Refined turbulence model in 12.1.0

[Franz007 505]

That looks promising! 😊

 

[Sethos 4,452]

Looks great. Now I just want to see 12.1.0 on my computer.

[Murmur 3,944]

Great work Laminar! @jcomm will appreciate this too. 😁

[Franz007 505]

53 minutes ago, Sethos said:

Looks great. Now I just want to see 12.1.0 on my computer.

We all want that 😄

[jcomm 5,961]

Me want's it!!! C'mon LR, give it away ... Now!!!

[Epikk 63]

3 hours ago, jcomm said:

Me want's it!!! C'mon LR, give it away ... Now!!!

I second that 😁

[flightwusel 10]

Nice. Turbulence is broken since a year or so. It was ridiculous how some special figures defended it on the forums. Maybe they got it better this time.

[Franz007 505]

5 hours ago, flightwusel said:

Nice. Turbulence is broken since a year or so. It was ridiculous how some special figures defended it on the forums. Maybe they got it better this time.

Can you explain what exactly was « broken »? They got tuned by Jan and were already quite good imo. 

Edited April 29 by Franz007

[flightwusel 10]

1 hour ago, Franz007 said:

Can you explain what exactly was « broken »? They got tuned by Jan and were already quite good imo. 

It feels ok for airliners. For slower aircraft the effect feels too strong in the same weather. It is too uniform (they seem to have addressed that) and it just randomizes air effects sinus-like a bit without any first principles. Live weather adds it in places where you would not find it while flying in reality (at sunrise without any thermal activity for example, or on a day with just stratus clouds, or directly above clouds).

And because it is not local (but instead just bound to the altitude aka wind layer), it does not work with the aerotow physics: the tow plane should show you the local wind 2s before you enter the same effect. In fact, aerotows feel very unrealistic currently (not only due to the wrong wing vortex effect, but also due to funny turbulence).

By the way: Turbulence in how the word is used here does not exist in meteorology. What they are trying to model here is different local wind (which includes vertical wind) or "turbulent eddies" inside a wind flow. The kind of local wind that does not change much at a single position within the flow, but its effect feels like changing because you move through different pockets of air. Strong "Turbulent flow" is just between different wind layers, right next or inside a thermal column for example or on the lee side of a mountain.

Edited April 29 by flightwusel

[BobFS88 648]

11 hours ago, flightwusel said:

Nice. Turbulence is broken since a year or so. It was ridiculous how some special figures defended it on the forums. Maybe they got it better this time.

I think some of the defensive position comes from their understanding that most of the weather issue have been reported that will be address after the foundational work is done with the graphics, in how all these things are being process. e.g., moving weather process from the cpu to another core or to the gpu. Without that work being completed, it would have been frugal to try to fix something that may depend on this design change in order for the effects to work as you would expect them to be. It maybe too premature to contemplate over these issue with Xplane 12.1 just around the corner when have a chance to see if they address any of yours or other's concern.  

Edited April 29 by BobFS88

[Franz007 505]

8 hours ago, flightwusel said:

It feels ok for airliners. For slower aircraft the effect feels too strong in the same weather. It is too uniform (they seem to have addressed that) and it just randomizes air effects sinus-like a bit without any first principles. Live weather adds it in places where you would not find it while flying in reality (at sunrise without any thermal activity for example, or on a day with just stratus clouds, or directly above clouds).

And because it is not local (but instead just bound to the altitude aka wind layer), it does not work with the aerotow physics: the tow plane should show you the local wind 2s before you enter the same effect. In fact, aerotows feel very unrealistic currently (not only due to the wrong wing vortex effect, but also due to funny turbulence).

By the way: Turbulence in how the word is used here does not exist in meteorology. What they are trying to model here is different local wind (which includes vertical wind) or "turbulent eddies" inside a wind flow. The kind of local wind that does not change much at a single position within the flow, but its effect feels like changing because you move through different pockets of air. Strong "Turbulent flow" is just between different wind layers, right next or inside a thermal column for example or on the lee side of a mountain.

Thanks for the explanations. It sounds a bit like you are expecting a 100% simulation of every wind effects. Of course there will always be some limitations because weather-simulation is a topic that is hugely specialised and very complex.

I personnally haven‘t noticed exagerated turbulences with GA‘s. Thermics are being simulated when some criterias are fulfilled (one of them being warm temperatures during the day). Up- and downwinds over mountainous areas are also simulated. To test that I reproduced the exact same flight over the Alps I did in real. And it felt pretty close (flying closer to the mountain-tops made the aircraft shake pretty similar to how it did in real).

With XP12 I‘m having turbulences most of the time below or inside clouds-layers. Above them you can encounter clear air turbulences that I also experience from time to time, but I couldn‘t tell if the are just randomised or following logical (and realistic) laws. Because I think it would be difficult to know, since there are no visual references.

One of the limitation is that there is no way for X-Plane to tell the stability of the temperature gradient (allowing "blue" thermal activity). So for now it assumes "if cumulus, then vertical rising air“. And there doesn‘t seem to have turbulences crated by obstacles like forests or buildings.

Even with these limitations, I wouldn‘t call it „broken“ because it imitates the effects quite good imo. Can it be improved? Of course. And that‘s also what they seem to have done in some aspects for the next version. But beetween „not perfect“ and „broken“ there is a big difference in my opinion. So it doesn‘t surprise me if some disagreed that it is „broken“. 

Edited April 29 by Franz007

[flightwusel 10]

I just think the lower atmosphere is very important. And simulating a wind vector field by using overlapping turbulent eddies with some granularity is not too hard technically. But yes, maybe my expectations are too high. I have been waiting for the weather API ever since to try it myself. But I don't even know if they are planning to expose that.

+1 that by omitting the temperature gradient they miss out on a very valuable potential input variable.

Maybe they are working on that? I really can't tell from the bullet points that have been released on Twitter. There is so little information coming out. Time for a more technical blog post... 🙂 

[jcomm 5,961]

My feeling is that both XP and now MSFS 2020 too, modeled turbulence in a rather incomplete / unrealistic way.

Typical bounce from unstable air is modeled by default much more covincingly in FSX / P3D, I mean the usual shaking resulting from crossing unstable air...

If we look at the effect caused on most aircraft in XP10/11/12 we see that they bounce madly when there's wind variation, of any intensity. The effect is present evenn at light winds if any kind of shear is present.

Turbulennce shakes the aircraft in 3 DOF in a rather weird way, and this is common to MSFS 2020 with it's new turbulence model which even includes a modulation GUI....

I look fwd to test this refined turbulence simulation in XP12.1 !!!

I confess that it's one of the main showstoppers for me whenever I come back to XP12 and try to enjoy it after gettng frustrated with the rather poor "feel oif flight" that MSFS 2020 provides...

Both simulators also model thermal / convective lift, poorly, but MSFS 2020 is by far the most inconsistent - ridiculous to be honest 😕  - this is yet another aspect of weather modelling that I would love to see fine tuned in X-Plane.

[mSparks 2,091]

its hinted at the end of the video.

But the basic story behind this is the kind of thing you only get from a group like LR.

Austin did his usual custom builds he throws around just because he loves this stuff that included the new g forced camera he presented in Montreal in Feb.

That worked great - except it was then obvious how broken the turbulence was. several backs and forths with anyone with cockpit experience got it tuned beautifully. 

Last I heard it wasnt coming in 12.1 because Austin of course broke lots of other things in his build. Pitchforks were raised and torches were lit.

It wasnt until watching that video I actually understood what all the fuss was about tbh, I think this could actually be a gamechanger for flight sim, pretty sure asobo will be copying it as soon as they get their hands on it.

 

[flightwusel 10]

15 hours ago, mSparks said:

That worked great - except it was then obvious how broken the turbulence was. several backs and forths with anyone with cockpit experience got it tuned beautifully. 

...

It wasnt until watching that video I actually understood what all the fuss was about tbh, I think this could actually be a gamechanger for flight sim, pretty sure asobo will be copying it as soon as they get their hands on it.

 

Thanks for the insight 😉 . I hope it will come out significantly better.