Nose high GSX pushback by others on VATSIM/vPilot

[virtuali]

But it could be even simpler - don't use a towbarless pushback truck, when flying online?

Or disable the "Pushback Raise" in the airplane profile.

But yes, I guess we might detect Multiplayer (or vPilot), and "cheat" on the Pushback preference and always select Towbar in that case.

[Rick Maclure]

Umberto,

Some interesting investigation there and I thank you for investing time in some analysis. I have a much better understanding of what is going on.

Firstly, are we confident that using the towbar does not result in the nose high/low behaviour on vPilot? Similarly with disabling Pushback Raise? If this is true, then we have a workaround and I would ask that, for the benefit of the VATSIM community, you strongly advise GSX Pro users to limit themselves to the towbar when pushing back online.

If you are able to detect the use of vPilot on the client and enforce a 'non-raise' pushback, that would be a stronger solution.

We are desperately keen to avoid witnessing this behaviour...

[virtuali]

Firstly, are we confident that using the towbar does not result in the nose high/low behaviour on vPilot? Similarly with disabling Pushback Raise? If this is true, then we have a workaround and I would ask that, for the benefit of the VATSIM community, you strongly advise GSX Pro users to limit themselves to the towbar when pushing back online.

I haven't tested with vPilot but, another user said it happens "the same" with standard Multiplayer so, I verified that and while it's not exactly "the same" (the airplane jumps in the air, and it's pitched up more or less correctly, so it's different to what's in your screenshot), I definitely can see that, WITHOUT GSX, setting some variables on a plane that has been frozen DOES create issues with Standard Multiplayer and, USING GSX with Standard Multiplayer with a TOWBAR tug, worked perfectly fine, so I'm confident that, even if it's not exactly the same on vPilot, the underlying reason might be similar.

If you are able to detect the use of vPilot on the client and enforce a 'non-raise' pushback, that would be a stronger solution.

From my last post, it should be clear that, from GSX side, we only have TWO options to fix this, both being less than ideal:

- Disable the simulation while pushing with a Towbarless tug. This is way worse than "less than ideal", it's basically unacceptable because, during that time, you wouldn't be able to do anything on the airplane, engines wouldn't start, controls wouldn't move, the airplane will be as dead for the whole pushback duration. So, we can scratch that.

- Enforce the use of a Towbar when using vPilot. A bit annoying, but at least it won't affect the simulation.

On the other hand (assuming it will work), there's the 3rd option, which can be done only from vPilot's side, and it's disabling the simulation of the object it created to represent the remote airplane, for as long it's pushed by GSX, which can be known by checking its frozen status. Sure, you might notice minor things, like if the remote guy lowers the flaps or test controls, you won't see it, but at least it won't be something so obvious like the whole airplane turned down.

[HeicoH]

And of course there is a third option. This one would be ideal for the users, but less than ideal for FSDT.

The developers of GSX could re-write the code of the pushback part of the app so it isn't necessary anymore to use the respective variables that seem to be the reason for the whole phenomen.

[Captain Kevin]

And of course there is a third option. This one would be ideal for the users, but less than ideal for FSDT.

The developers of GSX could re-write the code of the pushback part of the app so it isn't necessary anymore to use the respective variables that seem to be the reason for the whole phenomen.

But would it. Because in one of his previous posts, he said this:

Instead, since vPilot is receiving them and it can verify GSX is pushing, it can fix this in a way simpler way we could ever do, which would, as you correctly guessed, require to change the pushback engine completely, not only losing features, but making all the existing custom pushback useless.

If all the customized GSX profiles become useless, that would be far from ideal for users if you consider the amount of time it takes to create one for one airport, let alone many airports.

[HeicoH]

@Captain Kevin. ("If all the customized GSX profiles become useless...")

I think there is a misunderstanding. I understand Umberto's statement to mean that by "custom pushback" he doesn't mean "customized GSX [airport] profile". Umberto should clarify this.

[Captain Kevin]

@Captain Kevin. ("If all the customized GSX profiles become useless...")

I think there is a misunderstanding. I understand Umberto's statement to mean that by "custom pushback" he doesn't mean "customized GSX [airport] profile". Umberto should clarify this.

The custom pushbacks would still be part of the GSX airport profile anyway, so my point still stands. Unless you've done it before, you have no idea how long it can take to get the custom pushbacks just right for one spot, let alone for an entire airport and multiple airports.

[Rick Maclure]

I haven't tested with vPilot...

Firstly, thank you for confirming the scope of your network testing to date.

I do not have access to GSX Pro so I cannot conduct extensive testing on multiplayer networks. However, I did spent around an hour observing pushbacks on VATSIM.

For each pushback I witnessed, I privately contacted the pilot and asked them what pushback method they were using. Out of the 18 pushbacks I observed, the majority reported that they were using methods other than GSX: Better Pushback (X-Plane), Toolbar Pushback, and integral pushbacks  - namely the FBW A320Neo and the Fenix A320.

There were five pilots who replied stating that they were using GSX. One was using GSX with P3D, and the remaining four were using GSX Pro with MSFS.

If the pilot replied confirming his/her use of GSX, I made a further enquiry asking whether they were using a tow-bar or a tug. Every pushback I witnessed looked normal except two pilots who confirmed they were using the tug in GSX Pro.

Not, I admit, conclusive evidence, but sufficient to suggest that the lifting tug method causes the problem and the tow-bar does not.

Enforce the use of a Towbar when using vPilot. A bit annoying, but at least it won't affect the simulation.

My emphasis. I would suggest it is far more annoying witnessing aircraft burrowing their noses into the tarmac!

The developers of GSX could re-write the code of the pushback part of the app so it isn't necessary anymore to use the respective variables that seem to be the reason for the whole phenomen.

I agree with @HeicoH here. This would be the industrial strength solution. However, I strongly suspect that GSX was developed with the single, non-networked user in mind. I further suspect that absolutely no network testing was undertaken by FSDT around the behaviour of pushbacks.

That said, I understand the cost in completely rewriting the bespoke pushback method and so, I suggest the most effective, and cheapest, method is to restrict the pushback method the tow-bar when on a network (at least the VATSIM network). It would also eliminate the need for third parties (such as the vPilot network client) to make changes at their end.

Respectfully

[Darewell]

Still, it could be difficult to have this method working.

What if I activated GSX before launching VPilot, the tug would already be there, I can’t imagine GSX being able to swap it with a different one as soon as I launch VPilot.

In the meantime a « VATSIM » setting in GSX global settings, forcing the towbar tug, could help, but it will require good will from everyone, and being aware of this issue.I bet a lot of VATSIM/GSX users don’t even have a clue about GSX being the reason of this behavior.

I won’t lie I always laugh when I see other aircraft’s diving in the ground, I’m used to it now

[virtuali]

And of course there is a third option. This one would be ideal for the users, but less than ideal for FSDT.

The developers of GSX could re-write the code of the pushback part of the app so it isn't necessary anymore to use the respective variables that seem to be the reason for the whole phenomen.

And what makes you think this option is "ideal for users"?

Not using the standard limited pushback system, like other apps do, will dumb down GSX pushback to the level of the other system which, incidentally, is one of the reasons users say they bought GSX in the first place. While I think GSX offers a lot more than just the best Pushback possible, it's a fact many users use it for Pushback only, and why do you think that is?

And what about all those hundreds of custom airport profiles, many containing lots of custom pushback routes, do you think it would be "ideal" for users and creators to tell them that, because Vatsim users can't stand seeing *other* user airplanes in the wrong attitude, they would have to trash all that work away?

Are you trying to suggest we should cripple the features which make GSX stand out (custom pushback and a large community of creators), just because there are issues with Multiplayer?

While I don't think the ones with Standard Multiplayer can be fixed, I'm cautiously optimistic it might be possible to fix the ones with vPilot, because if the program works as I think it does (creating its own airplanes independently), the fix might be as simple as disabling the flight model on remote airplane models. It might even have some beneficial effect on performance, since the sim won't have to do any physics calculations for those.

So no, your "third option" is out of the question.

We'd rather consider the "fourth option" before doing that: which is writing a replacement for vPilot ourselves! I'm not kidding: a GSX-aware client might offer extra things, for example being able to see the remote GSX Tug (assuming both users have GSX installed), instead of just the airplane moving by magic...

[virtuali]

What if I activated GSX before launching VPilot, the tug would already be there, I can’t imagine GSX being able to swap it with a different one as soon as I launch VPilot.

No, but it might easily disable the Pushback Raise function in that case. The only issue might be if you started vPilot in the middle of the pushback, when the airplane is already raised, but why would you do that?

[Fragtality]

We'd rather consider the "fourth option" before doing that: which is writing a replacement for vPilot ourselves! I'm not kidding: a GSX-aware client might offer extra things, for example being able to see the remote GSX Tug (assuming both users have GSX installed), instead of just the airplane moving by magic...

If you pair that with an Auto-Gain for the Mic, you would even be able to sell such a Client to non GSX Users (cause everybody likes their Eardrums not being busted) 

[MrRoper]

I have been investigating this over the past few days as I was intrigued to see if we could set the sim state on the aircraft as Umberto suggested. Unfortunately its bad news, whilst you can set this flag on the aircraft it does not fix the issue.

It would seem that this is an underlying simulator issue where the velocity vectors are being continually calculated when the aircraft is frozen by GSX. These values are what vatsim / multiplayer are using to create the smooth animation we see between each update.

Here is a picture of the CRJ resting on the tug whilst there is no movement, notice the velocity vectors and also reported speed.

Not sure where the fix lies, probably with Asobo but its possible that the vPilot guys could change how they predict motion when aircraft are on the ground.

[Atrius]

I have been investigating this over the past few days as I was intrigued to see if we could set the sim state on the aircraft as Umberto suggested. Unfortunately its bad news, whilst you can set this flag on the aircraft it does not fix the issue.

It would seem that this is an underlying simulator issue where the velocity vectors are being continually calculated when the aircraft is frozen by GSX. These values are what vatsim / multiplayer are using to create the smooth animation we see between each update.

Here is a picture of the CRJ resting on the tug whilst there is no movement, notice the velocity vectors and also reported speed.

Not sure where the fix lies, probably with Asobo but its possible that the vPilot guys could change how they predict motion when aircraft are on the ground.

Thank you for looking into this. We as players really don't care where the problem lies. My hope is that the developers of GSX and vPilot/xPilot can talk together and figure out a fix. Sitting on different ends of the problem and blaming each other really does not contribute to the issue.

If not already done, i challenge Umberto and the team to reach out to the vPilot developer(s) and start an initial discussion on the matter to find a common solution. In case you agree that this is something Asobo needs to fix, you are the best people to make a report for them as well with relevant data.

Appreciate any response and attention to this post.

[virtuali]

If not already done, i challenge Umberto and the team to reach out to the vPilot developer(s) and start an initial discussion on the matter to find a common solution.

If you followed my replies, you would have known this already happened, and mr. Carlson has been informed with my latest post, which confirmed the underlying issue is inherent in the way the sim works, and a solution.

However, I don't think it is correct from my side, continuing to update users about any progress, or the eventual lack of it: if mr. Carlson decides to act on this or not, it's his own prerogative, in his own time, and it's not up to me to constantly report this publicly, or even soliciting him in any way.

I guess there MIGHT be a possible solution on our side, in case vPilot won't be updated for some reason, but the solution would be very cumbersome, bad for performance (Simconnect traffic) and unnecessarily complex, worse for everybody compared to a much cleaner solution from vPilot because it would require the following:

- A separate Simconnect app, that we would have to write from scratch, that must be run in parallel to vPilot. We couldn't done it from GSX, since it would force to install GSX even for those that don't use it, so it must be a separate app.

- This separate app, will need to *constantly* monitor ALL AI planes created by vPilot, check when they become frozen by GSX, and Disable their simulation, so it won't interfere with GSX movement.

This MIGHT work (in theory at least), but it's so much worse than a vPilot solution, because we would need to constantly scan all created airplanes, increasing traffic over Simconnect and putting extra strain on the sim, with the added annoyance of another app to start.

Clearly, vPilot could make it much better, since it already has a list of airplane it has created, so it can do it way more efficiently and reliably, without any need for a separate app to start.