Quote from Dissident :

Here's an idea: create the device as an open helmet which covers the user's head and shoulders.

Whenever you need to apply forces, motors inside the device apply the forces on the user's head using the shoulders as leverage point - you can then simulate front, back, left and right forces.

Bonus if you include headphones inside the helmet for sound, and allow the user to turn his head to change POV.

No bonus if the device snaps the user's spinal chord when hitting the wall at 300 km/h

Probably can be made for under $500. And people can finally justify wearing a helmet while driving a sim

If I remember well, this idea was patented during the 70s or 80s, by the british army. In this patent, they used two pulleys, one on each shoulder, and a helmet with a rope attached on both sides going through the shoulder pulleys. Pulling the ropes pulls the head left and right.

There is somewhere in Japan a very similar device with a formula one cockpit.

Nice idea, but I think it is very dangerous.

Yes you are right after studying about plastic sheet (ABS or Acrylic or even PP) forming, I came back to metal sheet.

This is the best thing for a limited budget, and metal sheet can make the device a lot stronger, while you can not expect plastic sheet to be a strong part of the structure, unless you use very thick sheet that are a nightmare to form.

Thank you very much for voting. It seems such a product could find some customers, if I manage to do it

The funny (or not) thing is that the most difficult part to do is the part with no functionality...the plastic part for the look.

A few people with a low budget can put together mechanical parts without any problem.
They can do electronics too, and have a small PCB produced too.
They can develop drivers.

But try to have a plastic part produced... as soon as you put plastic in your product, you have to buy molds, order thousands parts.

Quote from Shotglass :

but they still deliver very low torque and whats more importantly they dont turn very fast both of which is a real issue with wheels

With a proper reductor, three such motors are enough to turn a wheel in 1/4th of a second and give 10 pounds/5 kg of force on your hands at the same time. This is very strong.

Anyway...from the pictures I see, their power supply is not able to give enough current for three such motors (they require 300 watts!!). I suppose they use the same motors than G25, but this time three of them.

Quote from Shotglass :

this however is just a really bad joke for a 350$ wheel:
Powerful Mabuchi RS 550 Motor as used in RC cars delivers extra strong ForceFeedback effects

Usually, FFB devices use motors which are rated to give less than 30W output (mechanical power). I don't know what motor G25 uses, but G25 power supply can deliver 42watts (electric power).
Small motors are 70% efficient or less -> mechanical power of G25 is less than 30watts.

From the datasheet, RS 550 motors are rated to deliver 80 watts of mechanical power....for each motor!
http://www.mabuchi-motor.co.jp ... log.cgi?CAT_ID=rs_550pcvc

It means that with a proper power supply, this hardware can be several times more powerful than G25.

In fact I doubt this wheel use RS 550 motors...maybe RS 555. RS 550 are way too powerful for such application. 3 such motors require at least a 250 watts power supply to be used properly. RS 550 are not for RC cars, but garden tools and ride-on toys.

In fact I ask this question because it is difficult for me to guess what simracers really want. I would like to have a clearer picture before starting this adventure.

I want to make something affordable...$500 or less if I can. Mechanics, actuators, electronics, software...
It is obvious that I can not use carbon fiber, fancy aluminium profiles, parts labeled with well-known brands.

My first goal is to focus on force feedback functionality and robustness for low price, not look. And I wonder wether this is really what simracers want. Most products designed for simracers focus on look and expensive material used, not price. Carbon fiber addons for G25 with LCD display, fiber glass cockpit shell, hand made wheels, real racing seats, gauges...

Maybe I should not try to make it for $500, and rather use expensive and nice material.

UncleBenny: it would not require a racing seat or a cockpit, but it should work with it too!

Last edited by Juls, Fri, 5 Oct 2007 21:14.

I don't know where to post this, but this is hardware/controller-related.

After searching all possible ways to feel G-forces in racing simulation, I realized that there is no way to get convincing acceleration cues for less than $1000 if you DIY, or $2000 for commercial products. (I do not consider seat vibration is enough). But acceleration cues are badly needed in racing sims.
As I am currently unemployed, and I have some potentially interesting design for a motion simulator device, I would like to ask you:

If one day, there is a device with the following characteristics:
- less than $500
- small, portable, easy to mount/unmount
- does not require a racing seat or a cockpit
- provide very convincing acceleration cues, so you can physically feel the car
- robust
- less fancy-looking than mass-products but not ugly. Rather discrete. (fancy aluminium profiles and materials already cost hundreds of $ alone)
- works with LFS and rFactor first, and later other racing sims, flight simulators, and if possible all games where there is action/movement...like FPS!

would you like to buy such a device?

Last edited by Juls, Fri, 5 Oct 2007 20:03.

Just a little precision concerning FOV.

Let's say I have a 22 inchs, the screen is 30 inchs from my eyes. I use the formula to know what FOV my screen really represents....result: XX degrees.

If I use these XX degrees in the game, what I see on the screen is what I see in a real car if I cover everything with black fabric excepted a 22 inchs wide hole located 30 inchs away from my eyes.

Of course objects are undistorted. But is this realistic? Maybe for a tank driver...

Until we have surround, 180 degrees screens, chosing FOV is still a compromise between:
- an undistorted tank view....not realistic
- a distorted view, with a FOV closer from our real FOV (180 degrees)...not realistic too.

Quote from tristancliffe :

Aha, clever clogs, buy you alluded to a question.

OMG...where did I allude a question?? I already knew how to pit when I posted! This is plain crazy...as I can see you will never admit your attitude is wrong.

Quote from tristancliffe :

If you give up on buying a game because of one person on an internet forum, then you aren't going far in life. What are you going to do when someone says you drove like a fool in a race? How are you going to cope with being crashed out periodically? Just buy the game, enjoy it, and realise that LFS is really a very simple game. There aren't many complexities to 99.9% enjoyment.

There is no enjoyment if I get reactions like yours on the newbie forum. It kills game enjoyment if I know I have something to ask and I can expect to be insulted. Why is it so hard to understand?

We still are in the topic. I wrote pit is not so obvious to find the first time (first post ever in newbie forum), you wrote I am stupid, all people who don't find the pit and dare to ask for it are lazy and/or stupid, I am stupid again, and if I don't appreciate a game where I get flamed for a simple remark on the newbies forum I am not going far in life. No understanding at all, just flaming.

After this experience, you may understand I wonder if it is a good idea to play online with such gentle people. Often forums reflect the online atmosphere...and you are member since 2003.

Obviously your reaction does not represent LFS community, but it seems many people agree with the way new players are treated here as they don't react. And this is significant. I will not forget to share this little experience and this thread when talking about LFS.

This is my last post in this newbie forum, I understood the punishment.
Strange, here a guy asked how to pit (I did not ask)...and that time you did not tell him he is an idiot/lazy.
http://www.lfsforum.net/showthread.php?p=253948#post253948

Last edited by Juls, Wed, 18 Jul 2007 08:37.

Quote from tristancliffe :

Nope, sorry. Anyone with an IQ over 6 could probably work this out for themselves. When we were all pitting noobs (when the S2 Demo came out) we all managed. It was only a few days later that people to lazy/stupid to work stuff out for themselves started posting stupid questions.

But in this case I just think you are trying to defend the fact you asked a legitimate but rather daft question.

Anyone with an IQ over 2 could probably notice I did not start this thread, I did not ask any question here.
Here I just wrote about my experience to say it is not as obvious as it seems and it is a pity there is not even a single sentence in LFS manual about where to pit.
I first tried to stop at several places in pit lane and unfortunately even if I stopped in the yellow marks area I was misplaced and it did not trigger the pit. Then I tried to stop in pitlane and press shift+P, then I tried to find the garage I came from and stop inside or outside. Then I asked other players in game chat, not in forums.

Briliant as you are (and arrogant...engineer?), you may understand that it is totally different when you run a game for the first time and everything is new, and when you already know the game and upgrade to a new version where yellow marks and pit were added.

It is unacceptable to be treated as an idiot for that. And it is unacceptable to treat the author of this thread like that too. Specialy in a forum for newbies. It is even more unacceptable to insist heavily as you do. What's happening there? Newbies have to apologize to ask newbies questions or should not ask anything or they get insulted?

I was about to buy two licences of this game to play with my brother. If buying this game means having to be insulted every time I ask something in the new players forum, I won't do it.

Last edited by Juls, Wed, 18 Jul 2007 07:45.

Quote from tristancliffe :

Yeah, really funny. When pitting was added in the S2 Demo, it took thousands of players almost, what, 2 seconds to work it out for themselves. The though process goes a bit like this:

"Hmmm, I need to pit. I'll go to the pitlane, that sounds promising. Oooh, what are these little yellow marks that sort of make a car sized rectangle. Maybe if I stop in them then the... yep, it works."

However, I think some peoples thought processes go like this:

"Hmmm, I is needing 2b pitted innit. WTF do I is doin now. I nevah bee RTFMing cos I is uber cool dorifto man. I just rite on net chat ruum an hope someone homie helpz mes out innit. boy is i cool, all dah ladeeez will luv me an my totel ingorenz."

I often play other sims. I was expecting some kind of "activity" where I should place my car, like in most other sims. As I did not find anything, I then assumed I had to drive the car in (or in front of) the garage I started from. I saw the yellow lines but how could I assume that stopping on ANY of these places would be enough?
Do you really think this is so intuitive and obvious in a sim supposed to be one of the most realistic ones, when cars can pit anywhere?

Can you imagine in your infinite wisdom and perfect knowledge of the absolute purest English, that even people with a normal IQ following a rational logical process could have some difficulties finding the pit as it is different from most other sims?

Paradoxically enough, you can not stand newbies and their questions, but feel obliged to answer them in an impolite way. Does it bring you an easy satisfaction ?

*And of course, you will certainly answer that it was some humor I did not understand, usual process.*

Last edited by Juls, Tue, 17 Jul 2007 13:43.

Quote from yaper :

You are wrong Juls.
http://www.lfsforum.net/showth ... ontroller+rate#post470880

Nice, this is already a great improvement for smoothing problems, but it does not solve the problem I am talking about. I hope I will manage to explain it...

This improvement in X version means that every time the physics engine updates, it queries the controller state. So controller query, like physics engine, runs *in average* 100 times per second.

But don't forget that if you have 25 frames per second, frame is displayed and then physics/controller query runs 4 times in a raw to keep the 100hz pace.
3 times to simulate what happened during frame update, and 1 time for the new update. And every time it queries for controller state.

So, it means that the 3 physics updates simulating what happened during frame update use controller state AFTER the frame update, not controller state DURING the frame update. Changes in controller state during the frame update are lost in most games (unless in LFS they use another trick, but I doubt controller inputs can be acquired during frame update...but maybe they manage to do it?)->stroboscopic vision.

They did this change (query controller input every physics update) to have consistant smoothing on all computers. rFactor did it too. Simbin games (GTL/GTR/GTR2/Race) still have a framerate dependant controller smoothing...you have to change smoothing coefficient depending on your framerate to get the same feeling on different computers (tested it)

Maybe in LFS they manage to query controller in the middle of frame update...then it is perfect...but until now most games don't do that as it is slower, and then higher framerate in game gives a better responsiveness, less controller lag, better interaction with other network players.

Last edited by Juls, Mon, 16 Jul 2007 12:58.

I got the same problem...I did not find where to pit, I believed you have to find the same stand you started from.

Once you know where to stop it looks obvious, but first time....
It is funny, because it is not written in LFS manual....or search does not give any result.

In fact I would like to buy LFS licence mostly because of LX cars. There is this incredible slow motion video by DoN on youtube where a LX car is jumping.
Is it true that there are no server for LX cars?

Quote from boosterfire :

You don't need more than 25 or 30 fps to drive without a flaw. Under that, MAYBE you'll have some trouble, and that's if you're not used to it.

Personally, anything over 25 fps is fine. People with 120 fps don't gain much experience in 100 fps.

In almost all games (and I suppose LFS too), during frame update, physics engine does not run at all, and later, it runs several times in a raw to keep the pace. For example, when you have 25 fps and 100 hz physics engine (like LFS), physics engine will run 4 times in a raw after every frame, 3 times to simulate what happened during the frame update, and 1 time for next step.

It does not cause any inaccuracy but....The problem is that physics engine does not know how the controller input varied during frame update, and it will run 4 times with the same controller input.

Conclusion:
If you have 25fps, this is an average, and you can expect that sometimes the framerate drops to 10fps for a few frames...
If you have X fps, physics engine see your input (controller) and the network at Xfps. When it reaches 10fps it causes big problem because the engine sees your moves a lot more brutal than they really are (stroboscopic view) and misses quick moves you make.
And there is a one frame duration delay before the physics engine knows about your move, and another frame duration delay before you receive FFB corresponding to your move. At 25fps you have 0.08s delay between move and FFB, this is enough to cause crazy wheel oscillations, prevent the wheel from countersteering effectively...etc

Same for network activity...60+ fps is a LOT better than 25fps.

If you use 720 degrees and more, you should set wheel compensation to 1.0

A simple and nice improvement (without changing DirectX) would be to use alpha channel from the skins. So you can decide how much each part of skin reflects the sky...and that way simulate different materials.

+1 well said.

I just try to imagine what it looks like to buy an expensive real cockpit, integrate 3 monitors to use 140 degrees FOV, then sit, connect a server, and discover that 2/3rd of your monitors area is covered with another cockpit and another wheel.

Hehe, it may be funny....or not.

It is like spherical/cylindrical distorsion...it has to be done by graphics engine. DirectX projection matrix has to be changed before car is displayed. The sample is an edited screenshot.

Anyway, I think this dual FOV thing can work nice with cars like in example (wheels are hidden by car interior), but it will not work with cars like LX and single seat because car wheels will look weird and seem to have lost contact with the road. Or maybe it will...Can't check that.

I think the spherical distorsion is the way to go. With a simple graphical step (mapping the rendered view on a 1:1 subdivided grid with U,V coordinates calculated from a sphere or anything else) it allows to see a nice large FOV without almost any distorsion, and even to create a view perfectly fitted for 3 monitor displays with side monitors turned a lot.

I have problem to make a nice movie of it, because the spherize filter in premiere lowers too much the resolution, but it seems it has some advantages. I can see a lot better on the sides (large FOV) and immersion, sense of speed seem better. To do it properly I need a movie in 1280*768...fraps stops at half this resolution.

I think most racing sims use rather classical tricks to improve immersion. Such an advanced feature, if it works nice, could be cutting edge.

It seems that FOV is now a very discussed point.

I was still searching for an idea to improve the view, and I found something else. Main problem we have with FOV is that large FOV causes distorsion in objects with Y close from zero.
This is very visible within cockpit view: as you increase reasonably FOV, road remains mostly undistorted as it is far away, but car roof, doors and driver's arms are stretched and cover more and more the view. This happens because the projection used in game is perfect...in real life, wide angle lens always have a distorsion on the side which prevent this. For one time, perfection is a disadvantage

One way of getting rid of this is to use cylindrical or spherical distorsion, because this distorsion simulates a bit what happens in a wide angle lens or in our eye. For example the eyefish Quake simulates a view through a real, unperfect fisheye, and it looks far better for large FOV than the perfect DirectX projection.

Another way is to use different FOV. One for the world, one for the cockpit/car. It sounds strange but it works, because outside world and cockpit have different distances. World is mostly far away and large FOV is a better choice, it increases spatial awareness and sense of speed. Increase FOV and you see more.
On the contrary, cockpit is close, and you are inside of it...it looks nicer with lower FOV, closer from natural one.

Some driving simulators use a real car with a monitor on the windscreen.
In that case, cockpit is viewed using the natural FOV, and simulated environment uses another FOV adjusted for the monitor. It seems it does not cause any problem. Same for some flight simulators with real cockpit and monitor (there is a study like that where they use a real cockpit and they change the FOV on the display...from 22 degrees -natural FOV- to 80 degrees, and 80 degrees is better).


Here is an example. First a screenshot with 90 degrees FOV.
http://cracovia.free.fr/lfs_90_full.jpg

Then 130 degrees FOV (extreme setting). Better immersion for the road, but cockpit covers too much screen, door and arms are distorted.

http://cracovia.free.fr/lfs_130_full.jpg

Then a dual FOV screen. Road uses 130 degrees, and cockpit 90 degrees. You can see more road than in the 130 degrees FOV so the immersion is even better, and the cockpit looks better too.

http://cracovia.free.fr/lfs_130_90.jpg

With a dual FOV, as you see more road, you can get better immersion and sense of speed even for moderate road FOV. Increasing world FOV you can adjust sense of speed/distorsion, and reducing cockpit FOV you can decide if you want the cockpit to be more or less present/distorted.

Without moving point of view!

Last edited by Bob Smith, Tue, 10 Jul 2007 16:38. Reason : changed IMG tags to URL tags - images were too big and breaking forum layout

FOV is a very interesting subject. I found several studies about it, measuring performance of pilots/drivers/workers in simulators depending of FOV.

As usual, this is a compromise. Too high FOV causes heavy distorsion. Too low FOV causes a loss of spatial representation (position and speed). According to studies I read, FOV under 60 degrees becomes a disadvantage, and FOV over 140 degrees does not bring any advantage.


With a 17 inchs display, I tried several FOV values, and every time, I asked someone not familiar with video games to tell me what is the car speed.
115 degrees gave the best result for sense of speed.

Then I tested spatial representation (distance). I drive until the car is at the same level than a turn sign, and I ask how far is the next turn sign (turn signs for the reverse track, so figure is not visible). 115 degrees gave good results too. 50m look like 50m.

But you have to get used to distorsion.

Edit: There is somewhere a nice study of FOV effects on a driving simulator. They use a driving simulator with the exact simulation of a real road, and they compare difference between speeds (cornering, straight) and position on track between the real driving and the simulator with various FOV.
Result: with low FOV (50), people are slower in turns and faster in straight line than IRL. And they are placed differently on the track. And if the FOV increases, speed and position in simulator becomes closer from real driving.

Last edited by Juls, Tue, 10 Jul 2007 15:03.

Quote from eRaptor :

Since when showing your arguments is spamming?

I'm opposed to the change, the way it was so radically implemented, but I would accept it with some changes already suggested.

Don't know if you are spamming, but your avatar clearly demonstrates without any doubt that chase view may not be realistic, but can be highly entertaining.

Last edited by Juls, Mon, 9 Jul 2007 16:23.

Quote from ajp71 :

IMO pillars and so on should be in everyones way so bonnet views are out of the question.

This is a dead-end way. With triplehead2go you put FOV at 130-140 or more. With such FOV, the board looks 1 meter away, the interior takes 70% of display, the pillars are close from center, and you need to move at least 50 -70 cms front to have the pillars back at their right place (on side displays).
So you need to allow such move for triplehead2go users, but if you use such move with a 90 degrees FOV, you have a bonnet view.

The funny thing is that one reason why I wanted to buy LFS is the possibility to customize view. Until now only RBR with Camhack allowed me to achieve a satisfying view for my non-usual display (2 monitors, and maybe shutter glasses next as the framerate is very high). GTL, GTR2, rFactor are not satisfying. When I discover it, they start to change it...bad luck

Quote from Vykos69 :

There is NO reasonable roof view for a driver....

Sorry but IRL I can look at the road and see cars on the sides at the same time. I have both sense of speed and see details far away.

Elevated and pitched view may seems a strange idea at first glance, but it gives a more realistic sense of speed, and a more realistic side view. eye's position is not anymore realistic, but two other things become realistic.

With cockpit view you have realistic eyes position, but sense of speed and side view are sacrified, and you can't improve both, because as soon as you increase FOV, car interior stretches.

Cockpit view is a bit tricky...it looks 100% realistic, but it is not

Someone suggested that all view movements should be allowed as soon as you are at same Y-level or in front of driver eyes.

It prevents from using chase views, but allows all possible customisation for large FOV displays, shutter glasses, hood and reasonable roof views that bring no unfair advantage.

(People using 3 monitors have to increase FOV a lot, and they need to move view more than 20 cms front in order to avoid interior covering 70% of screen)