Quote from DarkKostas :

Scawen, as you said this will be multiplayer incompatible and you have some other things in priority. Does this means that the next patch won't include this? (I am happy anyway that this is being worked on, just saying)

I'm talking about an incompatible test patch in this run of test patches. I expect the next test patch will include this. But maybe not, if I do some other compatible fixes first (e.g. enabling physics on some track surfaces where you can fall underground in open configs) to be tested along with the tyre contact detection fix. It's possible those could be tested in one more compatible test patch before the incompatible one.

Quote from avetere :

I reckon this means mpr-format will be altered?
If so, would it be possible, you let me/us know, in what way (e.g. which packets are affected)?

No, it's just that the new object numbers would make old guests have all the wrong objects and missing objects, so they would go OOS. Same with the enabling of qualifying. So I just have to add something that forces incompatibility - does not allow Z33 and previous versions to connect.

Quote from Whiskey :

/\ that and InSim changes for making public the packets that manage the AX objects, as LFS already have them, but hidden for internal use only.

LFS has them but InSim does not - InSim has all its own packets, and does not yet contain anything for this. The new ones will share the object structures from the existing LFS packets and the layout file format. Of course I will add documentation when I add the packets to InSim.

You'll be pleased to know that one more "flying cars" issue seems to be fixed in my version. In the attached Z32 replay, I reverse my XR GT into a barrier so that it intersects (because the current car collision system doesn't notice that). Then the wheel detects a contact with the top of the barrier and the car goes flying (because the penetration is deep, the force is great).

The wheel contact test ray now starts from the centre of the wheel, instead of above the wheel, so this penetration is not detected and the car just gets nudged outwards. I guess this same type of collision might happen between cars sometimes. Also this will stop that strange looking bug where a car can be pinned to a fence by its tyres.

The main thing this week is that I'm trying to unify the autocross objects system into a single file that is shared between all tracks. This has a few advantages :

1) All objects will be available on all tracks.
2) Each object will have the same object number on all tracks, helpful for external autocross editors or InSim programs that can add or remove objects.
3) Some of the objects can be optimised for graphics and physics (CPU time reduction).
4) Helpful in the development track editor.
5) If a new object is added or improved, it will be available at all tracks.

It's a bit dangerous to say what I'm working on, just in case it turns out that it can't be done. But I think it will be OK. There needs to be a conversion function to repair all the old layouts to use the new object numbers, automatically when loading. It will be multiplayer incompatible, so it will be time to do some other things on my list as well, like allowing qualifying in open configs and enabling the randomised start lights time. So I guess a Friday patch looks unlikely at this point.

Quote from pearcy_2k7 :

After you've finished this scawen will be see some content implemented soon after or have you still got quite alot left to do?

After the official patch is released I will be back working on the tyre physics.
When the tyre physics is done it should not be long to the Scirocco.
Hopefully not long to some S3 content after that...

But... please ignore me when I say "soon" or "not long". That's just the belief in my mind that keeps me optimistic and on the case. It always takes much longer than I expect.

Quote from Squelch :

Quote from Tango :

im using z30 and was on the cargame s1 server,south city chicane route when i seemed to momentarily hit a hole in the road and it lifted my nose in the air after i passed over it,pic attached.
Tango

Quote from Kid222 :

I have suspected Z30 glitch. Happened in one endu race at AS5R last week to one of the participants _twice_. I tried to get some info from him, as what version was he running, but he told me only that his antivirus was updating, no answer since then, so i'm posting this, because it's at least funny to look at.
In case you want to look at it, watch the 002 car towards the end of the replay, approx. 1:25. Exactly at the place, where the tarmac turns dark-ish, he's shot in the air and flies out of bounds. The second case was also exactly at the same spot, happened only to him.

Looking at both of those replays makes me suspect they are very similar. In the first, the car is flipped up and to the left, and the second the car becomes airborne a couple of times before wiping out. They both have similarities in that it happens at a join in the track (or at least a change in texture) and they appear during a multiplayer session only. The first incident does not appear to have any forces involved, but the second definitely does when forces are viewed.

I postulated before that this might be caused by dropped packets or other networking problem causing a loss of sychronisation, and the report that an antivirus update was also happening during the second incident reminds me of a couple of similar incidents I have had while testing on my LAN. I have noticed in the past that some background processes can severely impact LFS during a multiplayer session, particularly where it coincides with network activity and high disk use. An antivirus update might just qualify in this regard, and even more so if it involves possible firewall changes or activity. The answer is to make sure no other applications or other background processes are running while playing LFS, but its interesting to note the problem appears at joins in the track.

My opinion is that this has possibly been a problem for some time, and has only just been noticed, or because of the recent changes to physics detection the problem has been exacerbated. In my case, I have dismissed this problem as I already knew that my system was under pressure and it didn't warrant further investigation, but it might catch some others unaware. I'd also hazard a guess that both of the systems in those reports are single core too.

I don't think it is a networking problem, because the car's physics runs locally, it shouldn't be affected by dropped packets or anything like that. And they weren't near enough to another car to suggest that another car's lag was a likely cause.

I guess, and I hope, that this was just a problem with the Z30 collision system, something like the Fern Bay gantries (which, unlike these, were easily reproducable).

I haven't heard of any false collisions with objects or falling through the road in Z31. I found a genuine cause for the Fern Bay type of false collisions in Z30, detected and eliminated them in a way that seems clean and good, so I hope these are gone now.

Quote from THE WIZARD DK :

i just experienced z25 isnt compatible with z31.. for some unknown reason tire physics just doesnt add up.. i was going to cargame.nl S2 and was gonna race in xrt at oval.. even a wr set up, wouldnt keep me straight.. well its possible to drive, but has absolutely no grip for racing.. like driving on ice.. also makes braking distances ALOT longer than expected.. which is not so good in races..

I really wouldn't expect that to be the case. Tyre physics should be nearly the same, only the surface detection is more accurate now, so the road surface might appear to be in a different place by 1 or 2 mm (or a few mm if you are on a steep uphill or downhill slope, or several cm in odd places where the old detection really went wrong - though that was worst at the end of segments like the pit garage walls, so wouldn't usually affect tyres at all). But that difference is usually small for tyres and shouldn't affect the driving feel or lap times. I think it would need a more thorough test. You would have to take the same setup, and try it once in Z28 and again in Z31, to do a direct comparison. And you could post the setup here in case anyone else wants to test.

Instead of waiting until tomorrow, I thought you might like to try these useful updates.

Many translation files have been updated
Connection list is now visible in SHIFT+U mode
Skid marks are now visible on open configurations
Multiplayer : Cars driving far out of bounds are spectated
Race positions list / map car colours now work on open configs
InSim : IS_BTN did not allow documented limit of 240 characters
FIX : False collisions that showed up near pit lanes at Fern Bay

http://www.lfsforum.net/showthread.php?t=74012

NOTE :
I have not yet tried to do anything about missing barriers or ground without physics enabled.
I still intend to look at some of those issues but this is an intermediate update.

But that (GTR number plates) is totally off topic for this patch.

I'm not doing any unrelated requests, I'm just trying to get this patch finished, so I can carry on with the tyre physics.

Quote from vistaman :

I guess this was reported already. Check the video from 0:40.

http://www.youtube.com/watch?v=u65OeDZty7g

Quote from Squelch :

That problem is common to all instances of those gantries on Fern Bay, even standard configs. Hit at the right angle, and the car gets fired backwards violently, or shot forward at incredible speed.

I know the reason for this now - the segments of those gantries are constructed in a way that is unsuitable for the physics detection. It was unsuitable in the old version as well but that just didn't result in an out of bounds collision. I need to write a little code to detect and switch off physics on unsuitably constructed segments.

Quote from Tango :

im using z30 and was on the cargame s1 server,south city chicane route when i seemed to momentarily hit a hole in the road and it lifted my nose in the air after i passed over it,pic attached.
Tango

That looks quite nasty, and I can't reproduce it with my test program. Because it's an MPR, it's not clear exactly how this happened. I would like to see an SPR of this, if anyone can reproduce it (a wheel falling underground on proper road) at this place or anywhere else.

Hello Racers. The new open configurations open up some areas that have problems, because the collision detection has been switched off on some segments, for optimisation reasons.

Though I haven't investigated further, it may be possible for LFS.exe to switch on the contact detection in some areas. I hope that there aren't TOO many places like that... I can't just switch on *everything* - that would include overhead power cables, some inaccessible buildings and would reduce frame rates.

But if you could post a screenshot here of areas where you fall through the ground, maybe I can put some code into the exe that switches on those particular "no physics" segments.

For a start - this area at the Fern Bay mini-oval (posted by kart-36) :

FIXED IN Z33

Last edited by Scawen, Fri 13 May 2011, 10:45.

WARNING : THIS IS A TEST

NOTE : THIS DOES NOT CONTAIN NEW TYRE PHYSICS / VW SCIROCCO / ROCKINGHAM

Hello Racers, here is a new Test Patch : Z34.

You can read the story behind this thread in the main forum thread :
http://www.lfsforum.net/showthread.php?t=73874

Z34 is semi-compatible with the official version Z28.

Z34 can play version Z28 replays.
Z34 can join Z28 hosts.
Z28 cannot join Z30 - Z34 hosts.

You cannot upload hotlaps made with Z30 - Z34 because they are test patches.

Please keep a backup of your LFS.exe from version Z28 so you can go back to Z28 if there are any problems.


Changes from Z33 to Z34 :

FIX : Occasional false collision detections near segment ends
FIX : Endless auto car reset after falling into the underworld
Blue flag detection function for racing on open configurations
Dedicated host window now adds new messages without flickering
KY : No humps / tyres in open configs entering GP Long from Oval

Translations :

Greek translation is now up to date for the SHIFT+U mode updates
New Ukrainian translation included for testing (nearly finished)
Hungarian translations of training lessons updated


Changes from Z32 to Z33 :

Improved wheel contact detection to reduce chance of flying cars
Speed limit no longer enforced if autocross start position exists
All languages updated except Trad. Chinese / Greek / Belarussian
BL : Repaired floating laptop on tool box in the first pit garage
BL : Enabled all the rallycross section barriers in open configs
SO : Repaired a sunken barrier closing pit lane (Sprint Track 1)
SO : Enabled various barriers that were missing from open configs
FE : Enabled contact detection around mini-oval and nearby walls
FE : Enabled a barrier in open configs (at corner near a bridge)
KY : Enabled some barriers at entry / exit to infield / outfield
KY : Enabled speedhumps and tyres at entry to GP Long from Oval
AS : Enabled missing barrier across side road near AS3 chicane
Attempted FIX : Occasionally refuelling did not match set amount
FIX : Momentary loss of TV camera at Fern Bay Club X turn one


Changes from Z31 to Z32 :

Optimised transparent objects (e.g. fences) to help open configs
Adding / removing autocross objects no longer resets all objects
Much better frame rate with many autocross objects (faster draw)
Reduced Z-buffer flickering of chalk objects and pit stop markers
Custom layouts race progress / history now displayed on LFS World
FIX : On joining host, cars not past finish line showed wrong lap
FIX : SHIFT+U follow mode now stays on car if you rewind a replay
Command /modified=yes allows a private host to avoid CP checks

InSim :

Command /cv now only cancels game votes (restart / end / qualify)
OutGauge : OG_SHIFT and OG_CTRL (keys) bits added to OutGaugePack
Lap timing info added to IS_RST (standard / custom / checkpoints)


Changes from Z30 to Z31 :

Many translation files have been updated
Connection list is now visible in SHIFT+U mode
Skid marks are now visible on open configurations
Multiplayer : Cars driving far out of bounds are spectated
Race positions list / map car colours now work on open configs
InSim : IS_BTN did not allow documented limit of 240 characters
FIX : False collisions that showed up near pit lanes at Fern Bay


Changes from Z28 to Z30 :

Contact detection and collisions :

Improved contact detection with road / walls / pit garages
Improved collisions with unmovable objects (e.g. red barriers)
Unmovable objects (e.g. ramps) now have concrete friction level

SHIFT+U mode :

New free view mode replaces the old "low" and "high" path modes
Object or ground targeted by the mouse pointer is now detected
Left mouse button moves view point by "dragging" target point
Double click to move view point directly towards target point
Mouse wheel moves view point towards or away from target point
Left + Right mouse buttons rotate view point around the target
Arrow key movement follows ground below view point when possible
Store up to 10 views per track configuration with SHIFT+NUMBER
Recall a stored view with CTRL+NUMBER (or click on the button)
Click on a car to follow it / look away to stop following
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

Autocross editor :

Maximum number of autocross objects is now 800 on all tracks
Object positioning is much better - using mouse target detection
Less CPU / better checks when adding / moving / deleting objects
FIX : Could click on invisible replay slider when buttons visible
FIX : Could add marshall circles out of bounds then undeletable
FIX : N key in SHIFT+U with options on caused overlapping text
FIX : High altitude layout objects appeared under ground on load

Open configurations :

New "drive anywhere" open configs (CTRL in track select screen)
Open configs can be selected with text command like /track=SO1X
Open configs can be used as racing circuits by adding checkpoints
Checkpoints can now be up to 62 metres wide to include pit lane

Optimisations :

Removed continual small memory leaks (materials list corruption)
Number plate generation is much faster - reducing pit-out glitch
Frame rate increased - drawing most world objects is much faster

Interface :

Unlock screen now renames unnamed player to user name on exit
Clicking selected track loads that track (like pressing ENTER)
Windows messages processed in a cleaner way (affects key presses)
Mouse movements are processed more efficiently (drag / slide)
FIX : Mouse look info could go off screen in some languages
FIX : An invalid view could be selected when no cars in race
FIX : Meeting room scroll bar was invisible in recent versions
FIX : Text commands could be used to load layout in hotlap mode

Graphics / Audio :

Improved display of help text / lesson text / welcome messages
F9 tyre diagram and SHIFT+L suspension diagrams now antialiased
Improved tyre optimisation - no missing parts of deflected tyres
Entry screen logo and frame are now drawn if a dialog box is open
FIX : In Driver Options the driver intersected with world objects
FIX : Reset from behind a barrier at South City could stop sound
FIX : Missing shadow on ground near fences at Autocross track

Multiplayer :

Wear limit for changing tyres can be set to 100% (never)
List of hosts can now show hosts you are not licensed to join
Host name is displayed in task bar and window title (if Latin)
Implemented VOB mod protection (physical changes will cause OOS)
Host options are now readable when opened from game setup screen
Default host IP address is now 127.0.0.1 (address of local host)
CTRL+SHIFT now displays time and date in all multiplayer replays
Removed TCP filter that allowed old versions in the list of hosts
FIX : Some problems resulting from multiple requests to join race
FIX : Pit stop did not end if car was knocked out of the pit lane
FIX : Some crashes that could result from an invalid unlock state
FIX : Some missing images (e.g. track select) caused LFS to crash
FIX : Suspension could incorrectly stay broken on remote computer
FIX : Host restart exploit found by Luigi Auriemma (join packets)

InSim :

New command /ck cancels any kick or ban votes in progress
New command /cv cancels any votes (kick, ban, restart, etc)
NLP / MCI minimum time interval reduced to 40 ms (was 50 ms)
New collision packet gives details of contact between two cars
Longer IS_MST now up to 128 chars / specify sound / send to all
FIX : Clutch axis / button was not reported from Controls screen

Misc :

Misc options : F9 / F10 accelerometer can be shown as one value
Improved detection of invalid window positions when starting LFS
URL command is now case insensitive - LFS:// or lfs:// both work
Security : InSim can no longer be initialised from a URL command
Cruise : no need to drive a lap before serving DT / SG penalty
Cruise : current lap is not displayed if lap timing is disabled
FIX : InSim camera with vertical pitch would cause LFS to crash
FIX : OutGaugePack ID was always zero regardless of ID in cfg.txt


DOWNLOADS :

NOTE : Z28 (or later) must already be installed! If it is not, please install Z28 before installing this test patch.

PATCH Z28 to Z34 (SELF EXTRACTING) (If you already have Z28) :
www.liveforspeed.net/file_lfs. ... =LFS_PATCH_Z28_TO_Z34.exe (1.6 MB)

PATCH Z28 to Z34 (ALTERNATIVE ZIP) (If you already have Z28) :
www.liveforspeed.net/file_lfs. ... =LFS_PATCH_Z28_TO_Z34.zip (2.0 MB)

NOTE : Z30 - Z34 are multiplayer compatible so if your host is running Z30 - Z33 you do not need to update it

DEDICATED HOST Z34 (non-graphical version for hosting only) :
www.liveforspeed.net/file_lfs.php?name=LFS_S2_DEDI_Z34.zip (1.6 MB)

Last edited by Scawen, Mon 16 May 2011, 16:46.

Quote from EQ Worry :

I wonder if this also covers contacts between cars, so that there is less crazy results of small touches, especially in open wheelers. That would be awesome, but even if that's not the case, it is great to see a serious progress in LFS.

No, collisions between moving objects (including cars) are not affected.

There are three collision / contact changes :

1) Collisions with movable objects - they now have the same hardness as segments (track / walls / etc). Fixes the red barrier bug.

2) Objects (like autocross objects) have the friction of "concrete" when touching a tyre - was previously "ice". Helps on Ramp1 and Ramp2.

3) Accurate segment collisions - mainly affecting the contact detection at the *end* of segments and this fixes the pit garage exit bug.

Quote from Skagen :

Does this also mean improved contact detection with the speed bumps in i.e KY3 chicane?
Often it seems like it's up to luck if you get suspension damage or not when "cheating" (or just missing) the chicane and fly over the speed bumps. If it was spot on every time (meaning suspension destruction if hitting the bumps at 230 km/h), the chicane would be even more intense.

PS. Would also be interesting if the front spoilers on the GTRs and open wheels could be damaged or even completely run off if they smash into curbs and bumps.

Those speed bumps are now hard if you hit them with the car body so I don't think you will do that in a race. However, there is probably still luck involved, as they are thin you will probably still "miss" them sometimes, if they happen to be in a position between physics updates.

Quote from vicbel :

One more shoutout to Scawen to give host/server admins the possibility to limit the amount of gas and/or the type of tyres to use. This would spice up the possibilities in terms of increasing the strategy factor in races.

Sorry, but I won't be doing that kind of thing. I'm not really taking general requests. Not meaning to single out your post in particular, just replying to you but the same applies to a lot of requests that have been made in this thread. I've done loads for this patch in a short time, I'm quite tired and want to get it out. So I'm working very hard to clear my list of bugs and add some remaining improvements that I have noted, as quickly as possible. I do not want to add more things which could create more bugs. I know I did the tyre wear 100% option request, but that was easy to do, in a compatible way. And I did some other suggestions because they were easy enough and very on topic.

But we are getting to the point where people are starting to come up with more and more suggestions, which are really nothing to do with the subjects of this patch, while I'm struggling away trying to finish the things that are already in the patch!

Quote from [Audi TT] :

My message would be helpful.
Found another bug. On the road SO3. The same situation as with the track FE6r machine goes on the air in a certain area. Screenshots are attached.

I can't fix any of those kinds of bugs. That would be a simple thing to edit, if we were releasing a track update. But this patch will be just the exe and a few files. The tracks are remaining unchanged to retain compatibility, leaving all the hotlaps valid and so on. It's not time to delete the hotlaps tables - the time for that will be when the real physics updates come. So there are no track updates in this patch. The open configurations are a function of the EXE only, they are auto-generated so no tracks need to be released.

- Objects in escape road at BL chicane removed from open configs
- List of Hosts now shows X or Y if an open config is selected
- Car reset now works on open configs (if reset is enabled)
- Wear limit for changing tyres can be set to 100% (never)
- Compatibility : new packets not sent if connected to old host
- Compatibility : new host does not allow old versions to connect

Quote from Eclipsed :

How about this roundabout - will it be accesible fully? Might be some nice addition to change usual SO4R layout or just have some banging fun in super small oval. I guess the other road up to highway isn't complete or is it?

Sorry, but no, because that roundabout is closed off with "segments" rather than "objects". That SO section would need reworking in the track editor to allow access to all of that roundabout and the motorway.

Quote from boothy :

What about at Kyoto when running the KY3 config, the roads into the KY2 track are blocked off with the tyre walls, similarly with the Aston configs and their tyre walls - e.g. on historic bend?

Well, those do vary between configs - that means they are automatically not included in the open configs.

Quote from kimd41 :

Why it is so important to make this patch compatible with Z28? Can't it just be another obligatory patch and make the needed changes to the game without any worry?

As the physics have not changed, there is not point making things incompatible (and deleting all the hotlaps) at this point - so no new tracks. And although this patch has taken me quite a few weeks, it's still a "quick" update, which it wouldn't be if we started to edit the tracks. It's quite convenient and helpful that when people get the test patch, they can still join all the old hosts.

Quote from GP4Flo :

Yes please! Some retro tracks without chicanes would be cool.

I opened the Blackwood one this morning, and the extra Fern Bay road I had already done.

All I can do in this version is make extra "objects" disappear, but not "segments" which are the curvy things used for walls, fences, tracks. The open configs are basically auto generated. Any objects which differ between configurations are excluded from an open configuration.

But, for example, at the BL chicane I just added some hard code to detect those cones and barriers when BL loads. It then switches them off in the Rallycross config (not affecting anything). So that means they are no longer shown in the open config.

If there are some more places where something that easy can be done (removal of "objects" that aren't needed in one of the real configs) do tell me about it.

Quote from aroX123 :

Also Scawen, will "drive anywhere" also work for blackwood carpark ?
Will you remove the red and white barriers so it can lead us to the carpark?

Also the Autocross track, Im sure the road between Autocross, skidpad and drag would work.

No, I've had a look and there's nothing simple I can do in those places. Two problems : (1) they are not closed off simply by objects and (2) some areas were never designed to be driven on, some are incomplete or have physics collisions switched off.

I agree that there is a lot that could be done if we started editing the tracks specifically for this. I'm sure we will consider the open configs in future track releases and updates.

Quote from Storm_Cloud :

Whilst we're on about opening FE and BL, is the same going to happen at SO?

There's one I would like to see opened, but it's not part of any config so not sure whether it will be or not. There's an off ramp part way down the straight on SO Long before the last tight corner. That would lead into downhill braking into a tight corner which looks fun and would also cut down the length of the first and espcially final "straights", cutting average speed and making it a bit more Monaco-esque.

SO does have open configs (all tracks do, even if it is pointless, e.g. Westhill). I see that part you are talking about and it is a good idea but unfortunately those barriers are made of "segments" and there are some other problems I can see too.

Dear Programmers, as you probably know there is a new test patch coming.

http://www.lfsforum.net/showthread.php?t=73874

Old InSim programs will still work with the new patch, but you can make use of some new functions.

1) Note the Track[6] fields will indicate the new configs like SO1X and FE3Y etc. Those two would indicate an open config using SO Classic Forward start grid / Fern Bay Gold Reversed grid.

2) New collision reporting. Please take a look at this packet specification and have a think about how you would use it, and if it needs any more info :


3) IS_MTC now can have up to 128 characters (including a zero at the end) in a variable sized packet. You can also specify the sound effect of the message and send a message to all connections with a single call.


4) I've had a look at EQ_Worry's suggestions in this post :

http://www.lfsforum.net/showthread.php?p=1337106#post1337106

Particularly suggestion (2) a traction control bit I'd like to include because that is something missing. And I'll have a look at (3). I have not studied them yet so I can't say for sure.

I really must keep any updates to the minimum, because I want to release the test patch as soon as possible so I can get back to the tyre physics. So if you have suggestions, please try to stick to very simple ones and think through how they would work.

Thanks

Dear LFS Racers,

Since the release of Z28 there have been various fixes and improvements that were not related to the new physics. Some of them were already added into my Z28 compatible version as well as the new version. But until now there weren't enough significant updates to release a new version, while Z28 was so stable.

NOTE : This patch is mainly compatible with Z28, so does NOT have the new tyre physics, the VW Scirocco or any S3 content. It is an intermediate update with some interesting improvements for LFS while I continue with the tyre physics. The new version can connect to Z28 hosts, but Z28 cannot connect to the new version's hosts.

The main things in this patch :

- New free view in SHIFT+U mode - not bounded by the path
- Improved collisions with unmovable objects (e.g. red barriers)
- Number of autocross objects increased to 800 (was 512)
- Open track configurations allowing you to drive anywhere
- Collision detection and reporting system (via InSim)
- Graphical optimisations improving frame rates
- Many other fixes and improvements

The story behind the new patch :

A few weeks ago I was working on some track editor improvements that Eric had requested, when I came across a fix that could possibly affect key presses in multi byte languages. An LFS user in Japan kindly tested the change, to make sure the fix didn't break the text input. I added a few updates from the new version into the patch before sending it to him.

Talking to Victor about possibly releasing the updates in a patch, he asked if it was possible to do a quick fix in SHIFT+U mode to help with his LFS Record program. I have long been a little frustrated in SHIFT+U mode so I got a bit carried away with that and rewrote all the camera control. It took a few weeks (that made me nervous as I really want to finish the new physics) but the result is a lot nicer to use. You can move the camera around freely, not bounded to the track path. The editor is fully aware of where your mouse is pointing and that helps with mouse movement and object positioning.

Because the new view is not locked to the path of the track, LFS does not know which objects are visible, so the whole track must be drawn. I managed to find several optimisations that improved the frame rate in this situation, and that gave me a thought : if you can be in SHIFT+U mode and draw everything, then how about allowing "open configurations" where you can drive around anywhere in a track area, with the barriers removed? I thought that would make the patch more interesting, and managed to get it done in a few days.

To make the open configurations useful, you need to put barriers up using the autocross system. But the unmovable barriers were famously troublesome, often throwing your car several metres in the air if you touched a red and white barrier. Looking into this, I found that I was able to improve these collisions a lot without any major changes. It doesn't seem to affect the hotlap replays (tested on all the world records).

Over last weekend I added a collision detection system, that can send information about collisions to InSim. And today I've increased the maximum length of message packets, which will be helpful for InSim programmers.

If all goes well, it should be possible to release this as a test patch within one to two weeks. There are a few things left on my list, including VOB mod protection and some small InSim updates. I am aiming for one week - I do want to get back to the physics - but I know that finishing things off sometimes takes longer than expected.

Here is the full list of updates :

SHIFT+U mode :

New free view mode replaces the old "low" and "high" path modes
Object or ground targeted by the mouse pointer is now detected
Left mouse button moves view point by "dragging" target point
Double click to move view point directly towards target point
Mouse wheel moves view point towards or away from target point
Left + Right mouse buttons rotate view point around the target
Arrow key movement follows ground below view point when possible
Store up to 10 views per track configuration with SHIFT+NUMBER
Recall a stored view with CTRL+NUMBER (or click on the button)
Click on a car to follow it / move away to stop following

Autocross editor :

Maximum number of autocross objects is now 800 on all tracks
Object positioning is much better - using mouse target detection
Less CPU / better checks when adding / moving / deleting objects
FIX : Could click on invisible replay slider when buttons visible
FIX : Could add marshall circles out of bounds then undeletable
FIX : N key in SHIFT+U with options on caused overlapping text

Open configurations :

New "drive anywhere" open configs (CTRL in track select screen)
Open configs can be selected with text command like /track=SO1X
Open configs can be used as racing circuits by adding checkpoints
Checkpoints can now be up to 62 metres wide to include pit lane

Optimisations :

Removed continual small memory leaks (materials list corruption)
Number plate generation is much faster - reducing pit-out glitch
Frame rate increased - drawing most world objects is much faster

Interface :

Unlock screen now renames unnamed player to user name on exit
Clicking selected track loads that track (like pressing ENTER)
Windows messages processed in a cleaner way (affects key presses)
Mouse movements are processed more efficiently (drag / slide)
FIX : Mouse look info could go off screen in some languages
FIX : An invalid view could be selected when no cars in race
FIX : Meeting room scroll bar was invisible in recent versions
FIX : Text commands could be used to load layout in hotlap mode

Graphics / Audio :

Improved display of help text / lesson text / welcome messages
F9 tyre diagram and SHIFT+L suspension diagrams now antialiased
Improved tyre optimisation - no missing parts of deflected tyres
Entry screen logo and frame are now drawn if a dialog box is open
FIX : In Driver Options the driver intersected with world objects
FIX : Reset from behind a barrier at South City could stop sound
FIX : Missing shadow on ground near fences at Autocross track

Multiplayer :

List of hosts can now show hosts you are not licensed to join
Host name is displayed in task bar and window title (if Latin)
Host options are now readable when opened from game setup screen
Default host IP address is now 127.0.0.1 (address of local host)
CTRL+SHIFT now displays time and date in all multiplayer replays
Removed TCP filter that allowed old versions in the list of hosts
FIX : Some problems resulting from multiple requests to join race
FIX : Pit stop did not end if car was knocked out of the pit lane
FIX : Some crashes that could result from an invalid unlock state
FIX : Some missing images (e.g. track select) caused LFS to crash
FIX : Suspension could incorrectly stay broken on remote computer

InSim :

NLP / MCI minimum time interval reduced to 40 ms (was 50 ms)
New collision packet gives details of contact between two cars
Longer IS_MST now up to 128 chars / specify sound / send to all

Misc :

Misc options : F9 / F10 accelerometer can be shown as one value
Improved detection of invalid window positions when starting LFS
URL command is now case insensitive - LFS:// or lfs:// both work
Security : InSim can no longer be initialised from a URL command
Cruise : no need to drive a lap before serving DT / SG penalty
Cruise : current lap is not displayed if lap timing is disabled
FIX : Improved collisions with unmovable objects (red barriers)
FIX : InSim camera with vertical pitch would cause LFS to crash
FIX : OutGaugePack ID was always zero regardless of ID in cfg.txt

OK, thanks for the comments.

I think that nearly all possible points have been made now (from the community's side).

I'm sure it would have been different if we had known in advance how long this development stage would take, I guess we could have tried to plan other releases in the meantime. But it has always seemed like I was near to finishing the tyre physics, so no change of plan was required.

Remember it was the Scirocco that showed up the issues in the tyre physics (there were already known tyre physics issues but we couldn't get the Scirocco to drive as safely as a real one does while restricting ourselves to a realistic setup). So, releasing the Scirocco has been out of the question until the tyre physics system is quite close to reality.

Releasing Rockingham has also been out of the question. We want to release a real track when the behaviour of the cars is right (not perfect, just close enough so the bumps feel right and the lap times are just like real ones).

Who knows, if I had known in advance that the tyre model would take so long, maybe Eric would have thought that he should focus on some S1 / S2 content updates. With the benefit of hindsight, it's easy to pick holes in what we have done and point out the mistakes we have made. But as I said, we always thought that in a few weeks or a couple of months, the new tyres would be ready to roll.

But the carcass simulation, which I had to look into because the tread simulation model clearly ignored too much of how a tyre actually behaves, took me in a whole new direction for a good part of last year. A long time of studying and finding out a lot more about tyres than I knew before, and had never thought I'd need to know. It made my head spin quite often, I'm not an eighteen year old on a university course. I couldn't even read this stuff for full eight hour days, had to take it a bit easy at times.

But still I usually thought I was near the end. So that is probably the reason for our dreadful "strategy" and lack of communication. It wasn't a strategy at all, and I didn't want to communicate until I got solid results.

I am fully aware that this post isn't going to convince many people. But I thought I'd write it anyway, to help you see slightly more from my point of view.

I will unsubscribe from this thread now, which I followed up to this point. But it's a bit too distracting to receive the email notifications and too tempting to come in here and read the latest posts, which takes me away from research and development.

News from www.lfs.net :

Dear LFS Racers,

We hope you have had a good year.

The stable Z28 has been the latest version for a long time now, while we have been working on things for the long term future of LFS. As many of you know, the tyre physics system is being updated so all the cars will behave that bit closer to reality. It has proved to be a challenging and unpredictable project, involving a lot of studying, mathematics, testing and rewriting.

Sometimes, we had a version that felt good to drive and it seemed that soon we would be able to release it to the community. But then we found situations where the simulation produced incorrect results, leading to further investigations into the way tyres perform in various circumstances and why they do so. One thing tends to lead to another and for a while it seems that the more you learn about tyres, the more you realise that there is so much more to find out!

Scawen has written a little more about tyre design and simulation in case you would like to get a bit more technical.

The VW Scirocco and Rockingham have seemed quite close to release for some time now, but we decided to wait for the updated tyres because these things from the real world will be best experienced with the most realistic physics system. Eric has been working on some more S3 content but he would like to get nearer to finishing before we show any new pictures.

Victor has recently produced a special program "LFS Record" which can help you to produce smooth movies from Live for Speed, using advanced camera control. It is still in a beta stage at the moment. The program has a lot of features - too much to describe here, so if you do make movies or would like to have a go, take a look at the LFS Record thread on our forum.

- LFS Developers

Quote from Amynue :

Why we can't see them in Z28 patch?

I don't know, you'd have to ask Eric.

As I said before, this thread is Tyre Physics Progress Report + Scirocco discussion.

If you are starting to think this is random question time for Scawen, please think again... maybe it's time to close the thread.

Quote from cargame.nl :

Maybe Scawen also finally learns that it's not the best idea to put TEXT MESSAGES in the middle OF THE SCREEN.. Blocking all view.

Why the sarcastic comment?

The tone of your comment suggests that you think I am really quite stupid. But maybe your comment is the thing that needs attention, because I really have no idea what you are talking about.

I am not aware of putting any text messages in the middle of the screen which block all view. That sounds like a very serious bug - maybe it should be in the bug reports section? Or maybe you are exaggerating and really you have noticed some text that could be in a better position? In which case you should be writing in the improvement suggestions. Post a screenshot as well so we know what you are talking about.

Either way, your comment is not about tyre physics or the Scirocco. Sarcastic posts like yours and scipy's are not welcome here. From the tone of your text, your intention is to annoy me. You are wasting my time. If you want to post here, stay on topic.

Quote from rediske :

See Scawen? ...didn't hurt at all. or did it...?

What counts is that you stopped by and said something...

I said something because I had something to say, not because I was forced to. I am quite immune to the moaning that goes on, it doesn't disturb me much at all - it's just what happens on forums. And there are many people here with a perfect understanding of what we are doing, and they do explain it to others round here sometimes.

For a long time there was nothing to say, with some half developed tyre physics and unfinished AI drivers. When I had something to say, I said it. Hope you understand, as many others do!

Quote from rc10racer :

I don't think Eric's work rate is the issue because no one has a clue how fast he can make tracks/cars without him telling us, my opinion is that Eric has plenty of work that could be made public but the final decision ends with Scawen on what gets released.

It's up to Eric what of his work gets released, never have I tried to stop something of his being released!

Quote from Amynue :

Thanks Scawen for that post. Good to know that AI updates are already done and they are better than before. Sorry to ask but how is the VW stability control, suspension updates and limited setup system? If the tyre physics is the last thing to polish then it seems like we are getting close.

Those things are not done yet. I am not expecting them to be too difficult though. But where have I heard that before?

- I'm not planning to try and make the VW stability systems as refined or complete as the real thing. I tried something before, when we were apparently close to release at the end of 2008. Already learned something by taking some wrong turns at that point and I hope that it won't be too hard to get a couple of reasonable systems in place that enhance the handling.

- The suspension updates I hope won't be too difficult, though that is an area where one thing could lead to another, so I don't really know. I think it will be a case of "where to draw the line". It's doesn't need to be the final update to suspension systems but if I could get a bit of anti-dive and anti-squat working then it would help with the default car setups which are more softly sprung now.

- Limited setup system is just a whole bunch of logic programming, meaning it's a load of rules, nothing complicated like physics and geometry. So it shouldn't be too hard. But there will be a few things to sort out, like how it will work in the interface and deactivate an "extreme" setup when you join a "limited" setup host, and you'll need to specify if a setup you are working on should conform to "limited" or "extreme" setup rules.

After the updates are done and the beta testers are quite happy with it, I think there will be a decent period of public testing. The limited setup system should be quite interesting. Class balancing will also need to be sorted out, with new tyre physics and default setups on limited setup hosts, we'll probably find that some adjustments are needed. And of course, public testing will usually bring up other issues we can't think of yet.

Last edited by Scawen, Wed 14 Jul 2010, 10:28.

Quote from Macfox :

We're now at the point where the devs show a lot of contempt for those who question LFS progress.

I've just scanned through the last few pages of the thread. I haven't looked at the thread for a long time, as I've been working a lot recently.

I think you are not correct about the contempt. We're just improving this simulator. I'm trying to program the tyres to behave as realistic as possible. It takes a lot of investigation and testing.

The reason I don't make progress reports is not contempt. It just doesn't make any sense to me to do, say, monthly progress reports. One of them says "I'm working on the tyre physics" and then the next month "I'm still working on the tyre physics" and so on. I just want to say something when there is something interesting to say. But the tyre physics has been a lot of development, a lot of testing and trying new things, finding it did some things well and others not so well. Changing some things, looking for reference material to refine constants, taking a break, coming up with a new idea to test, trying it out, taking some wrong turns, investigating something that turned out impossible, rewriting something, finding a better result, and so on...

Maybe you would suggest that all those twists and turns along the journey of development is exactly what I should be writing in the weekly or monthly progress reports, but actually I don't want to. When I am following a line of investigation, I don't want to say what I am looking at, instead I just want to get on with it and see if it turns out right or wrong in the end. It's kind of a private activity, not sure how to explain that.

Quote from Macfox :

The fact we've seen very little quantifiable evidence of progress for so long, is truly worrying. Sure the champions will hang around for whatever update materialises, but the damage to reputation and future prospects of LFS, will be a big hurdle to overcome if the old prosperous days are to return.

The statistics of this thread speak for themselves. The devs need to show some initiative and put LFS back on track otherwise they are just wasting their time pursuing a dream that will never meet reality.

I don't want to give any quantifiable evidence of progress at the moment. I suppose that would be a video or some graphs of tyre forces? No, I just want to get on with it. For example today I woke up with an idea to try which might have a subtle effect on the force outputs from tyres in some situations. It's a pleasing one because it can be coded in a few minutes and will be easily visible on the generated force graphs, unlike some things that take days to write. I don't want to say what it is, it's just something I'm looking into. Sorry to everyone who would like insight into every step of the tyre physics development but it's something I do for my own interest and to come up with really nice handling cars that our community will enjoy, and that's nice because it should give us more sales eventually as well. I'm not interested in doing something like writing a book on tyre physics development, if you see what I mean.

But what I can say is...

I think the tyre physics have come along a lot recently. There were some significant changes over quite a long time this year and when I found the handling was quite interesting and certain old issues were sorted, I started to look into the tyre heating, trying to get a version ready again for the testers. That meant sorting out the AI drivers, who only knew the old physics system. If they could drive well in the new tyre physics, they could drive consistent laps and so inform me if the tyre heating was working well.

The AI ended up being about three weeks of work, and I found it personally quite an interesting development time - which is good because that's what I'm here for! Anyway their grip prediction systems are a lot better and they can now drive quite close to the limit, the best and most reliable AI drivers yet seen in LFS. At some point I'd really like to work on their overtaking code, though that is not a high priority for this physics update patch.

So the AI could be crossed off that published list of big things to be done, and I feel the tyres are getting close. I've been testing them out with Eric and Victor for the last two weeks. They have been coming up with interesting comments and leading me to further investigations and improvements. I'm trying to get the version ready for our private beta testers to have a go. Also looking at notes I have here, specific comments from our testers, the last time they did some testing. I try to deal with each point then cross it off the list. No point giving them back a version that still has issues they mentioned before.

From experience I have learned that it's not worth giving the beta testers a version that I still know is flawed. While there is still plenty for me to work on full time, and Eric and Victor are pleased with what they see but are still able to demonstrate situations where they think the handling is not quite like reality, the best thing to do is just carry on. For how long, I don't know. Not everything goes as expected. Some things I try make strangely little or surprisingly large differences and sometimes they open up a new area of investigation. Personally I am surprised overall how much there has been to learn and understand about the behaviour of tyres.

Today I thought I'd just have a look at this thread, but there are too many new pages for me to read in detail. Anyway, you may be assured we have not stopped development. I guess if we stopped development then we'd make a post about it. I'm not sure how that would happen, maybe if I got run over or contracted a fatal illness or perhaps someone will come and offer me a flight to the moon but I have to train at NASA for a couple of years. Unlikely, and there's really nothing better I can do than work on this simulator. I just don't want to write about every step on the development path, so that's why there is a long time between progress reports.

Last edited by Scawen, Wed 14 Jul 2010, 09:48.

There is some misreading of my post going on. LFS development is not on the back burner, and I never said I don't like my work.

I said, I am working on the physics and I look forward to releasing it because of the excitement of it and I added that of course also it will be good for earnings!

To be entirely clear... since our "official" progress report we discovered bugs and further analysis showed other flaws in the tyre model. Further development of the tyre model and further analysis showed up more issues as well. This is that model which we think feels very nice to drive in most sutuations. I am not interested in "fudging" or "bodging" the tyre model and it takes time, research and thinking space to come to the right solutions.

Around Christmas and New Year, time spent with family and so on also causes some delays. I don't regret that.

My post was supposed to let you know what's going on, after dozens of people were complaining about a wall of silence, developers hiding under a rock and all that.

I hope you now have a better understanding of what I was saying. Some guys, please don't try to take the worst possible view of everything. Most people did understand!

Hi there.

I'll "break the silence"... though I do think I said something quite recently.

There's really nothing to report, that's why we don't report anything. I've been continuing with the tyre physics. It's a mathematically based model as you know. That involves assumptions and simplifications to make it workable. During testing we found some extreme situations where the assumptions used caused a breakdown of the output forces. It turned out that the extreme situations can actually come up quite often. I don't really want to go into detail, I've been working on the model and studying the situations where things go wrong.

My work rate has not been very high these few months, I admit. For a while I seemed to fix something, it started to look good then on further investigation and analysis, another problem showed up. That kind of stumbling block caused me to take a break while I tried to get my head around it, to try another approach. The new approach might work or might not and I needed to think again. In those thinking times I didn't just sit here pushing myself until i came up with answers, instead I just sorted out other things unrelated to LFS. At this point in life in am not motivated to sort of sit there and pull my hair out and bang my head against a wall until the solution appears, I'd prefer to just do something else, running the problem over in my mind until I know what do do. That way I can make progress gradually without the frustration that can be experienced by someone who is stuck but needs to produce solutions in a given time.

It's not worth releasing a model with fatal flaws that make it worse than the current model in some situations, even though the on-track normal driving feel is better than the old tyre model.

Unfortunately the slow progress with the physics does delay the release of the Scirocco and Rockingham, but it can't really be helped. What I really want to do is get those improvements out there and then work on some more fun things I've been thinking about. You'll understand I can't comment on future things that may or may not happen, but there are things I want to do but can't until this hard stuff is out of the way.

About slow progress... some people comment how progress is so unbelievably and ridiculously slow... well I guess that means there must be a whole lot of sims out there that have overtaken LFS in nearly every way. If so, you can use them for a while, that's why we have an email notification when a patch is released. If people "leave" LFS that is fine. That kind of thing really can't be something I can get worried about. All I can do is work on the things I need to do, to get LFS to a level we are happy with, then release the updates. Sometimes that is delayed by obstacles as I described before, and sometimes by other things that come up in life generally. With two children and a house to look after, things that go wrong, guests that visit occasionally and that kind of thing, it's true that I don't do the same hours that I used to in the early days, when LFS was my hobby, my passion and my work all in one. Now it's more my work, I have other interests as well and there are now other things in life that are more important, but LFS is something I want to be proud of. Something that I want out there and being enjoyed.

Anyway, we have not died and we are not sipping cocktails on a tropical beach. Actually we are working on LFS. I'm developing the physics, Victor is keep things running, Eric is working on things too. We've done a load of work that we want to release, partly for the excitement of it, but we also have mortgages and other bills to pay and the releases will increase sales, so financially it would also be a great idea to get it out there!

Nothing else I can do do really, I'm just trying to get my current LFS projects finished and released. Threats of people "leaving" LFS, predictions of impending doom and so on don't really make any difference. I'm not about to release unfinished stuff and I will still get plently of sleep and relaxation time, and look after the other things I need to look after and develop LFS physics in whatever time it takes. Then we'll get there in the end and we'll send a notification email when the patch is ready, so you can all have a go on an improved LFS with more content.