system was based on voting and LFS community knows how to vote.
There are still 2 categories left. One of them is the Flat tyre award. You can bet all your money to the "Flat tyre & LFS" combo.
So what? Those are just bugs. Those can be fixed more easily than creating the whole driving feel from scratch to Simbin/ISI engine which it lacks.
As far as I know Scawen is the only one who quit his dayjob for LFS. Am I right?
I didn't know that.
Yes, but this is taking so long to be fixed, and Simbin didn't stop developing their engine. They can even do that before Scawen finish fixing those bugs (if he is working on that, what at this point, I don't know if he is doing that or doing research for other features), since they have more people and more money involved, and I bet that Simbin has good programmers.
And I believe that going against who complains about it here in the forum changes even less.
Ok, some months ago, I agree that it was pointless to be complaining, but c'mon, those bugs are since S2 L, that I even don't remember when it was released (before August 2005 for sure, because patch P was released on August, that is 7 months ago, and after patch P it was only minor bug fixes and language packs). We don't have even a single progress report (apart from the Scawen's kitchen progress report).
And how long you think it takes to develope ISI-based engine to something different than it has been since F1C? Already three games based on same engine: GTR, rFactor and GT Legends and it still has the same flaws.
And by the way GTR 2 by Simbin (or "Blimey Games" because of the recent hullaballoo) estimated release is summer 2006. I hope we see LFS physics patch before that.
But don't get me wrong. I agree that we need a progress report. I would be happy if it was just one line saying "The physics issues are under work".
I'm sure we will get a comprehensive physics update soon (next few months). Scawen said that is what he will be working away on from January and he said that patient people will see progress when the time comes.
And I hope that I'm still alive when it happens...... There is even patient people waiting for Jesus to come again to Earth, and they strongly believe that it will happen.
Me either, and if possible, a estimate date for the next patch release (even if he says 5 months or so)
But they don't know that what they are waiting for is actually being worked on at the moment. Why is it that whenever someone that is complaining about the nature of development of LFS and they are told to be patient they come out with the arguement "but what if I die before an LFS patch comes out?". Don't be thick, you won't be dying anytime soon through natural causes so unless you get run over by a car, there is still pleny of time. And even if you do die because you got run over by a car it is not like you died because LFS wasn't released. You can be sure that a patch will be out sooner rather than later (my educated-ish guess is for the middle/end of March) and it will be comprehensive and will make LFS' physics light years better than anything else out there today - its already the best as the award shows, now it just needs to be improved. And if you think Scawen would just stop working on physics because LFS got an award for it, then you are completely wrong and haven't read any of his posts on the forum - he works on what he feels inspired to work on and this has brought LFS so far. Not a money-hungry approach but rather enjoying your work and making a living at the same time. This is why the physics are so comprehensive even if they are not perfectly accurate - they weren't rushed but rather, everything is calculated in real time and no short-cuts were taken to save time/money. Nagging Scawen like you are will not help. Seing as this IS Scawen's "job" and his only source of income and not just a hobby, LFS is being worked on. Scawen said he'd be working on physics between January and the release of the patch and he said he would go quite for 3 or so months because physics takes a long time. The physics patch will come when it is ready. Seing as it is February and there is no physics patch out, you can be pretty sure it is being worked on. So just be patient or find another hobby.
LFS deserves what it won :thumbs:
Shame they don't leave the contest open to older sims, GPL/N2003 could have given LFS good competition in the multiplayer and physics departments as well as probably wiping up the sound catagory.
Shame they don't leave the contest open to older sims, GPL/N2003 could have given LFS good competition in the multiplayer and physics departments as well as probably wiping up the sound catagory.
Well, if he said that, why not on a progress report but underneath in this forum ? This is kind of a progress report, and this I believe that would calm down the people that is unsatisfied with LFS physics (or better, the way that ppl have been exploiting bugs) today.
I didn't see all Scawen's posts on this forum, so now we can expect by April or May a new physics patch. If it just fixes the aero bug and makes a workaround for the differentials, I will be happy. For now, I think that we don't need more than that.
Game Of The Year!
(95 posts, started )
FGED GREDG RDFGDR GSFDG