GPL force feeback - setting latency in core.ini

I've been tweaking my core.ini settings for force feedback for my new Logitech Momo wheel and I have been having some trouble with latency. I understand that the current thinking is not to set latency by bumping kerbs, but to swerve back and forth on a straight and try to get the feedback to match. What exactly are you trying to feel in this exercise? Should the wheel get light during this? How fast do you drive when you do this test? When I try this going very slowly (around 20 mph), I can get what I think is okay, but if I step it up to 40 mph the wheel gets completely light and I have absolutely no control. Does that indicate a latency that is too long? Any help with this would be appreciated. Dave Ewing -- ***************************************************** David A. Ewing Xilinx Boulder [email hidden] *****************************************************

From what I've read to have latency free reactions GPL has to predict what you should be feeling instead of what the wheel is actually doing. If you increase the latency GPL has to do less predicting and you get better feedback but you'll be feeling the road behind your car. Can somebody clear this up? Are any super low latency wheels coming out? "David Ewing" <[email hidden]> wrote in message news:[email hidden]...

Quoted text (23 lines)

> I've been tweaking my core.ini settings for force feedback for my new > Logitech Momo wheel and I have been having some trouble with latency. > I understand that the current thinking is not to set latency by bumping > kerbs, but to swerve back and forth on a straight and try to get the > feedback to match. What exactly are you trying to feel in this > exercise? Should the wheel get light during this? How fast do you drive > when you do this test? When I try this going very slowly (around 20 > mph), I can get what I think is okay, but if I step it up to 40 mph the > wheel gets completely light and I have absolutely no control. Does that > indicate a latency that is too long? > > Any help with this would be appreciated. > > Dave Ewing > > -- > ***************************************************** > David A. Ewing > Xilinx Boulder [email hidden] > ***************************************************** > > >

What you are looking for is for the forces to feel natural and in phase with your steering motions. Admittedly, it can be easy to fool yourself this way. After playing with is some more, I think the best way to test it is to make a single lane-change type maneuver, as if pulling out to pass someone, rather than constantly swerving back and forth. The reason is that even with real cars, if you swerve back and forth at the wrong frequency you can get out of phase with the chassis frequency, and then end up swerving wider and wider, much like what happens with the wrong latency correction. But with a single passing type maneuver the feeling is immediate and very obvious. In this type of maneuver you should feel resistance to the initial motion, the force should change direction as you bring the wheel back the other way to straighten out, and then should die out very quickly as the car straightens out. With inadequate latency correction the force will continue in the same direction after you have already reversed the direction of the wheel, and will often throw the wheel back the other way. Try it at a medium speed like maybe 60 to 80 mph, so that you do not have to deal with loss of traction, which will change what you are feeling. When it is correct it should feel as natural as changing lanes to pass someone on the highway. When it is incorrect it feels obviously wrong. BTW if you also have NASCAR 4 save yourself some trouble and try setting it there first. Do the same exercise as above, but at Watkins Glen in Simulation mode with the "<fast>" setup, so that there is some similarity in the steering ratio. This is quicker because you can change it from the options menu without even getting out of your test session. For me at least, the value that worked right in N4 also worked right in GPL. "David Ewing" <[email hidden]> wrote in message news:[email hidden]... > I've been tweaking my core.ini settings for force feedback for my new > Logitech Momo wheel and I have been having some trouble with latency. > I understand that the current thinking is not to set latency bybumping > kerbs, but to swerve back and forth on a straight and try to get the > feedback to match. What exactly are you trying to feel in this > exercise? Should the wheel get light during this? How fast do youdrive > when you do this test? When I try this going very slowly (around 20 > mph), I can get what I think is okay, but if I step it up to 40 mphthe > wheel gets completely light and I have absolutely no control. Doesthat

Quoted text (14 lines)

> indicate a latency that is too long? > > Any help with this would be appreciated. > > Dave Ewing > > -- > ***************************************************** > David A. Ewing > Xilinx Boulder [email hidden] > ***************************************************** > > >

Hehe .. just a casual aside thrown in here... I 've got two pc's set up to race and so have two ff wheels.. one act labs which is sorted but the early guillemot ferrari wheel on the other machine is really hard to keep straight !! It reminds me of that old van we all set off for morocco in.. the one with the half inch play either way in the steering... in order to keep in a straight line you have to describe a gentle series of arcs along the track, valiantly trying to keep up as the ff wanders back and forth... while my co teamie has a tried and successful way with this drawback, which involves the honda .and turning the ff waaay down...., he has most success on tracks without long straight bits.... "Haqsau" <[email hidden]> wrote in message news:a2ksed$11umac$[email hidden]...

Quoted text (60 lines)

> What you are looking for is for the forces to feel natural and in phase > with your steering motions. Admittedly, it can be easy to fool yourself > this way. After playing with is some more, I think the best way to test > it is to make a single lane-change type maneuver, as if pulling out to > pass someone, rather than constantly swerving back and forth. The > reason is that even with real cars, if you swerve back and forth at the > wrong frequency you can get out of phase with the chassis frequency, and > then end up swerving wider and wider, much like what happens with the > wrong latency correction. But with a single passing type maneuver the > feeling is immediate and very obvious. In this type of maneuver you > should feel resistance to the initial motion, the force should change > direction as you bring the wheel back the other way to straighten out, > and then should die out very quickly as the car straightens out. With > inadequate latency correction the force will continue in the same > direction after you have already reversed the direction of the wheel, > and will often throw the wheel back the other way. Try it at a medium > speed like maybe 60 to 80 mph, so that you do not have to deal with loss > of traction, which will change what you are feeling. When it is correct > it should feel as natural as changing lanes to pass someone on the > highway. When it is incorrect it feels obviously wrong. > > BTW if you also have NASCAR 4 save yourself some trouble and try setting > it there first. Do the same exercise as above, but at Watkins Glen in > Simulation mode with the "<fast>" setup, so that there is some > similarity in the steering ratio. This is quicker because you can > change it from the options menu without even getting out of your test > session. For me at least, the value that worked right in N4 also worked > right in GPL. > > > "David Ewing" <[email hidden]> wrote in message > news:[email hidden]... > > I've been tweaking my core.ini settings for force feedback for my new > > Logitech Momo wheel and I have been having some trouble with latency. > > I understand that the current thinking is not to set latency by > bumping > > kerbs, but to swerve back and forth on a straight and try to get the > > feedback to match. What exactly are you trying to feel in this > > exercise? Should the wheel get light during this? How fast do you > drive > > when you do this test? When I try this going very slowly (around 20 > > mph), I can get what I think is okay, but if I step it up to 40 mph > the > > wheel gets completely light and I have absolutely no control. Does > that > > indicate a latency that is too long? > > > > Any help with this would be appreciated. > > > > Dave Ewing > > > > -- > > ***************************************************** > > David A. Ewing > > Xilinx Boulder [email hidden] > > ***************************************************** > > > > > > >

Thanks, Haqsau, I'll give this a try. Dave Ewing Haqsau wrote:

Quoted text (20 lines)

> What you are looking for is for the forces to feel natural and in phase > with your steering motions. Admittedly, it can be easy to fool yourself > this way. After playing with is some more, I think the best way to test > it is to make a single lane-change type maneuver, as if pulling out to > pass someone, rather than constantly swerving back and forth. The > reason is that even with real cars, if you swerve back and forth at the > wrong frequency you can get out of phase with the chassis frequency, and > then end up swerving wider and wider, much like what happens with the > wrong latency correction. But with a single passing type maneuver the > feeling is immediate and very obvious. In this type of maneuver you > should feel resistance to the initial motion, the force should change > direction as you bring the wheel back the other way to straighten out, > and then should die out very quickly as the car straightens out. With > inadequate latency correction the force will continue in the same > direction after you have already reversed the direction of the wheel, > and will often throw the wheel back the other way. Try it at a medium > speed like maybe 60 to 80 mph, so that you do not have to deal with loss > of traction, which will change what you are feeling. When it is correct > it should feel as natural as changing lanes to pass someone on the > highway. When it is incorrect it feels obviously wrong.

-- ***************************************************** David A. Ewing Xilinx Boulder [email hidden] *****************************************************