torcs-users Mailing List for TORCS - The Open Racing Car Simulator

Dear Muhammet Ozcetin <tun...@ho...>,=0A =0A=0AUnsubscribing fr=
om the Torcs-users mailing list is easy, and if you have any problems unsub=
scribing you can contact the list administrator at torcs-users-owner@lists.=
sourceforge.net, so calm-down and please cut down on the "!".=0A=0A=0AThe m=
ain way of unsubscribing is to go to your subscription page, your's is:=0A=
=0Ahttps://lists.sourceforge.net/lists/options/torcs-users/tuning%2Dnitro%4=
0hotmail.com=0A=0A=0AAlso, when you send emails to mailing lists, it is pre=
ferred that you send them in plain text.=0A=0A=0A-Timothy=0A =0A=0A----- Or=
iginal Message ---- =0A=0AFrom: Muhammet Ozcetin <tun...@ho...>=
 =0A=0ATo: tor...@li... =0A=0ASent: Saturday, October =
27, 2007 3:53:00 AM =0A=0ASubject: Re: [Torcs-users] Torcs-users Digest, Vo=
l 17, Issue 9 =0A=0A =0A=0A    .hmmessage P { margin:0px;padding:0px;} body=
.hmmessage { FONT-SIZE:10pt;FONT-FAMILY:Tahoma;}    =0A=0A can you stop wit=
h mailing me ????!!!!!!!!!!!!!! =0A=0A   =0A=0A Tank you !!!!!!!!!!!!!!!!!!=
!!!!!!! =0A=0A =0A=0A =0A=0A =0A=0A =0A=0A    =0A=0A > From: torcs-users-re=
qu...@li... =0A=0A> Subject: Torcs-users Digest, Vol 17, Is=
sue 9 =0A=0A> To: tor...@li... =0A=0A> Date: Thu, 25 O=
ct 2007 12:13:29 -0700 =0A=0A>  =0A=0A> Send Torcs-users mailing list submi=
ssions to =0A=0A> tor...@li... =0A=0A>  =0A=0A> To sub=
scribe or unsubscribe via the World Wide Web, visit =0A=0A> https://lists.s=
ourceforge.net/lists/listinfo/torcs-users =0A=0A> or, via email, send a mes=
sage with subject or body 'help' to =0A=0A> tor...@li...=
eforge.net =0A=0A>  =0A=0A> You can reach the person managing the list at =
=0A=0A> tor...@li... =0A=0A>  =0A=0A> When reply=
ing, please edit your Subject line so it is more specific =0A=0A> than "Re:=
 Contents of Torcs-users digest..." =0A=0A>  =0A=0A>  =0A=0A> Today's Topic=
s: =0A=0A>  =0A=0A> 1. Re: how does torcs determine if a car should be draw=
n =0A=0A> onscreen (kC) =0A=0A> 2. Re: how does torcs determine if a car sh=
ould be drawn =0A=0A> onscreen (Mart Kelder) =0A=0A>  =0A=0A>  =0A=0A> ----=
------------------------------------------------------------------ =0A=0A> =
 =0A=0A> Message: 1 =0A=0A> Date: Thu, 25 Oct 2007 05:42:23 -0700 (PDT) =0A=
=0A> From: kC <yak...@ya...> =0A=0A> Subject: Re: [Torcs-users] how =
does torcs determine if a car should be =0A=0A> drawn  To: =
Mart Kelder <ma...@ke...>, tor...@li... =0A=0A> M=
essage-ID: <313...@we...> =0A=0A> Content-Type: =
text/plain; charset=3D"iso-8859-1" =0A=0A>  =0A=0A> thanks mart for the inf=
ormation. the reason i need to determine which cars are in view is because =
i need to highlight those cars that are in view. =0A=0A>  =0A=0A> how do i =
determine the distance away, and the angle of the cars? i'm quite lost in t=
orcs. i know i can use car->sx, car->sy to determine the locations of the c=
ars with respect to the track coordinates, but as i need to highlight them =
(using circles) from a cam perspective, how do i translate the x,y,z cam co=
ordinates? =0A=0A>  =0A=0A> thanks once again for helping a mate out! =0A=
=0A>  =0A=0A> Mart Kelder <ma...@ke...> wrote: Hello kC (and others), =
=0A=0A>  =0A=0A> Op Tuesday 23 October 2007 14:37:03 schreef kC: =0A=0A> > =
Thanks Mart for replying. =0A=0A> > =0A=0A> > the problem i'm facing is tha=
t i've to determine what cars are being shown =0A=0A> > on the camera (i'm =
only dealing with the in-car camera), and if that car is =0A=0A> > shown, i=
'll highlight information such as the speed of that car etc. i'm =0A=0A> > =
clueless how is the camera view constructed, and how which cars are drawn =
=0A=0A> > on the camera scene. =0A=0A>  =0A=0A> I don't get why you want to=
 know which cars are drawn. =0A=0A>  =0A=0A> If you want to display the inf=
ormation above the cars. Then you can put that  =0A=0A> information on all =
the cars. For all cars you know how far it is away from  =0A=0A> the object=
 car (you can adjust the font size on that such that it is readable  =0A=0A=
> on larger distances, and you do know the angle (so you can make the  =0A=
=0A> orientation such that the viewing area is maximal).  =0A=0A>  =0A=0A> =
If you want to show it in some kind of 2d-leaderboard, you only need to kno=
w  =0A=0A> which cars is in front and how far it is in front. If that car i=
s displayed  =0A=0A> or not does not change the value of that data being sh=
owed. It is possible to  =0A=0A> see for each car how far it is in front. =
=0A=0A>  =0A=0A> To answer the question: I think it is quite hard to determ=
ine that, because  =0A=0A> the cars are drawn in the scene. The scene itsel=
f it's drawn by plib (IIRC).  =0A=0A> The camera also is transferred to pli=
b. For plib, there isn't much difference  =0A=0A> between a car and a wall,=
 so it will proberbly be pretty hard to get  =0A=0A> information from that.=
 =0A=0A>  =0A=0A> The camera's are defined in src/modules/graphic/ssggraph/=
grcam.cpp. The  =0A=0A> screens are drawn in src/modules/graphic/ssggraph/g=
rscreen.cpp. Information  =0A=0A> about plib can be found at [1]. =0A=0A>  =
=0A=0A> > i'm using MS Visual C++ 6, and i can't really find the links betw=
een =0A=0A> > different files as they're not in the class wizard. when i tr=
ied to =0A=0A> > construct the class wizard file, there's still nothing in =
them. =0A=0A>  =0A=0A> I do not use MS Visual C++, so I can't really say a =
lot about it. But it  =0A=0A> sounds strange to me that it is possible to c=
ompile it, but not get the  =0A=0A> information about the objects used. Giv=
en the source files (which are given,  =0A=0A> elsewise it was not possible=
 to compile), it isn't too hard to collect  =0A=0A> structs, functions, cla=
sses, ... . It is certainly easier then compiling it  =0A=0A> (and it can d=
o that). =0A=0A>  =0A=0A> > this is not a major issue, since i can still se=
arch for the files in the =0A=0A> > directory manually, but it's just more =
painful. =0A=0A> > =0A=0A> > oh, lastly, is it possible to use debugging mo=
de for torcs? i tried but the =0A=0A> > MSVC++6 always stopped before the o=
pengl window is drawn. =0A=0A>  =0A=0A> As above: I don't use MS Visual C++=
. =0A=0A>  =0A=0A> Regards, =0A=0A>  =0A=0A> Mart =0A=0A>  =0A=0A> [1] http=
://plib.sourceforge.net/ =0A=0A>  =0A=0A> ---------------------------------=
---------------------------------------- =0A=0A> This SF.net email is spons=
ored by: Splunk Inc. =0A=0A> Still grepping through log files to find probl=
ems? Stop. =0A=0A> Now Search log events and configuration files using AJAX=
 and a browser. =0A=0A> Download your FREE copy of Splunk now >> http://get=
.splunk.com/ =0A=0A> _______________________________________________ =0A=0A=
> Torcs-users mailing list =0A=0A> Tor...@li... =0A=0A=
> https://lists.sourceforge.net/lists/listinfo/torcs-users =0A=0A>  =0A=0A>=
  =0A=0A>  =0A=0A>  =0A=0A> _______________________________________________=
___ =0A=0A> Do You Yahoo!? =0A=0A> Tired of spam? Yahoo! Mail has the best =
spam protection around  =0A=0A> http://mail.yahoo.com  =0A=0A> ------------=
-- next part -------------- =0A=0A> An HTML attachment was scrubbed... =0A=
=0A>  =0A=0A> ------------------------------ =0A=0A>  =0A=0A> Message: 2 =
=0A=0A> Date: Thu, 25 Oct 2007 16:52:38 +0200 =0A=0A> From: Mart Kelder <ma=
rt...@ke...> =0A=0A> Subject: Re: [Torcs-users] how does torcs determine=
 if a car should be =0A=0A> drawn  To: kC <yakumo_sg@yahoo.=
com> =0A=0A> Cc: tor...@li... =0A=0A> Message-ID: <200=
710...@ke...> =0A=0A> Content-Type: text/plain; charset=
=3D"iso-8859-1" =0A=0A>  =0A=0A> Hello kC (and others), =0A=0A>  =0A=0A> Op=
 Thursday 25 October 2007 14:42:23 schreef kC: =0A=0A> > thanks mart for th=
e information. the reason i need to determine which cars =0A=0A> > are in v=
iew is because i need to highlight those cars that are in view. =0A=0A>  =
=0A=0A> You can make an estimate guess about if a car is visible of not. If=
 a car if  =0A=0A> next to you, it is certain that it isn't draw (if isn't =
in front and it isn't  =0A=0A> in the mirror). So you can reduce the number=
 of highlighted cars by looking  =0A=0A> to the angle of the car in respect=
 to the direction of the car. =0A=0A>  =0A=0A> Also, if the car is 2 km in =
front and not straight before your car, it also is  =0A=0A> very likely tha=
t it isn't drawn. =0A=0A>  =0A=0A> However, these kind of restriction is ha=
rd to get perfect.  =0A=0A>  =0A=0A> > how do i determine the distance away=
, and the angle of the cars? i'm quite =0A=0A> > lost in torcs. i know i ca=
n use car->sx, car->sy to determine the locations =0A=0A> > of the cars wit=
h respect to the track coordinates, but as i need to =0A=0A> > highlight th=
em (using circles) from a cam perspective, how do i translate =0A=0A> > the=
 x,y,z cam coordinates? =0A=0A>  =0A=0A> The coordinates of the track are e=
nough to give a distance and an angle. Let  =0A=0A> car be the (tCarElt*) o=
f the focused car and ocar the (tCarElt*) of the car  =0A=0A> you want to g=
et information about. =0A=0A>  =0A=0A> You can find out the position of car=
 with: car->_pos_X, car->_pos_Y and  =0A=0A> car->_pos_Z. This also for oca=
r. The distance between two points in the  =0A=0A> three-dimentional real s=
pace (of course it isn't a real space, but it is  =0A=0A> close enough), ca=
n be calculated by substracting the two vectors from each  =0A=0A> other an=
d then take the norm. So, in practice: =0A=0A>  =0A=0A> dist =3D sqrt(sqr(c=
ar->_pos_X - ocar->_pos_X)+sqr(car->_pos_Y -  =0A=0A> ocar->_pos_Y)+sqr(car=
->_pos_Z - ocar->_pos_Z)); =0A=0A>  =0A=0A> sqr is defined as #define sqr(r=
) (r)*(r). You can leave out the sqrt; then you  =0A=0A> get the square of =
the distance. =0A=0A>  =0A=0A> For the angle, you just have to project  the vectors to make a  =0A=0A> two-dimentional vector space. Then can ea=
sily be done by ignoring the  =0A=0A> Z-value. Then you consider the line t=
hough car (car->_pos_X and car->_pos_Y)  =0A=0A> and ocar. Now you make a s=
econd line which is perpendicular to that line (you  =0A=0A> can do that by=
 rotating the line 1/2*pi radials around the z-axis. Then you  =0A=0A> tran=
slate the line such that it goes through ocar. Now make it a plane by  =0A=
=0A> adding the z-axis. Your information should now be printed in that plan=
e (it  =0A=0A> has the maximal viewing angle). Of course, you must take car=
e of that you  =0A=0A> draw on that part of the plane that is close to the =
car (and not below the  =0A=0A> track). =0A=0A>  =0A=0A> The place of the c=
amera can slightly change from the position of the car. To  =0A=0A> get the=
 real camera position, you can read in the code how that position is  =0A=
=0A> calculated. It proberbly depends on the xml-file of the car type. =0A=
=0A>  =0A=0A> > thanks once again for helping a mate out! =0A=0A>  =0A=0A> =
Regards, =0A=0A>  =0A=0A> Mart =0A=0A>  =0A=0A>  =0A=0A>  =0A=0A> ---------=
--------------------- =0A=0A>  =0A=0A> ------------------------------------=
------------------------------------- =0A=0A> This SF.net email is sponsore=
d by: Splunk Inc. =0A=0A> Still grepping through log files to find problems=
? Stop. =0A=0A> Now Search log events and configuration files using AJAX an=
d a browser. =0A=0A> Download your FREE copy of Splunk now >> http://get.sp=
lunk.com/ =0A=0A>  =0A=0A> ------------------------------ =0A=0A>  =0A=0A> =
_______________________________________________ =0A=0A> Torcs-users mailing=
 list =0A=0A> Tor...@li... =0A=0A> https://lists.sourc=
eforge.net/lists/listinfo/torcs-users =0A=0A>  =0A=0A>  =0A=0A> End of Torc=
s-users Digest, Vol 17, Issue 9 =0A=0A> ***********************************=
******* =0A=0A =0A=0A =0A=0AWindows Live Messenger het beste van de toekoms=
t Download NU! Windows Live Messenger! =0A=0A =0A=0A =0A=0A=0A=0A=0A=0A=0A_=
_________________________________________________=0ADo You Yahoo!?=0ATired =
of spam?  Yahoo! Mail has the best spam protection around =0Ahttp://mail.ya=
hoo.com 

=20
can you stop with mailing me ????!!!!!!!!!!!!!!
=20
Tank you !!!!!!!!!!!!!!!!!!!!!!!!!



> From: tor...@li...> Subject: Torcs-users Dig=
est, Vol 17, Issue 9> To: tor...@li...> Date: Thu, 25 =
Oct 2007 12:13:29 -0700> > Send Torcs-users mailing list submissions to> to=
rcs...@li...> > To subscribe or unsubscribe via the Worl=
d Wide Web, visit> https://lists.sourceforge.net/lists/listinfo/torcs-users=
> or, via email, send a message with subject or body 'help' to> torcs-users=
-re...@li...> > You can reach the person managing the lis=
t at> tor...@li...> > When replying, please edit=
 your Subject line so it is more specific> than "Re: Contents of Torcs-user=
s digest..."> > > Today's Topics:> > 1. Re: how does torcs determine if a c=
ar should be drawn> onscreen (kC)> 2. Re: how does torcs determine if a car=
 should be drawn> onscreen (Mart Kelder)> > > -----------------------------=
-----------------------------------------> > Message: 1> Date: Thu, 25 Oct =
2007 05:42:23 -0700 (PDT)> From: kC <yak...@ya...>> Subject: Re: [To=
rcs-users] how does torcs determine if a car should be> drawn onscreen> To:=
 Mart Kelder <ma...@ke...>, tor...@li...> Message=
-ID: <313...@we...>> Content-Type: text/plain; c=
harset=3D"iso-8859-1"> > thanks mart for the information. the reason i need=
 to determine which cars are in view is because i need to highlight those c=
ars that are in view.> > how do i determine the distance away, and the angl=
e of the cars? i'm quite lost in torcs. i know i can use car->sx, car->sy t=
o determine the locations of the cars with respect to the track coordinates=
, but as i need to highlight them (using circles) from a cam perspective, h=
ow do i translate the x,y,z cam coordinates?> > thanks once again for helpi=
ng a mate out!> > Mart Kelder <ma...@ke...> wrote: Hello kC (and other=
s),> > Op Tuesday 23 October 2007 14:37:03 schreef kC:> > Thanks Mart for r=
eplying.> >> > the problem i'm facing is that i've to determine what cars a=
re being shown> > on the camera (i'm only dealing with the in-car camera), =
and if that car is> > shown, i'll highlight information such as the speed o=
f that car etc. i'm> > clueless how is the camera view constructed, and how=
 which cars are drawn> > on the camera scene.> > I don't get why you want t=
o know which cars are drawn.> > If you want to display the information abov=
e the cars. Then you can put that > information on all the cars. For all ca=
rs you know how far it is away from > the object car (you can adjust the fo=
nt size on that such that it is readable > on larger distances, and you do =
know the angle (so you can make the > orientation such that the viewing are=
a is maximal). > > If you want to show it in some kind of 2d-leaderboard, y=
ou only need to know > which cars is in front and how far it is in front. I=
f that car is displayed > or not does not change the value of that data bei=
ng showed. It is possible to > see for each car how far it is in front.> > =
To answer the question: I think it is quite hard to determine that, because=
 > the cars are drawn in the scene. The scene itself it's drawn by plib (II=
RC). > The camera also is transferred to plib. For plib, there isn't much d=
ifference > between a car and a wall, so it will proberbly be pretty hard t=
o get > information from that.> > The camera's are defined in src/modules/g=
raphic/ssggraph/grcam.cpp. The > screens are drawn in src/modules/graphic/s=
sggraph/grscreen.cpp. Information > about plib can be found at [1].> > > i'=
m using MS Visual C++ 6, and i can't really find the links between> > diffe=
rent files as they're not in the class wizard. when i tried to> > construct=
 the class wizard file, there's still nothing in them.> > I do not use MS V=
isual C++, so I can't really say a lot about it. But it > sounds strange to=
 me that it is possible to compile it, but not get the > information about =
the objects used. Given the source files (which are given, > elsewise it wa=
s not possible to compile), it isn't too hard to collect > structs, functio=
ns, classes, ... . It is certainly easier then compiling it > (and it can d=
o that).> > > this is not a major issue, since i can still search for the f=
iles in the> > directory manually, but it's just more painful.> >> > oh, la=
stly, is it possible to use debugging mode for torcs? i tried but the> > MS=
VC++6 always stopped before the opengl window is drawn.> > As above: I don'=
t use MS Visual C++.> > Regards,> > Mart> > [1] http://plib.sourceforge.net=
/> > ----------------------------------------------------------------------=
---> This SF.net email is sponsored by: Splunk Inc.> Still grepping through=
 log files to find problems? Stop.> Now Search log events and configuration=
 files using AJAX and a browser.> Download your FREE copy of Splunk now >> =
http://get.splunk.com/> _______________________________________________> To=
rcs-users mailing list> Tor...@li...> https://lists.so=
urceforge.net/lists/listinfo/torcs-users> > > > > _________________________=
_________________________> Do You Yahoo!?> Tired of spam? Yahoo! Mail has t=
he best spam protection around > http://mail.yahoo.com > -------------- nex=
t part --------------> An HTML attachment was scrubbed...> > --------------=
----------------> > Message: 2> Date: Thu, 25 Oct 2007 16:52:38 +0200> From=
: Mart Kelder <ma...@ke...>> Subject: Re: [Torcs-users] how does torcs=
 determine if a car should be> drawn onscreen> To: kC <yak...@ya...>=
> Cc: tor...@li...> Message-ID: <200710251652.39233.ma=
rt...@ke...>> Content-Type: text/plain; charset=3D"iso-8859-1"> > Hello =
kC (and others),> > Op Thursday 25 October 2007 14:42:23 schreef kC:> > tha=
nks mart for the information. the reason i need to determine which cars> > =
are in view is because i need to highlight those cars that are in view.> > =
You can make an estimate guess about if a car is visible of not. If a car i=
f > next to you, it is certain that it isn't draw (if isn't in front and it=
 isn't > in the mirror). So you can reduce the number of highlighted cars b=
y looking > to the angle of the car in respect to the direction of the car.=
> > Also, if the car is 2 km in front and not straight before your car, it =
also is > very likely that it isn't drawn.> > However, these kind of restri=
ction is hard to get perfect. > > > how do i determine the distance away, a=
nd the angle of the cars? i'm quite> > lost in torcs. i know i can use car-=
>sx, car->sy to determine the locations> > of the cars with respect to the =
track coordinates, but as i need to> > highlight them (using circles) from =
a cam perspective, how do i translate> > the x,y,z cam coordinates?> > The =
coordinates of the track are enough to give a distance and an angle. Let > =
car be the (tCarElt*) of the focused car and ocar the (tCarElt*) of the car=
 > you want to get information about.> > You can find out the position of c=
ar with: car->_pos_X, car->_pos_Y and > car->_pos_Z. This also for ocar. Th=
e distance between two points in the > three-dimentional real space (of cou=
rse it isn't a real space, but it is > close enough), can be calculated by =
substracting the two vectors from each > other and then take the norm. So, =
in practice:> > dist =3D sqrt(sqr(car->_pos_X - ocar->_pos_X)+sqr(car->_pos=
_Y - > ocar->_pos_Y)+sqr(car->_pos_Z - ocar->_pos_Z));> > sqr is defined as=
 #define sqr(r) (r)*(r). You can leave out the sqrt; then you > get the squ=
are of the distance.> > For the angle, you just have to project one of the =
vectors to make a > two-dimentional vector space. Then can easily be done b=
y ignoring the > Z-value. Then you consider the line though car (car->_pos_=
X and car->_pos_Y) > and ocar. Now you make a second line which is perpendi=
cular to that line (you > can do that by rotating the line 1/2*pi radials a=
round the z-axis. Then you > translate the line such that it goes through o=
car. Now make it a plane by > adding the z-axis. Your information should no=
w be printed in that plane (it > has the maximal viewing angle). Of course,=
 you must take care of that you > draw on that part of the plane that is cl=
ose to the car (and not below the > track).> > The place of the camera can =
slightly change from the position of the car. To > get the real camera posi=
tion, you can read in the code how that position is > calculated. It prober=
bly depends on the xml-file of the car type.> > > thanks once again for hel=
ping a mate out!> > Regards,> > Mart> > > > ------------------------------>=
 > ------------------------------------------------------------------------=
-> This SF.net email is sponsored by: Splunk Inc.> Still grepping through l=
og files to find problems? Stop.> Now Search log events and configuration f=
iles using AJAX and a browser.> Download your FREE copy of Splunk now >> ht=
tp://get.splunk.com/> > ------------------------------> > _________________=
______________________________> Torcs-users mailing list> Torcs-users@lists=
.sourceforge.net> https://lists.sourceforge.net/lists/listinfo/torcs-users>=
 > > End of Torcs-users Digest, Vol 17, Issue 9> **************************=
****************
_________________________________________________________________
http://www.live.com/?mkt=3Dnl-nl
Live.nl=

Hello kC (and others),

Op Thursday 25 October 2007 14:42:23 schreef kC:
> thanks mart for the information. the reason i need to determine which cars
> are in view is because i need to highlight those cars that are in view.

You can make an estimate guess about if a car is visible of not. If a car if 
next to you, it is certain that it isn't draw (if isn't in front and it isn't 
in the mirror). So you can reduce the number of highlighted cars by looking 
to the angle of the car in respect to the direction of the car.

Also, if the car is 2 km in front and not straight before your car, it also is 
very likely that it isn't drawn.

However, these kind of restriction is hard to get perfect. 

> how do i determine the distance away, and the angle of the cars? i'm quite
> lost in torcs. i know i can use car->sx, car->sy to determine the locations
> of the cars with respect to the track coordinates, but as i need to
> highlight them (using circles) from a cam perspective, how do i translate
> the x,y,z cam coordinates?

The coordinates of the track are enough to give a distance and an angle. Let 
car be the (tCarElt*) of the focused car and ocar the (tCarElt*) of the car 
you want to get information about.

You can find out the position of car with: car->_pos_X, car->_pos_Y and 
car->_pos_Z. This also for ocar. The distance between two points in the 
three-dimentional real space (of course it isn't a real space, but it is 
close enough), can be calculated by substracting the two vectors from each 
other and then take the norm. So, in practice:

dist = sqrt(sqr(car->_pos_X - ocar->_pos_X)+sqr(car->_pos_Y - 
ocar->_pos_Y)+sqr(car->_pos_Z - ocar->_pos_Z));

sqr is defined as #define sqr(r) (r)*(r). You can leave out the sqrt; then you 
get the square of the distance.

For the angle, you just have to project one of the vectors to make a 
two-dimentional vector space. Then can easily be done by ignoring the 
Z-value. Then you consider the line though car (car->_pos_X and car->_pos_Y) 
and ocar. Now you make a second line which is perpendicular to that line (you 
can do that by rotating the line 1/2*pi radials around the z-axis. Then you 
translate the line such that it goes through ocar. Now make it a plane by 
adding the z-axis. Your information should now be printed in that plane (it 
has the maximal viewing angle). Of course, you must take care of that you 
draw on that part of the plane that is close to the car (and not below the 
track).

The place of the camera can slightly change from the position of the car. To 
get the real camera position, you can read in the code how that position is 
calculated. It proberbly depends on the xml-file of the car type.

> thanks once again for helping a mate out!

Regards,

Mart

thanks mart for the information. the reason i need to determine which cars are in view is because i need to highlight those cars that are in view.

how do i determine the distance away, and the angle of the cars? i'm quite lost in torcs. i know i can use car->sx, car->sy to determine the locations of the cars with respect to the track coordinates, but as i need to highlight them (using circles) from a cam perspective, how do i translate the x,y,z cam coordinates?

thanks once again for helping a mate out!

Mart Kelder <ma...@ke...> wrote: Hello kC (and others),

Op Tuesday 23 October 2007 14:37:03 schreef kC:
> Thanks Mart for replying.
>
> the problem i'm facing is that i've to determine what cars are being shown
> on the camera (i'm only dealing with the in-car camera), and if that car is
> shown, i'll highlight information such as the speed of that car etc. i'm
> clueless how is the camera view constructed, and how which cars are drawn
> on the camera scene.

I don't get why you want to know which cars are drawn.

If you want to display the information above the cars. Then you can put that 
information on all the cars. For all cars you know how far it is away from 
the object car (you can adjust the font size on that such that it is readable 
on larger distances, and you do know the angle (so you can make the 
orientation such that the viewing area is maximal). 

If you want to show it in some kind of 2d-leaderboard, you only need to know 
which cars is in front and how far it is in front. If that car is displayed 
or not does not change the value of that data being showed. It is possible to 
see for each car how far it is in front.

To answer the question: I think it is quite hard to determine that, because 
the cars are drawn in the scene. The scene itself it's drawn by plib (IIRC). 
The camera also is transferred to plib. For plib, there isn't much difference 
between a car and a wall, so it will proberbly be pretty hard to get 
information from that.

The camera's are defined in src/modules/graphic/ssggraph/grcam.cpp. The 
screens are drawn in src/modules/graphic/ssggraph/grscreen.cpp. Information 
about plib can be found at [1].

> i'm using MS Visual C++ 6, and i can't really find the links between
> different files as they're not in the class wizard. when i tried to
> construct the class wizard file, there's still nothing in them.

I do not use MS Visual C++, so I can't really say a lot about it. But it 
sounds strange to me that it is possible to compile it, but not get the 
information about the objects used. Given the source files (which are given, 
elsewise it was not possible to compile), it isn't too hard to collect 
structs, functions, classes, ... . It is certainly easier then compiling it 
(and it can do that).

> this is not a major issue, since i can still search for the files in the
> directory manually, but it's just more painful.
>
> oh, lastly, is it possible to use debugging mode for torcs? i tried but the
> MSVC++6 always stopped before the opengl window is drawn.

As above: I don't use MS Visual C++.

Regards,

Mart

[1] http://plib.sourceforge.net/

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Torcs-users mailing list
Tor...@li...
https://lists.sourceforge.net/lists/listinfo/torcs-users




 __________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

Hello kC (and others),

Op Tuesday 23 October 2007 14:37:03 schreef kC:
> Thanks Mart for replying.
>
> the problem i'm facing is that i've to determine what cars are being shown
> on the camera (i'm only dealing with the in-car camera), and if that car is
> shown, i'll highlight information such as the speed of that car etc. i'm
> clueless how is the camera view constructed, and how which cars are drawn
> on the camera scene.

I don't get why you want to know which cars are drawn.

If you want to display the information above the cars. Then you can put that 
information on all the cars. For all cars you know how far it is away from 
the object car (you can adjust the font size on that such that it is readable 
on larger distances, and you do know the angle (so you can make the 
orientation such that the viewing area is maximal). 

If you want to show it in some kind of 2d-leaderboard, you only need to know 
which cars is in front and how far it is in front. If that car is displayed 
or not does not change the value of that data being showed. It is possible to 
see for each car how far it is in front.

To answer the question: I think it is quite hard to determine that, because 
the cars are drawn in the scene. The scene itself it's drawn by plib (IIRC). 
The camera also is transferred to plib. For plib, there isn't much difference 
between a car and a wall, so it will proberbly be pretty hard to get 
information from that.

The camera's are defined in src/modules/graphic/ssggraph/grcam.cpp. The 
screens are drawn in src/modules/graphic/ssggraph/grscreen.cpp. Information 
about plib can be found at [1].

> i'm using MS Visual C++ 6, and i can't really find the links between
> different files as they're not in the class wizard. when i tried to
> construct the class wizard file, there's still nothing in them.

I do not use MS Visual C++, so I can't really say a lot about it. But it 
sounds strange to me that it is possible to compile it, but not get the 
information about the objects used. Given the source files (which are given, 
elsewise it was not possible to compile), it isn't too hard to collect 
structs, functions, classes, ... . It is certainly easier then compiling it 
(and it can do that).

> this is not a major issue, since i can still search for the files in the
> directory manually, but it's just more painful.
>
> oh, lastly, is it possible to use debugging mode for torcs? i tried but the
> MSVC++6 always stopped before the opengl window is drawn.

As above: I don't use MS Visual C++.

Regards,

Mart

[1] http://plib.sourceforge.net/

Thanks Mart for replying.

Mart Kelder <ma...@ke...> wrote: Hello kC (and others),

Op Monday 22 October 2007 04:23:19 schreef kC:
> As the topic states, i need to know how does torcs determine if a
> particular should be drawn on the screen. So far, I've only found the
> coordinates of a car with respect to the track. I quite certain that it
> depends on the camera and cars' position. However, I'm quite lost with how
> to get this information.

AFAIK, there are two ways to get something in the screen. One is putting it 
somewhere in the 3d-world (such as is done with cars), and the other is 
putting it directly on the screen (such as done with gauches, leaderboard, 
etc.). The 3d-world (scene) is converted to what is shown on the display by 
the library which is used (IIRC).

the problem i'm facing is that i've to determine what cars are being shown on the camera (i'm only dealing with the in-car camera), and if that car is shown, i'll highlight information such as the speed of that car etc. i'm clueless how is the camera view constructed, and how which cars are drawn on the camera scene.

> Also, I'm using the windows version, so, is there a class wizard file
> somewhere I can use, so that the class wizard works properly? Thanks!

Most of the code is written in C, and C doesn't have classes. Also, I don't 
know how your editor (or IDE) you work with works, but the compiler can get 
all the information out of the available source files, so any IDE should be 
able to do so (if that is a feature of a particular IDE is another story, of 
course).

A seperate file which contains the classes and class properties is very soon 
out of date (probably as soon as the next change is made).

i'm using MS Visual C++ 6, and i can't really find the links between different files as they're not in the class wizard. when i tried to construct the class wizard file, there's still nothing in them.

this is not a major issue, since i can still search for the files in the directory manually, but it's just more painful.

oh, lastly, is it possible to use debugging mode for torcs? i tried but the MSVC++6 always stopped before the opengl window is drawn.

Regards,

Mart

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Torcs-users mailing list
Tor...@li...
https://lists.sourceforge.net/lists/listinfo/torcs-users




 __________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

Hello kC (and others),

Op Monday 22 October 2007 04:23:19 schreef kC:
> As the topic states, i need to know how does torcs determine if a
> particular should be drawn on the screen. So far, I've only found the
> coordinates of a car with respect to the track. I quite certain that it
> depends on the camera and cars' position. However, I'm quite lost with how
> to get this information.

AFAIK, there are two ways to get something in the screen. One is putting it 
somewhere in the 3d-world (such as is done with cars), and the other is 
putting it directly on the screen (such as done with gauches, leaderboard, 
etc.). The 3d-world (scene) is converted to what is shown on the display by 
the library which is used (IIRC).

> Also, I'm using the windows version, so, is there a class wizard file
> somewhere I can use, so that the class wizard works properly? Thanks!

Most of the code is written in C, and C doesn't have classes. Also, I don't 
know how your editor (or IDE) you work with works, but the compiler can get 
all the information out of the available source files, so any IDE should be 
able to do so (if that is a feature of a particular IDE is another story, of 
course).

A seperate file which contains the classes and class properties is very soon 
out of date (probably as soon as the next change is made).

Regards,

Mart

As the topic states, i need to know how does torcs determine if a particular should be drawn on the screen. So far, I've only found the coordinates of a car with respect to the track. I quite certain that it depends on the camera and cars' position. However, I'm quite lost with how to get this information.

Also, I'm using the windows version, so, is there a class wizard file somewhere I can use, so that the class wizard works properly? Thanks!



 __________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

Hi

 

In the xml-file for each car model an "efficiency" is stated for each gear
in the gearbox. How is this utilized in the model of the transmission?

 (when I print out car->transmission.gearEff the answer is zero)

 

The same question goes for the efficiency of the differential?

 

Kind regards

Thomas

Hello Petter (and others),

Op Friday 12 October 2007 06:48:10 schreef Petter Johansson:
> My name is Petter Johansson, I am a cognitive scientist working at The
> University of Tokyo. I would like to use TORCS in an ongoing project
> of mine, but I have some problems. What I want to achieve is the
> following:
>
> The participant in the experiment drives a car in the simulator. After
> some 20 minutes, when they have learnt to trust the controls and the
> set-up, the steering wheel as well as the car on the screen will be
> manipulated to start turning to one side. Shortly after (in the order
> of milliseconds), something is presented on the screen exactly where
> the car would have been had the participant not steered away.
>
> The question is: will the outcome of the action influence how the
> event is perceived, i.e.  will the fact that the driver avoids to
> collide with the object make them believe the action was actually
> initiated by themselves?
>
> I would like to use TORCS to test this idea, but there are some
> programming difficulties:
>
> 1. how to introduce, control and monitor novel objects (e.g. a little
> girl walking on to the road)
> We looked at using the code for "other cars" and tried to tweak them
> into being "new objects", but it did not seem very promising. Does any
> one know if there is an easier way of doing this, or if there is
> already an implementation somewhere of inserting obstacles on the road
> that the drivers need to avoid?

A robot can't put it's car suddenly on a random place on the map, but the 
raceengingeclient can. However, I don't know how good it will actually work 
and if the collision is still handled.

There also is another approce. You can change the display by adding the object 
when it is drawn. This can be done in src/modules/graphic/ssggraph/. Maybe a 
2d object with texture is sufficient.

With just drawing, collision is not provided. But you can also change the 
collision code (it is in src/modules/simu/simuv2/collide.cpp).

Of course, you need some variables in some interface to make variables 
available to all objects you change. The interfaces are in src/interfaces.

> 2. how to override and manipulate the actual steering of the wheel.

You can change some files in src/drivers/human. The human driver is just a 
robot like all the computer programmed robot, except that it based it actions 
on the input of the user.

> all help much appreciated,
> Petter

Regards,

Mart

My name is Petter Johansson, I am a cognitive scientist working at The
University of Tokyo. I would like to use TORCS in an ongoing project
of mine, but I have some problems. What I want to achieve is the
following:

The participant in the experiment drives a car in the simulator. After
some 20 minutes, when they have learnt to trust the controls and the
set-up, the steering wheel as well as the car on the screen will be
manipulated to start turning to one side. Shortly after (in the order
of milliseconds), something is presented on the screen exactly where
the car would have been had the participant not steered away.

The question is: will the outcome of the action influence how the
event is perceived, i.e.  will the fact that the driver avoids to
collide with the object make them believe the action was actually
initiated by themselves?

I would like to use TORCS to test this idea, but there are some
programming difficulties:

1. how to introduce, control and monitor novel objects (e.g. a little
girl walking on to the road)
We looked at using the code for "other cars" and tried to tweak them
into being "new objects", but it did not seem very promising. Does any
one know if there is an easier way of doing this, or if there is
already an implementation somewhere of inserting obstacles on the road
that the drivers need to avoid?

2. how to override and manipulate the actual steering of the wheel.

all help much appreciated,
Petter

On Thu, 11 Oct 2007 07:38:24 +0200
Stephan Kempkes <ske...@xs...> wrote:

> On Thu, 2007-10-11 at 15:03 +1000, Paul Trevethan wrote:
> > On Thu, 11 Oct 2007 02:50:51 +0200
> > Bernhard Wymann <be...@bl...> wrote:
> > 
> > > Hi Paul
> > > 
> > > > checking for XOpenDisplay in -lX11... no
> > > > configure: error: Can't find libX11. Please check config.log
> > > > and if you can't solve the problem send the file to
> > > > tor...@li... with the subject "torcs
> > > > compilation problem"
> > > 
> > > If the configure script does not find the X11 library or include
> > > files you can pass the location as parameter to the configure
> > > script, example: ./configure --x-libraries=/usr/lib/
> > > ./configure --x-libraries=/usr/lib/ --x-includes=/usr/include/X11/
> > > 
> > > Bye, Bernhard.
> > > 
> > 
> > Hi Bernhard,
> > 
> > I added your suggested parameters and it got me past the original
> > error only to fall over on the following:
> > 
> > checking for XRenderSetSubpixelOrder in -lXrender... no
> > configure: error: Can't find libXrandr. Please check config.log and
> > if you can't solve the problem send the file to
> > tor...@li... with the subject "torcs
> > compilation problem"
> > 
> > I had a look and (same as before) I have the required library I
> > think:
> > 
> > $ locate libXrandr
> > /usr/lib/libXrandr.so.2
> > /usr/lib/libXrandr.so.2.0.0
> > 
> > >From the README I tried the configure again with the
> > >--disable-xrandr
> > and it still falls over; but disabling is different to finding in
> > the 1st place. It has to find it to disable it I guess.
> > 
> > I get a bit lost with all this required library stuff, I can follow
> > it when given instructions but I don't do enough to know how to
> > track down problems too well. I only try it when I really want an
> > app.
> > 
> > I have a .deb file from a Ubuntu distro. I might see if I can
> > "alien" convert it to an rpm that will work. As well, pclinuxos is
> > based on Mandriva so I might be able to find an rpm there that will
> > work. Finding a 1.3 version that is accessible would be nice.
> > 
> > I have no idea why the distro makers left out such a popular game -
> > I have asked them about that.
> > 
> > Thanks for the guidance. Any further thoughts most welcome.
> > 
> > Paul.

> Hi Paul,
> 
> Xrender is probably is missing. I added this to the bug tracker:
> http://sourceforge.net/tracker/index.php?func=detail&aid=1674545&group_id=3777&atid=103777
> 
> Stephan

Sorry Stephan, I moved your quote to the bottom ( I never top post).

Anyway, thanks for that. I did just check and I have the libXrender
files and they are installed in /usr/lib with all the others it seems
to be falling over on. In any case, I realised that my ISP also mirrors
Mandriva so I had a look in the new 2008 repo for that and found their
rpms for Torcs 1.3. I downloaded and installed them and they seem to be
working ok. At least I can get my daily fix of speeding :)

By rights I guess I should try to get the src rpms and rebuild them
to suit this system; but that can wait if these work out ok. I still
have no idea why such a great popular game is not in the repos?

Cheers,
Paul.

Hi Paul,

Xrender is probably is missing. I added this to the bug tracker:
http://sourceforge.net/tracker/index.php?func=detail&aid=1674545&group_id=3777&atid=103777

Stephan


On Thu, 2007-10-11 at 15:03 +1000, Paul Trevethan wrote:
> On Thu, 11 Oct 2007 02:50:51 +0200
> Bernhard Wymann <be...@bl...> wrote:
> 
> > Hi Paul
> > 
> > > checking for XOpenDisplay in -lX11... no
> > > configure: error: Can't find libX11. Please check config.log and if
> > > you can't solve the problem send the file to
> > > tor...@li... with the subject "torcs
> > > compilation problem"
> > 
> > If the configure script does not find the X11 library or include
> > files you can pass the location as parameter to the configure script,
> > example: ./configure --x-libraries=/usr/lib/
> > ./configure --x-libraries=/usr/lib/ --x-includes=/usr/include/X11/
> > 
> > Bye, Bernhard.
> > 
> 
> Hi Bernhard,
> 
> I added your suggested parameters and it got me past the original error
> only to fall over on the following:
> 
> checking for XRenderSetSubpixelOrder in -lXrender... no
> configure: error: Can't find libXrandr. Please check config.log and if
> you can't solve the problem send the file to
> tor...@li... with the subject "torcs compilation
> problem"
> 
> I had a look and (same as before) I have the required library I think:
> 
> $ locate libXrandr
> /usr/lib/libXrandr.so.2
> /usr/lib/libXrandr.so.2.0.0
> 
> >From the README I tried the configure again with the --disable-xrandr
> and it still falls over; but disabling is different to finding in the
> 1st place. It has to find it to disable it I guess.
> 
> I get a bit lost with all this required library stuff, I can follow it
> when given instructions but I don't do enough to know how to track down
> problems too well. I only try it when I really want an app.
> 
> I have a .deb file from a Ubuntu distro. I might see if I can "alien"
> convert it to an rpm that will work. As well, pclinuxos is based on
> Mandriva so I might be able to find an rpm there that will work.
> Finding a 1.3 version that is accessible would be nice.
> 
> I have no idea why the distro makers left out such a popular game - I
> have asked them about that.
> 
> Thanks for the guidance. Any further thoughts most welcome.
> 
> Paul.
> 
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc.
> Still grepping through log files to find problems?  Stop.
> Now Search log events and configuration files using AJAX and a browser.
> Download your FREE copy of Splunk now >> http://get.splunk.com/
> _______________________________________________
> Torcs-users mailing list
> Tor...@li...
> https://lists.sourceforge.net/lists/listinfo/torcs-users
> 

On Thu, 11 Oct 2007 02:50:51 +0200
Bernhard Wymann <be...@bl...> wrote:

> Hi Paul
>=20
> > checking for XOpenDisplay in -lX11... no
> > configure: error: Can't find libX11. Please check config.log and if
> > you can't solve the problem send the file to
> > tor...@li... with the subject "torcs
> > compilation problem"
>=20
> If the configure script does not find the X11 library or include
> files you can pass the location as parameter to the configure script,
> example: ./configure --x-libraries=3D/usr/lib/
> ./configure --x-libraries=3D/usr/lib/ --x-includes=3D/usr/include/X11/
>=20
> Bye, Bernhard.
>=20

Hi Bernhard,

I added your suggested parameters and it got me past the original error
only to fall over on the following:

checking for XRenderSetSubpixelOrder in -lXrender... no
configure: error: Can't find libXrandr. Please check config.log and if
you can't solve the problem send the file to
tor...@li... with the subject "torcs compilation
problem"

I had a look and (same as before) I have the required library I think:

$ locate libXrandr
/usr/lib/libXrandr.so.2
/usr/lib/libXrandr.so.2.0.0

=46rom the README I tried the configure again with the --disable-xrandr
and it still falls over; but disabling is different to finding in the
1st place. It has to find it to disable it I guess.

I get a bit lost with all this required library stuff, I can follow it
when given instructions but I don't do enough to know how to track down
problems too well. I only try it when I really want an app.

I have a .deb file from a Ubuntu distro. I might see if I can "alien"
convert it to an rpm that will work. As well, pclinuxos is based on
Mandriva so I might be able to find an rpm there that will work.
Finding a 1.3 version that is accessible would be nice.

I have no idea why the distro makers left out such a popular game - I
have asked them about that.

Thanks for the guidance. Any further thoughts most welcome.

Paul.

Hi Paul

> checking for XOpenDisplay in -lX11... no
> configure: error: Can't find libX11. Please check config.log and if you
> can't solve the problem send the file to
> tor...@li... with the subject "torcs compilation
> problem"

If the configure script does not find the X11 library or include files you can
pass the location as parameter to the configure script, example:
./configure --x-libraries=/usr/lib/
./configure --x-libraries=/usr/lib/ --x-includes=/usr/include/X11/

Bye, Bernhard.

-- 

Visit my homepage http://www.berniw.org
Official TORCS racing: The TORCS Racing Board, http://www.berniw.org/trb

I am trying to compile Torcs from source to run under PCLinuxOS 2007. I
have tried all the suggested variations of options and keep falling
over on the following error:

quote.....

checking for XOpenDisplay in -lX11... no
configure: error: Can't find libX11. Please check config.log and if you
can't solve the problem send the file to
tor...@li... with the subject "torcs compilation
problem"

......endquote

The config.log file is attached.

I have libX11 installed. It is in /usr/lib where you would expect, so I
have no idea what this is about?

I really love the game but I have moved to this new distro and no-one
seems to have developed an rpm for it yet.

FYI, as requested.
Paul.

Bernhard,  Thank you,  I finally found that info in the drivers 
instructions all the way at the bottom of the page.   I guess I got 
anxious to drive and didn't finish RTFM.


Bill

Hi William

> NEWB Alert.
> 
> First day on the track,  What are the user controls to keep the view on 
> the car you are driving and not switch to other camera shots?  I found 
> the obvious controls for the car but no other keyboard controls or 
> documentation of same.

Short: During the race select your car if the camera does follow another one
(page Up/Down to scroll through the cars), then press F2 several times to select
a car bound camera. You can get more infos with:

- Pressing F1 for help ingame (context sensitive, do it during the race)
- http://torcs.sourceforge.net/index.php?name=Sections&op=viewarticle&artid=10

Bye, Bernhard.

-- 

Visit my homepage http://www.berniw.org
Official TORCS racing: The TORCS Racing Board, http://www.berniw.org/trb

NEWB Alert.

First day on the track,  What are the user controls to keep the view on 
the car you are driving and not switch to other camera shots?  I found 
the obvious controls for the car but no other keyboard controls or 
documentation of same.

Bill

Hi Thomas

> We are a group of university students utilizing TORCS in a research project
> about platooning. In order to make our own controller which can maintain a
> given inter-vehicular spacing, first of all we need to know the simulation
> model for vehicle behaviors. Documentation on this mathematical-physical
> model immediately doesn't appear anywhere. We would very much appreciate to
> obtain any possible documentation. Does something exist?

No, but the source code is quite compact, have a look into
src/modules/simu/simuv2, function SimCarUpdate.

Kind regards

Bernhard

-- 

Visit my homepage http://www.berniw.org
Official TORCS racing: The TORCS Racing Board, http://www.berniw.org/trb

Hi, 

 

We are a group of university students utilizing TORCS in a research project
about platooning. In order to make our own controller which can maintain a
given inter-vehicular spacing, first of all we need to know the simulation
model for vehicle behaviors. Documentation on this mathematical-physical
model immediately doesn't appear anywhere. We would very much appreciate to
obtain any possible documentation. Does something exist?

 

Kind Regards, 

Thomas