Board index FlightGear Development Weather

Wind-shear

Everything related to weather simulation, visuals should be discussed in the shader subforum.

Re: Wind-shear

Postby Thorsten » Thu Aug 25, 2022 6:43 am

OK, if interpolation in AW works, then something other in AW is wrong, because only FG AW generates discontinuity and only in live METAR mode.


Well, your own sentence contradicts you, because obviously AW has no rapid changes when you run it in offline mode despite using the same set of interpolation routines.

So your logic is flawed - if the same routines handle generated data fine but fetched data not, the logical conclusion is that the difference has to do with how online data is fetched and thus differes from generated data. But again, that's been explained in a dozen of threads in detail, and just because you insist in ignoring the info given to you I won't bother to repeat it.

Yes other programs use the same data and don't exhibit the phenomenon.


How do you know they fetch METAR the same way (i.e. fetch only the nearest station whatever it is, and only if within a given distance)? I don't think that's a reasonable assumption (clearly it's not what I would do if I were interested in flying live METAR because, well, it causes all the trouble).


Well saying it's the Iinput data is just a copout.


Whether you like it or not, it's what the problem is. But you've so far been very successful in diverting attention from the real issue, and as a result nobody has tried to fix the real issue - congratulatoins :D (it's been known for nearly a decade, but I've also made clear repeatedly that I am supremely not interested in working on any functionality that requires FG to be online - be it MP or live weather - because I usually never let FG go online in the first place, so... somebody else has to do that kind of work).

But yeah, I know exactly how to do the algorithms for weather synchronization over MP or smooth live weather fetching (and have also sketched these in different threads) - that however doesn't make me responsible for implementing a feature you want and I do not.

Again - if you don't want things 'as they are' and don't want to get your sleeves up and code a change - use offline weather (apparently that works for you) - or do not use FG and be happy with whatever other sim you like.

But the takeaway message is - stop random finger-pointing at systems you do not understand. Assigning the blame to a system that isn't actually responsible doesn't help anyone, it just makes life worse for everyone.

Thus it should be considered if the interpolation should have a higher minimum transition time / distance so that it does not generate rapid (although smoothed) transitions that are most likely not really there.


Yawn...

Actually that time is in (and has been from the beginning, I actually know the math of interpolations really well...) - the issue is that there's also a maximum time you have that's given by the requirement that you need the weather reported by the METAR right at the station - if you set the timescale too long, you may end up landing in winds that have not adapted at all to what the weatherman said. So by easing one issue you always make the other issue worse and vice versa.
Last edited by Thorsten on Thu Aug 25, 2022 10:58 am, edited 1 time in total.
Thorsten
 
Posts: 12490
Joined: Mon Nov 02, 2009 9:33 am

Re: Wind-shear

Postby MariuszXC » Thu Aug 25, 2022 9:27 am

Thorsten wrote in Thu Aug 25, 2022 6:43 am: if you set the timescale too long, you may end up landing in winds that have not adapted at all to what the weatherman said. So by easing one issue you always make the other issue worse and vice versa.


How about it does not adapt at all, and you end up landing at your destination in weather from your departure airport, even though the arrival airport METAR has been fetched? This also does not conform to 'what the weatherman said'. Please see this: viewtopic.php?f=69&t=40711
INOP
MariuszXC
 
Posts: 1061
Joined: Tue May 18, 2021 5:38 pm
Location: Europe
Callsign: SP-MRM
Version: 2020.4
OS: Ubuntu 16.04

Re: Wind-shear

Postby Thorsten » Thu Aug 25, 2022 10:49 am

How about it does not adapt at all, and you end up landing at your destination in weather from your departure airport, even though the arrival airport METAR has been fetched?


That's a poor man's offline mode I guess - not sure what bearing it has on the current discussion though?
Thorsten
 
Posts: 12490
Joined: Mon Nov 02, 2009 9:33 am

Re: Wind-shear

Postby V12 » Thu Aug 25, 2022 11:05 am

MariuszXC wrote in Thu Aug 04, 2022 11:44 am:H
As flight progressed a new METAR from LGRP was received, with the usual 'turbulence' effect.

It is very nice description of the interpolation error between the 2 METAR stations. I am very glad that I am not the only one who sees this error. Same errors are visible in this captured video :



Check timeframe 19:40 and 33:35.
Fly high, fly fast - fly Concorde !
V12
 
Posts: 2757
Joined: Thu Jan 12, 2017 5:27 pm
Location: LZIB
Callsign: BAWV12

Re: Wind-shear

Postby MariuszXC » Thu Aug 25, 2022 11:11 am

Thorsten, I do not understand your answer. What does an 'offline mode' mean, when the whole test was online, with new METAR being fetched properly (and, in fact, other pilots flying around, as it was a multiplayer event)?

As to what bearing it has, the discussion took a direction into weather adapting (interpolating) to new inputs. My point was, that it did not even begin to adapt to a new input, at all.

Edit:
@V12: I do not consider the 'turbulence' effect as error, per se. In real flying pilots sometimes encounter turbulence seemingly 'out of nowhere', so it can be argued, that it even increases the perceived realism of the simulation. I do however consider the weather (more precisely: wind speed and direction) experienced during landing being vastly different from what the current METAR for the airport I am landing at says, as one.
INOP
MariuszXC
 
Posts: 1061
Joined: Tue May 18, 2021 5:38 pm
Location: Europe
Callsign: SP-MRM
Version: 2020.4
OS: Ubuntu 16.04

Re: Wind-shear

Postby Thorsten » Thu Aug 25, 2022 11:57 am

Thorsten, I do not understand your answer.


That is because I have no clue what the bearing of your experience on the current discussion is. Whatever issue caused your experience is for sure not related to the question whether the interpolation algorithm works. So please don't mix it.


It is very nice description of the interpolation error between the 2 METAR stations. I am very glad that I am not the only one who sees this error.


Look - are you unable to read and process any of the provided information?

We all know the effect exists - no one doubts that.

The issue is all about whether this is an 'interpolation error' - or something else.

All of us who have bothered to do the math also know where it comes from - it has in fact nothing to do with the interpolation routines - they do what they're supposed to do - it is caused by the pathological way input data is fed to the interpolation when the weather stations are very sparse and your weather is determined by a station 3000 km away when suddenly a new station pops up 100 km distant.

So - you've been given a complete analysis of the issue, including a path to fixing it - and all you can think of doing is posting videos convincing people that there is an issue? I mean, yeah - by all means, prevent any constructive steps ad infinitum by dragging this discussion into its next and next iteration, always back to the beginning...
Thorsten
 
Posts: 12490
Joined: Mon Nov 02, 2009 9:33 am

Re: Wind-shear

Postby MariuszXC » Thu Aug 25, 2022 12:18 pm

Thorsten wrote in Thu Aug 25, 2022 11:57 am:
Thorsten, I do not understand your answer.


That is because I have no clue what the bearing of your experience on the current discussion is.


I explained but it seems your choice was to disregard what I wrote.
Probably a communication error on my part.
Apologies, won't disturb more.

Although I would love to understand, why the 'new station that popped up nearby' DID NOT update the wind direction and speed properties to new values at all.
If you feel this not a right place, there is the other thread I referenced, or a bug report you may comment on.
INOP
MariuszXC
 
Posts: 1061
Joined: Tue May 18, 2021 5:38 pm
Location: Europe
Callsign: SP-MRM
Version: 2020.4
OS: Ubuntu 16.04

Re: Wind-shear

Postby V12 » Thu Aug 25, 2022 12:20 pm

MariuszXc :
I consider turbulence effects as error, because these effects are always connected with METAR changes. Then I can assume that abrupt changes are not caused by turbulence simulation in the AW, but interpolation bug some flaw in the AW.

Thorsten :
Again check the video - it is test flight from LOWI to LSGG with many METAR stations on the 220 nm route. It was not transcontinental flight. Flight from LGKP to LGRP is 75 nm - again relative short, with some other METAR stations in the vicinity, but not between departure and arrival point.
Fly high, fly fast - fly Concorde !
V12
 
Posts: 2757
Joined: Thu Jan 12, 2017 5:27 pm
Location: LZIB
Callsign: BAWV12

Re: Wind-shear

Postby Thorsten » Thu Aug 25, 2022 4:12 pm

Again check the video


Customer (C): My camera doesn't work! Look, there's all these white spots and the colors are funny.
Support (S): That'd be because you're pointing it directly into the Sun. Cameras don't deal well with that.
C: But it doesn't work!
S: That has nothing to do with the camera, the issue is with what you're pointing it at.
C: But I know the camera doesn't work - my friend also has a different camera, and he never has such problems.
S: Probably he doesn't point it directly into the Sun, or at least dials down exposure massively.
C: But my camera has such problems!
S: Probably not in overcast weather.
C: Well, no - but that's not the issue, the issue is that my camera is broken.
S: No, there's really nothing wrong with it.
C: Yes, there is - look, here's another picture that's gone bad!
S: Because you were pointing it into the Sun again... Again - not a camera problem, the issue is with what you point it at.
C: No, it must be the camera. Look, my friend never has such issues
(...) ad nauseam :D
Thorsten
 
Posts: 12490
Joined: Mon Nov 02, 2009 9:33 am

Re: Wind-shear

Postby Thorsten » Thu Aug 25, 2022 4:14 pm

If you feel this not a right place


Yes, that'd be my point - you already have a thread discussing the issue - cross-discussing two separate issues in one thread usually leads to all sorts of misunderstandings unfortunately because it's hard to be clear what response is to what issue. So let's do it in the other thread, not here. :D
Thorsten
 
Posts: 12490
Joined: Mon Nov 02, 2009 9:33 am

Re: Wind-shear

Postby wlbragg » Thu Aug 25, 2022 6:35 pm

I just thought the clouds looked cool!
Kansas and Ohio/Midwest scenery development.
KEQA, 3AU, KRCP Airport Layout
Intel i7/GeForce RTX 2070/Max-Q
User avatar
wlbragg
 
Posts: 7588
Joined: Sun Aug 26, 2012 12:31 am
Location: Kansas (Tornado Alley), USA
Callsign: WC2020
Version: next
OS: Win10/Linux/RTX 2070

Re: Wind-shear

Postby Delta5142 » Thu Aug 25, 2022 6:46 pm

jfc, you guys really can be the worst sometimes. Any thread that proposes a cool feature for FG ends up being a V12 vs. All argument. Every single one. Too much time is spent arguing that no progress is ever made. But who am I too speak, I'm just a user, not a dev. Maybe I should just do it myself, that's how an open source project works right? *sigh* if only every FG user knew how to code.

Kind regards,
Delta5142
Delta5142
 
Posts: 340
Joined: Thu Jul 02, 2020 4:18 pm
Location: Northern California
Callsign: DAL5142
Version: next
OS: Mac OS X

Re: Wind-shear

Postby Thorsten » Fri Aug 26, 2022 6:42 am

Any thread that proposes a cool feature for FG ends up being a V12 vs. All argument. Every single one. Too much time is spent arguing that no progress is ever made.


That somehow misses how progress is being made in the first place.

You seem to imagine that this works in a way that someone proposes a feature, it is discussed and then someone implements it. That's almost never the case though.

Take AW: The way this happened was that I had a fascination for cloudscapes and flying through them - and was unhappy by the rather homogeneous layers BW offered. So I started to work on some things to change that. The whole thing was met with violent opposition - too performance hungry, unnecessary, shouldn't be in Nasal, nobody needs that - that was pretty much the discussion that happened.

I just developed it on because I wanted to have it, and when it was sort of running, the first people realized 'hey, this isn't as slow as we thought, and it's kind of cool' - so then some helper routines appeared core-side to get things faster and the whole thing was committed to the repo and became official.

From my perspective - here's what proposing features looks like: I know meteorology pretty well. Not only flight meteorology, but also meteorology as a science. I have worked through a couple of books on atmosphere dynamics, and I have a background in numerical modeling, I know how to do fluid dynamics codes, what their properties are etc.

Unless you happen to be a professional meteorologist, chances are I know far more about common as well as exotic weather phenomena than you do.

In fact, I wanted AW because I am so interested in what happens in the atmosphere. So for most weather phenomena, I have already thought about how to implement them during the design phase. Case in point - the weather system will give you an only-red rainbow if the red light of a low sun falls under the right angle on a rainfront (Most people do not even know a phenomenon such as an all-red rainbow exists). It produces different droplet size of rain dependent on what clouds the rain comes from and produces different ice halo scattering dependent on humidity and temperature. It knows about lee effects and terrain elevations serving as convection sources, it has a complicated notion of gusty winds based on studying a few hours of real recorded gusts. It even provides diurnal ground temperatures for different terrain types, clouds and the sky so that an FLIR can show halfway realistic images.

Chances are that if you know it and it can be implemented at all, it already is. What isn't implemented is of the type that it would require a much more performance-hungry re-design of the whole system (not that I wouldn't have thought already about how to do that either - I did that a couple of years ago).

In fact, the system has (de-activated) slots for really complicated things - lee waves, complete convective cloud evolution cycles,... It has the option to drive it in different ways, for instance by inserting a complete and detailed pattern of aloft winds from a server or a high-level weather simulation.

A thing like windshear? It sometimes arises e.g. when passing a sharp boundary layer in strong winds in the mountains (that'd be one of the most common situation in RL), but if it happens for upper-level winds, it's too complicated to support automagically, so I discarded the idea a decade ago.

See - all the AW progress emerged from one person's desire to have a certain feature, the knowledge of what it should look like and the willingness to code. Feature proposals or discussions were completely irrelevant (or even made things more difficult).

And here we come to the issue of V12 - because all I have ever been interested is the offline simulation of weather. Compared with what I would like to know to model a weather situation, the information from METAR is ridiculously small. It's like 'here's 30 pixels - guess whom the picture shows. ' So compared to the offline weather simulation, METAR mode massively loses the complexity and details which make weather interesting for me.

In younger years, I (foolishly) let others talk me into writing a METAR mode for AW anyway - something I never wanted, never planned and which I know is not as faithful and high-quality as the offline engine. And there's not a year in which I haven't regretted that decision. Because magically that single act of kindness made it somehow my job to provide all the extended feature requests that have been coming.

So... here the reverse is true. The lack of progress on the METAR mode (basically the engine needs to fetch the five nearest stations, however distant they are, and always interpolate at least those) is not due to a lack of feature proposals and discussions - there have been plenty. It's not even due to a lack of tooling - core-side the problem of fetching other METAR stations besides the nearest ones is addressed.

Lack of progress is directly caused by the lack of interest of a single person - me. In the last couple of years, despite a clear path to develop, no single person has come forward and implemented a different station fetching scheme.

And so we go into the next iteration - but don't fool yourself that any amount of (even constructive) discussion will change the situation - it will change only when someone is interested enough to work to understand the issue and code a solution.
Thorsten
 
Posts: 12490
Joined: Mon Nov 02, 2009 9:33 am

Re: Wind-shear

Postby TheEagle » Fri Aug 26, 2022 11:07 am

Thorsten wrote in Fri Aug 26, 2022 6:42 am:In younger years, I (foolishly) let others talk me into writing a METAR mode for AW anyway - something I never wanted, never planned and which I know is not as faithful and high-quality as the offline engine. And there's not a year in which I haven't regretted that decision. Because magically that single act of kindness made it somehow my job to provide all the extended feature requests that have been coming.

I understand why you regret that, but without the Live METAR mode, AW would really be missing something. Despite it's limitations, it is one of the best things in FG !!! :D
Cessna 210 (Wiki)
My other aircraft: my wiki profile !
Other: FGTools (GitHub)
World tour: View on SkyVector
Please consider donating $1 / €1 to help me finance a new camera !
User avatar
TheEagle
 
Posts: 3413
Joined: Sat May 01, 2021 3:27 pm
Location: France
Pronouns: You, he
Callsign: F-EAGLE
IRC name: none
Version: Git next
OS: Ubuntu Studio 22.04

Re: Wind-shear

Postby S&J » Fri Aug 26, 2022 12:15 pm

So Eagle as you understand Thorsten so well, where has anyone asked Thorsten directly to do anything ?

All I see is users pointing out an issue, which was initially denied exist and after several posts I think we all now agree does exist.

I see a developer who wants everyone to know he wrote the original code and receive the kudos for doing so.
Knows the issue exists and has no interest in resolving it as it's an 'online' problem and has no desire to fix online issues.

Yet won't keep his nose out of the conversation and allow others to explore the issues and maybe develop a momentum to resolve it.

This is a reoccurring situation where this developer thinks that because they wrote the original code, they're the only person clever enough to understand the issues, and is personally responsible for improving it.
Surely in an open source community driven project others can take on what's been done and improve it ?
Without the original developers permission, maybe some pointers would be appreciated without disrespecting the efforts of others to investigate and understand the problems.

And it should be up to the community as to whether any changes are accepted or not.

But maybe you and others see things differently.
"Stay away from negative people.They have a problem for every solution." - Albert Einstein
S&J
 
Posts: 794
Joined: Wed Aug 26, 2020 7:31 pm

PreviousNext

Return to Weather

Who is online

Users browsing this forum: No registered users and 2 guests