I would not use Seated Passengers until SU4 is out. Umberto blames Asobo as usual and didn't look at finding the solution and optimising it so it works in SU3.
I don't "blame Asobo as usual", I pointed out an UNDENIABLE FACT that the problem STARTED with SU3, and this is beyond any doubt and easily proven. Well, you said that yourself...
In SU4 is improved A LOT, exactly as I've said, but that doesn't mean the impact of potentially hundreds attached Simobject is exactly 0%, and if you weren't so blinded by your useless crusade to contest everything I said, you should have understood the problem reported as the OP is COMPLETELY different than that:
You reported a failure in the sim to optimize for invisible attached object that
started with SU3, resulting in
"Terrible performances", THESE were your own words.
But you said that "GSX Seated Passengers are not optimized", without realizing how absurd this statement is and how you are contradicting yourself since (your own words), it "Started with SU3" so, how can possibly be a result of "GSX Passengers not optimized", when the problem "started with SU3" ? This is so obvious that I'm even feel bad having to say that, because every sane person would understand that.
Again, the "terrible performances" of attached objects in SU3 has been greatly improved in the current SU4 beta (as I PROVED in my video), but their impact is still not exactly 0, but surely the are not nearly has the same bad impact as it was in SU3.
So, let's recap the FACTS, shall we ?
- Up to SU2 (and in MSFS 2020), NOBODY had a problem with Seated passengers when they were not visible.
- With SU3, their impact became much worse ( "terrible performances" )
- With SU4, it has improved, maye not as good as they were in SU2 and before.
So, the facts are: there has been 3 main simulator updates: going from very good, to very bad to acceptable, the only variable that hasn't changed are...the GSX passengers and the only one that changed are the 3 simulator updates. But yes, of course "I'm blaming Asobo as usual" ? Right.
Suppose the thing were in reverse: between NO Simulator updates, a GSX update came out and makes performances worse, what would be YOUR reaction if I said "davidcherrie blames GSX, as usual", because THIS is what you are doing!
Now, having closed your absurd argument about Seated Passengers, which are undoubtedly having varying performances depending on the various simulator updates, let's finally arrive at the subject of this post, which you haven't even realized is reporting something COMPLETELY different.
You were complaining about the very existence of Seated passengers affecting performances (known SU3 issue that didn't exists up to SU2, improved in SU4), the OP is reporting a completely different matter, which is:
micro-stutters when the Beacon lights is turned onThis has NOTHING to do with the very existence of Seated Passengers, because if it happens only when the Beacon light is turned on, NOTHING is happening to Seated Passengers at that time. If they were affecting performances before turning on the Beacon light, they would do it in the exact same manner with the Beacon light enabled.
As explained so many times, once the passengers sit, they are NO LONGER UNDER GSX CONTROL. In fact, they have been always part of the airplane, the only thing GSX did was to set their visibility variable to turn on/ff individual seats. I don't even have to explain this: you know better than me how GSX attaches Seated Passengers on the airplane, so I don't have to explain that GSX has zero control over them.
The only possible "bugs" in GSX would be failing to set the correct seat LVar (but then you might see all seats taken, always) or failing to recognize the cockpit door, but then you would always see passengers even with the door closed (and most importantly, you won't see them disappearing the moment you close it), so if you are NOT seeing either of these, it means GSX is doing its job, then it's up to the simulator to properly optimized object that are failing their Visibility test.
What REALLY happens when turning on the Beacon light ( in the Fenix, of course ), is that GSX is preparing for departure, so it creates the Pushback crew that will walk to the airplane, so it's possible those "micro stutters" are simply the normal effect of asking the sim to create new objects, and its impact is very different depending on so many variables like how many objects are already there, how taxed is the sim, how is the frame rate (too high and unlocked would make it worse) but, in general, creating an object through a Simconnect command is NOT a "free action", some micro stuttering will always happen.
Then there's the other observation: that
those stutters are related to the flashing interval of the Beacon light. This is quite interesting but, how can I provide a PROPER explanation without you having *again* saying "Umberto blamed Asobo ?", I really can't but, try to guess what that might be ?
A Beacon light is a dynamic light. It's should be intuitive that, dynamic lights have an impact on performances that can change depending how complex the scene they are lighting.
This is easily proven by the fact that, GSX Passengers (or ANY attached object) are not lit by the standard emissive Cabin lights, so we had to add internal Dynamic lights to prevent passengers to be pitch black, which of course was reported as a "GSX bug" rather than what really is: the simulator not calculating lighting on attached objects *unless* there's a full-blown dynamic light, not just a simple emissive texture.
So, the FACT here is:
- Cabin lights don't lit attached objects, Dynamic lights do.
- The Beacon light IS a Dynamic light, not an internal cabin emissive texture.
This obvious conclusion is:
- The Beacon light is lighting attached objects = Seated Passengers.
Would you reasonably expect that lighting 150+ objects that wouldn't have been there otherwise, with a truly dynamic light (flashing/rotating) is completely free ?
Do you think we can do something to PREVENT Passengers to be lit by the Beacon light ? No, of course we can't, there's no such thing as a setting in the SDK that tells "this object shouldn't not be lit by dynamic lights".
There IS something we can try: we might remove the extra dynamic lights attached inside the airplane so, we would reduce the number of dynamic lights by 12 (that's for the A320, it's different for every airplane and it's related to the cabin length) so MAYBE this might be enough to not to feel the impact of adding a new one ( the Beacon light ).
Of course, doing this would revert passengers to be pitch black again so, I wonder what would be the point then: if you would end up seeing distracting pitch black passengers for the entire flight (assuming it happens at night) to possibly save from a micro-stuttering when the Beacon light is enabled, wouldn't be just easier to NOT USE THEM ?
Because that's the main point that everybody is missing:
Seated Passengers are OPTIONAL and are not even enabled by default (and that was of course a precise choice, so nobody would miss the point), if you don't like micro-stutters, or you don't like pitch black passengers or you don't like the fact that SOME performance cost is associated with them, the solution is very easy, don't enable them in the first place.
