Thoughts / Feelings on Flight Models...

[SteveJ]

Very excited at the propoect of a new WWII era sim in a modern engine, it's something I'm sure any fan of the genre has dreamt about or been frustrated by (games with great content in dated engines).

My background - I read physics at uni, work with numbers, dabbled in sims for years, have a project in Unreal whereupon I felt it was a great idea to teach myself aeronautics & develop a flight model. Guess that puts me in the space of an enthusatic dabbler with a few ideas...

Feels like lots of projects tout "REAL PHYSICS, SEE THE PHYSICS, OMG THE FLIGHT MODEL", while quietly the reality is not "THE PHYSICS" it is what compromises are made in their approximation. I would encourage making that apparent, engage the community where appropriate. I feel like this would be very warmly received.

In that vein: resolving forces on a rigid body (and without actual insight I'm hypothesizing)... many competitors flight models seem to handle certain elements in a similar way, and there's a visual clue to one which is seems to me very common. With my limited knowledge I would surmise that it is common for flight models to resolve distributed force calcs and moments down to single vectors about the aircraft CoG for translation into the game engine.  This to my mind explains the tendancy for aircraft models, in many examples, to almost appear as though they pivot about a mid-point, and sometimes in quite a jarring way. This quirk is something I would dearly like to see the back-end of, deferring to a model with greater leaning towards distributed weights and forces, even at the expense of "absolute accuracy" from an aeronutics perspective. To my mind there's greater potential for return (in terms of emergent features, as a means to differentiate, improved behaviours, much scope to tie into damage modelling.. in sort GAMEPLAY) in such endeavors than in chasing an extra few percent fidelity out of how those forces are caluclated in the first place.

I feel like there's more mileage (at least for release) in saying your aeronautical model has to be good enough (you don't have to break new ground), and that efforts may be better spent on a system to maximise damage modelling, ground handling, audio-visual (the sounds of the aircraft in different states, engines, wind noise) etc. etc.

Anyone else care to chime in?

Edited November 6, 2023 by SteveJ

[Mitthrawnuruodo]

17 hours ago, SteveJ said:

I would surmise that it is common for flight models to resolve distributed force calcs and moments down to single vectors about the aircraft CoG for translation into the game engine.

This by itself shouldn't be a problem if implemented properly (which I imagine isn't necessarily the case).

Every set of forces and moments on a rigid body can be simplified to an equivalent pure moment and force pair at an arbitrary point. Similarly, every motion of an aircraft behaving as a rigid body can be described as a translation of the centre of gravity plus a rotation about the centre of gravity.

I do agree that sim dynamics usually feel a bit odd at times (refer to the endless discussions of "on rails" or "wobbly" flight models in the past), but it's really difficult to pinpoint any causes without inspecting the source code.

 

 

 

 

Edited November 7, 2023 by Mitthrawnuruodo

[SteveJ]

I agree that in principal resolving a distributed set of forces / moments to a pair about the CoG isn't in itself broken, in fact it's the standard approach to solve a system right?

What concerns me is that with this approach may come the supposition / simplification that the CoG / CoL won't change much, if at all.

If instead you start out expecting them to be dynamic, then you'd likely lean in different ways when implmenting any particular system.