FSHostClient for FSX > FSHostClient for FSX

Sticky: Jittery-Jerky planes in formation flying

(1/2) > >>

There was some initial comment about jittery remote planes discussed here (http://www.chocolatesoftware.com/forum/index.php?topic=649.msg3159;topicseen#msg3159) but nothing else has been mentioned since.

I have FSHost 3.2 with FSHostClient 1.1. As a test, I had 3 PCs (1 host, 2 clients) on the same LAN segment. Planes in the remote PC (ie the other plane that you are not flying) appears jerky when in close formation. This same 'bug' appears when testing with another distant client (ping of 110ms). The amount of jerkiness is about the same for the 'local remote' and 'distant remote'.

I wonder how M$ are doing this in FSX multiplayer? Is this another case of where M$ are using 'un-documented' APIs?

Has nothing whatsoever to do with undocumented API's.

Russell Gilbert:
Hi guys,

Yes, this is still a big pain in FSHostClient, and something that I definitely want to get fixed.  The problem is that when FSHostClient displays the remote "fake" plane to represent the other player, it tries to move it quickly enough that it'll look smooth, but it's unfortunately jumping around and looking all jittery.  I've seen the problem myself also.

There are several places where the problem could be, and I've spent quite a lot of time trying to figure it out.  So far though, I've been unable to find the solution.  First, it gets all the location data from the remote plane.  Then it moves the fake plane to the new location.  But if the location data isn't coming quickly enough from the other player, then it needs to do some hairy calculations to try to predict the movement, so it can do updates several times per second.  I think actually that the predicting part of it is ok, but when it gets the next real location update, for some reason it's jumping the plane backwards and forwards, and sometimes even sideways.

I think the reason why it doesn't happen when using FSX's built-in multiplayer system is that it's not having to do all the same calculations, since it has direct access to the information.

Anyway, it's at the absolute top of my priority list.  Hopefully I'll get a break soon from "real life" and dig into it again and get it figured out.  Sorry I don't have a better answer at the moment, but I'm definitely aware of the problem.


Kenny Fox:
Hi Russell.

Russell a big thanks for your work on the FShost Client for FSX ;D There maybe a few bug, but it is still far better that going to Microsoft's gamespy! Anyways a big thanks agent.

VATA Kenny
Senior Admin VintageATC ;D

Hi Russel,

first of all thanks a lot for this wonderful program. We area large FSX community in Spain and are currently evaluating the possibility of getting rid of MS IP network sessions and migrating our server to your FSHost systems. Now, the only issue we see at the moment concerns this jumping effect, as we often fly formations.

I'm not familiar with the algorithms used by FSX and FSHost to position planes, but what you describe in the various posts about this "bug" sounds to me like a problem of location updates when communication between hosts fails.

Now this is nothing new, but I was wondering if the solution to this could be to ignore the updated real position and always use the forecasted (and according to the latest available coordinates adjusted) values. As I understand the problem, this has to do with the fact that in case of forecasted position (which will never be the actual no matter how good the algorithm is) your program sends the outdated packages of data to the hosts and these apply outdated coordinates causing these jumping planes.

OK this was my humble knowledge about algorithms used in real time traffic modelling. Hope it is helpful,

cheers, keep up the great work



[0] Message Index

[#] Next page

Go to full version