Wind smoothing effects

[funkyp]

Hi Pete,

I'm working on a new soaring plugin for FSX, in tradition of CCS2004 of Eric Carden. I'm registered user of FSUIPC4, although I do not need it for my project.

I hit on a potential problem with wind smoothing. I'm trying to develop "natural" thermal patterns from the weather condition, in particular I'm busy with realising streeting patterns, that occur a medium wind speed. In order to keep the thing multiplayer compatible, every party has to experience the same weather conditions, which is normally fulfilled, when you have joined in a multi-player session.

What happens now with FSUIPC's wind smoothing? Could it happen, that the multiplayer engine and FSUIPC starts fighting each other? I suppose, at the end winds are different over time for the different parties. But are the wind data in average the same? In this case I could again smooth it out, to have in average the same wind conditions for all participants.

Another question is, whether weather remains totally unaffected in unlicensed versions.

Or is the only reasonable recommendation to turn off weather settings entirely, as long as the multiplayer session runs, and if wind conditions in particular remain unaffected then?

best regards,

Peter

BTW: Is there a way to scale the joystick axis such that there is less than 100% full scale output? Input range reduction is simple.

[Pete Dowson]

What happens now with FSUIPC's wind smoothing? Could it happen, that the multiplayer engine and FSUIPC starts fighting each other? I suppose, at the end winds are different over time for the different parties. But are the wind data in average the same? In this case I could again smooth it out, to have in average the same wind conditions for all participants.

Sorry, I don't know how to answer you.

If wind smoothing is enabled (and it is an option) it tries to smooth the wind changes at the rate defined by the user. This isn't new. It tried to do this before, by changing the winds in the WX station METARs. That wasn't very successful at all, but it did mean that if a Flight was saved, the WX file it created would have included been the changes, not the proper winds. (Ugh).

The current method is almost identical to the system used in FS2000, FS2002 and FS2004. Didn't you ever have to deal with any of those?

Another question is, whether weather remains totally unaffected in unlicensed versions.

Unaffected by FSUIPC, you mean? It is not affected at all if none of the weather options are enabled -- particularly the one to change FS's own weather (even with just that enabled, FSUIPC still tries to reduce the ridiculous number of spurious wind and temperature layers FSX creates over a period of time).

An unregistered user cannot enable the options, so they are in the same boat, weather-wise, as a registered user who doesn't enable any of the options.

Or is the only reasonable recommendation to turn off weather settings entirely, as long as the multiplayer session runs, and if wind conditions in particular remain unaffected then?

No idea, sorry. Obviously any weather option used differently by two people will result in differences. What else can I say? It always has been, it is no different now.

BTW: Is there a way to scale the joystick axis such that there is less than 100% full scale output? Input range reduction is simple.

You can write a program to disconnect axes, read the original value, scale it and write to on. You can do that directly in SimConnect or via an FSUIPC client.

Regards

Pete

[funkyp]

Hi Pete,

thank you for your rapid answer.

No, I haven't any programming experience with FSUIPC yet and don't have any insight how FSUIPC did wind smoothing before.

But from your reply I conclude, that you don't overwrite FSX weather stations anymore, as according to your release notes was the case before.

That is good to know, since it is a chance to use wind from a weather station, or from interpolated observation to determine the wind conditions for the thermals situation, even with wind smoothing is on.

My problem is, that all connected clients have to produce the same set of thermals, which tends to become very difficult, the more influences are engaged.

Regarding the axis questions, writing a program is of course possible. In effect I would like to suggest to consider a shrinking of the effective control range from the FSUIPC control panel. Sometimes controls appear overefficient and the sensitivity parameters of FSX seem to do just damping, or a non-linear reduction around the neutral position of the controls, rather than a reduction of the full-scale deflection.

best regards,

Peter

[Pete Dowson]

But from your reply I conclude, that you don't overwrite FSX weather stations anymore, as according to your release notes was the case before.

Obviousy, it tried to influence the weather in the way SimConnect provides -- by setting METARs at the surrounding WX stations.

It will still do this if you enable any of the other options, like adding random turbulence, cirrus layers, visibility restrictions, and so on, and it will always try to write back METARs it reads with multiple fals spurious layers without those layers.

All that presumes you enable the option for it to change FS's weather. Without that option, nothing is changed except weather being set THROUGH FSUIPC.

That is good to know, since it is a chance to use wind from a weather station, or from interpolated observation to determine the wind conditions for the thermals situation, even with wind smoothing is on.

I'm not really understanding you here. The wind smoothing now doesn't change the METARs, so you won't see what is happening -- unless you read the ambient winds at the aircraft. It is those which are being directly manipulated. They are rarely related to any specific WX station in any case.

My problem is, that all connected clients have to produce the same set of thermals, which tends to become very difficult, the more influences are engaged.

Of course. That's always been the case. If you want everyone to have the exact same conditions then no one can use any FSUIPC weather options at all. Naturally. As it was in all previous versions of FS.

Regarding the axis questions, writing a program is of course possible. In effect I would like to suggest to consider a shrinking of the effective control range from the FSUIPC control panel.

I would most certainly not be in favour of that at all. You should always be able to reach the extreme deflections defined for that aircraft. If those deflections are defined too great for the model, it is the aircraft configuration which needs changing.

Sometimes controls appear overefficient and the sensitivity parameters of FSX seem to do just damping, or a non-linear reduction around the neutral position of the controls, rather than a reduction of the full-scale deflection.

Quite right too, in my opinion. Anyway, FS's sensitivity does effectively reduce the range, which is why I always tell folks to set that to maximum if they calibrate in FSUIPC -- otherwise all they are doing is spreading a diminishing amount of axis movement to the full deflection range.

Surely what you want is a flatter (less sensitive) response over the central area, going steeper at the extremes so that the maximum deflection can still be achieved. That is exactly what the slope facilities in FSUIPC are for!

You really also always want to change the FS stick sensitivity mode, in FS's CFG file, as instructed in the FSUIPC user guide.

Regards

Pete

[funkyp]

Thanks for the clarifications, obviously I had still some misunderstandings.

Obviousy, it tried to influence the weather in the way SimConnect provides -- by setting METARs at the surrounding WX stations.

and it will always try to write back METARs it reads with multiple fals spurious layers without those layers.

From your "horrible hack" remarks I assumed that there was still something else. That was one reason for my questions. Still, even when it is done through METARs (what I meant with FSX Weather stations), then it's ok for my purpose if the principal METAR winds (Edit, was: surface winds) stays unaffected, which I understand is the case.

It's important for me to achieve compatibility with FSUIPC, and not urging the user to enable or disable important options.

I would most certainly not be in favour of that at all. You should always be able to reach the extreme deflections defined for that aircraft. If those deflections are defined too great for the model, it is the aircraft configuration which needs changing.

Principally I agree with you, with respect to the main controls.

Surely what you want is a flatter (less sensitive) response over the central area, going steeper at the extremes so that the maximum deflection can still be achieved. That is exactly what the slope facilities in FSUIPC are for!

In this particular case this is not exactly what I want to do. I found it very conveninet to map the twist handle to the Pan Heading axis, but I would like to restrict the pan angle from -90° to +90° (or so), rather than the normal +/-180°, to get a better control when looking sideways. Scaling down the sensitivity in FSX makes the axis gain very low around the center, but even more sensitive the more you turn the handle until you reac +/- 180° again.

In this situation, the convenient user interface of FSUIPC suggested to me, that there might be a chance to make the adjustment here, but of course there was not. (? or:is not?)

You really also always want to change the FS stick sensitivity mode, in FS's CFG file, as instructed in the FSUIPC user guide.

Thanks for the hint and I will have a look into it. That's the thing with MS Software: Most of the essential support is from third party.

best regards,

Peter

[Pete Dowson]

From your "horrible hack" remarks I assumed that there was still something else.

All the recent implementation of wind smoothing is by hacking. One part is a tidy call interception between on FS module and another. That nearly worked except that the odd incorrect wind slipped though and caused "ratcheting" (horrible jerks) on the ASI and other instruments. To get rid of that i had to do something really horrible -- patch SIM1 (dynamically, in memory of course) with a blister into a routine of my own. that is my "horrible hack".

it's ok for my purpose if the principal METAR winds (Edit, was: surface winds) stays unaffected, which I understand is the case.

The METAR reports should stay the same. Whether the surface wind at the aircraft is the same as the local METAR report (assuming there is one close by -- that is never guaranteed) is in the lap of FSX.

In this particular case this is not exactly what I want to do. I found it very conveninet to map the twist handle to the Pan Heading axis, but I would like to restrict the pan angle from -90° to +90° (or so), rather than the normal +/-180°, to get a better control when looking sideways.

Ah, right. Panning. never did like it -- it works well with TrackIR provided you have an FS virtual cockpit for orientation. In a cockpit like mine, with a real window onto a projected scenery view, panning is abysmal, so i never use it.

Scaling down the sensitivity in FSX makes the axis gain very low around the center, but even more sensitive the more you turn the handle until you reac +/- 180° again.

Yes, that's the idea with the normal flattened centre slopes in FSUIPC too. I didn't know FS did it so well.

In this situation, the convenient user interface of FSUIPC suggested to me, that there might be a chance to make the adjustment here, but of course there was not. (? or:is not?)

Not as it stands. I'm not really sure how I could implement such an option .. just for panning?

Regards

Pete

[funkyp]

All the recent implementation of wind smoothing is by hacking. One part is a tidy call interception between on FS module and another. That nearly worked except that the odd incorrect wind slipped though and caused "ratcheting" (horrible jerks) on the ASI and other instruments. To get rid of that i had to do something really horrible -- patch SIM1 (dynamically, in memory of course) with a blister into a routine of my own. that is my "horrible hack".

Hm, is the Sim1.dll used by all aircraft? I'm currently experiencing some weird behaviour with the DG808, where wind effects in the Shift-Z and by aircraft and gauges behaviour is actually are much worse with smoothing on as without.

It appears as an unstable siutation, when the ambient wind speed oscillates with increasing amplitude, while interpolated METAR info remains stable. I have still 4.241 in place, but I will start with 4.250 tomorrow.

I found this with "Thunderstorms" theme (or so) near Sion, simulated date 21.06.07 around 13:00 LT, in appr. 5000ft.

regards,

Peter

[Pete Dowson]

Hm, is the Sim1.dll used by all aircraft?

SIM1.DLL is the main simulation engine of FS. There's no other. I think originally (going waaaay back) the idea was to have different modules for different types of aircraft, but that is certainly not implemented in any recent version, as far back as I can remember.

I'm currently experiencing some weird behaviour with the DG808, where wind effects in the Shift-Z and by aircraft and gauges behaviour is actually are much worse with smoothing on as without. It appears as an unstable siutation, when the ambient wind speed oscillates with increasing amplitude, while interpolated METAR info remains stable.

That'll be the turbulence emulation -- either suppress it (in Winds and Clouds tabs), or move on to version 4.251 and try that. I'm still developing the wind effects.

I'm amazed that as a 4.241 Beta user you've not read anything here first about all the developments we've been doing and testing on all this!?

Regards

Pete