Lately  have been using the gtn750 a lot and accessing the METAR for the waypoints and nearby airfields is usually reasonably close - though not always.

5 minutes ago, Glenn Fitzpatrick said:

Lately  have been using the gtn750 a lot and accessing the METAR for the waypoints and nearby airfields is usually reasonably close - though not always.

I use ASN4 for the same thing and it's reasonably close--you just have to enter the airport into the CONDITIONS field quickly before it realizes it's not interacting w/ P3D.

2 hours ago, JRBarrett said:

The only time this is not the case is when the LiveWeather server gets stuck providing 3-day-old weather as happened two weeks ago. 

Maybe that's what's been happening, then, and I've just had unfortunate timing. I admit I don't do all that many longhaul flights these days because they usually either end in the memory leak framerate bug or a crash (software, not airplane). But the few times I've done them over the past few months, the winds aloft have been wildly divergent from reality.

2 hours ago, JRBarrett said:

LiveWeather does indeed use real METAR reports for surface wind direction - so at least you can plan (in advance) for which runway you will be landing on. But, clouds, precipitation and visibility might be very different than what the current METAR shows.

I'm forced to wonder why, if it's reading METARs at all, it doesn't read the whole METAR and adjust weather accordingly.

 

11 minutes ago, eslader said:

the memory leak framerate bug...

Has this ever been substantiated?  My 260nm flight in Deutschland just ended in a stutterfest and I had very modest CPU/GPU/VRAM load as well as a 26ms server connection.  Haven't seen this level of dysfunction for awhile now.

Edited May 6, 20215 yr by Noel

1 hour ago, Noel said:

Has this ever been substantiated?  My 260nm flight in Deutschland just ended in a stutterfest and I had very modest CPU/GPU/VRAM load as well as a 26ms server connection.  Haven't seen this level of dysfunction for awhile now.

I think there are a number of causes for framerate issues right now. For me, if I fly in certain areas (the Fort Lauderdale/Miami area is a good example), my ram usage goes from a norm of around 7 or 8 gigs to 20 or more. Whereas normally ram usage fluctuates up and down, once the leak hits it only goes up. Sometimes I can fly out of the affected area and then ram use starts going down again, and once it drops below 14 gigs or so things smooth out. But more often, I can't fly out of the impacted area before it slows to 1 or 2 fps and finally crashes to the desktop. I was able to escape the leak bug in the Fort Lauderdale area once by flying well east of the Bahamas, but that was in an overflight of Florida at FL400, so I was moving a lot faster over the ground than the other times when I haven't been able to escape it before a crash.

Further pointing to the idea that it's a memory leak is that if I start a flight in Fort Lauderdale, everything's smooth and memory usage is normal. But as I fly around the area, the ram use continues to increase, even if I just do circles over the airport, and a few minutes into the flight it starts getting noticeably jerky.

None of the usually suggested fixes (disable multiplayer, turn off various graphics settings, etc) helps the problem at all. And when the problem hits, even the settings menus are very slow to respond, and they should not be affected by other players or graphics settings. 

I have discovered that the eastern Caribbean islands are reliably smooth, but if I try to head toward South America, it only stays smooth if I stay to the western side of the continent. If I head toward, say, Guyana, I'll start stuttering right as I hit the Venezuelan coast near Trinidad, and will be down to a slideshow long before I get to the Guyana border. On the other hand, crossing into Venezuela in the Caracas area, it stays smooth.

So I believe there's a memory leak that's location-based, which suggests to me that they messed something up in the scenery files after update 3, or in the way the system handles certain scenery files.  But I don't understand the commonality between the bad areas as far as scenery files goes, because the scenery near Caracas is really similar to the scenery in the same country near Trinidad. And I get a stutterfest around the Denver area as well, which has very different scenery from the Miami area. 

 

Edited May 6, 20215 yr by eslader

2 hours ago, eslader said:

Maybe that's what's been happening, then, and I've just had unfortunate timing. I admit I don't do all that many longhaul flights these days because they usually either end in the memory leak framerate bug or a crash (software, not airplane). But the few times I've done them over the past few months, the winds aloft have been wildly divergent from reality.

I'm forced to wonder why, if it's reading METARs at all, it doesn't read the whole METAR and adjust weather accordingly.

 

I assume because the cloud graphics system is tightly integrated with the MB model, which covers a very large horizontal and vertical area with multiple layers and weather “cells”. If the model indicates a solid overcast at 3000 feet AGL over an airport, and the current METAR says “clear” you would probably end up with a circular “hole” in the model-generated cloud deck surrounding the airport. Interpolating clouds between METAR and the model forecast is undoubtedly an extremely complex task, that could lead to some really unusual graphics, so I am not surprised they have not tackled this area yet. Using METAR for wind, temperature and pressure is easy, because those values have no visual impact (except perhaps the wind effect on vegetation or water bodies)

Edited May 6, 20215 yr by JRBarrett

There’s also the issue that a METAR and ATIS are reports that usually appear at the top of the hour. The weather could significantly change by the time a new METAR/ATIS is issued. 

1 hour ago, amahran said:

There’s also the issue that a METAR and ATIS are reports that usually appear at the top of the hour. The weather could significantly change by the time a new METAR/ATIS is issued. 

SPECIs are generated much more rapidly though per hour.  Perhaps around every 6 minutes a new SPECI can come out(basically an updated METAR due to rapidly changing conditions).

1 hour ago, amahran said:

There’s also the issue that a METAR and ATIS are reports that usually appear at the top of the hour. The weather could significantly change by the time a new METAR/ATIS is issued.

Where I fly the  METAR ( and ATIS) is issued 20 and 50 minutes of the hour. And you will have  a METAR "SPECI" if weather deviates from previous observation - to avoid the issue you brought up.

Edit: @ryanbatcund, with the quicker draw 🙂

Edited May 6, 20215 yr by SAS443

Anyone not noticed the missing visibility settings in the sim? Aerosol density is completely missing from live weather. Its either cavok or clouds on the ground situation. Humidity parameters are not present in the sim to depict low visibility and haze. Is the weather engine not getting this info from Meteoblue ? I have seen multiple threads on this on the official forum aswell but no answer from the devs. Even FSX had visibility sliders. Besides this aerosol mumbo jumbo should be get rid off. There is no such thing as aerosol density in aviation. The phenomenon is knows as visibility to pilots.

2 minutes ago, Baber20 said:

. The phenomenon is knows as visibility to pilots.

This was said during alpha over and over again.  

17 minutes ago, Baber20 said:

Aerosol density is completely missing from live weather.

This is a personal bugbear for me. Have they said anything about fixing it?

8 minutes ago, scotchegg said:

This is a personal bugbear for me. Have they said anything about fixing it?

In the last Q&A they said that they plan on rebooting the atmospheric simulation. This is a limitation in their current weather engine. But its a long range fix which means by the end of this year.

Now that last part really surprised me. So basically they are going to take 6 months to get such a crucial weather parameter right ? Makes you wonder where their priorities are. 

@ 49:38

 

6 minutes ago, Baber20 said:

Makes you wonder where their priorities are....

I would guess their priorities are aimed at their perceived new market, console users who, on the whole, are going to be very likely far less critical of simulator fidelity, which fits with making scenery a top priority.  And then there is the DX-12 port which I thought I read was important for the console port.  All of the fabulous work on scenery will benefit all users, experienced PC simmers and newer console users, whereas maturing the SDK/Simconnect piece will mostly benefit the more hard-core among us.

6 hours ago, ryanbatcund said:

SPECIs are generated much more rapidly though per hour.  Perhaps around every 6 minutes a new SPECI can come out(basically an updated METAR due to rapidly changing conditions).

 

6 hours ago, SAS443 said:

Where I fly the  METAR ( and ATIS) is issued 20 and 50 minutes of the hour. And you will have  a METAR "SPECI" if weather deviates from previous observation - to avoid the issue you brought up.

Edit: @ryanbatcund, with the quicker draw 🙂

Eh, I’ll take my flying experience in Southern Ontario. The ATIS at 7PM in the summer can call for winds at 12 gusting to 18, but by the time you’re on final at 7:45, the tower wind check can be no more than 3 knots. That’s just an artefact of the dying thermals in the late afternoon.

And in my experience, these wind drops don’t get recorded anyways until the next scheduled ATIS since they’re not significant enough of a hazard to justify a SPECI outside of the top of the hour.

SPECIs do exist, I agree, but they aren’t used for changes that are considered minor in nature.

(On that note, I do recall a SPECI ATIS being issued after a 00.06” drop in altimeter setting...but that means that the ATIS was up to 00.06” off from actual conditions before the ATIS was reissued).

Before anyone takes offence to what I wrote, just remember, these are real-world observations from my own flying experience.