1 hour ago, Bobsk8 said:

I agree, when I fly with FSLTL and its injector, I see aircraft all over the skies, not just taxing around the arrival and departure airport. Big difference. 

i admire your positive attitude, bob

3 hours ago, Axis3600 said:

And do you also need AIFLOW and AIGROUND?

No.

1 hour ago, Speedbird 217 said:

People have asked how BATC performs with AI traffic and it's the first time in my 20 years of simming that I get full airports with moving AI traffic and no notable performance hit.

Not saying I don’t believe this, but surely it would take some processing power. Are you saying there’s absolutely no impact on performance?

With AIG TC at 100% at Heathrow (more than 200 aircraft in the area), there’s an FPS of about 30-40.

20 minutes ago, Cpt_Piett said:

Not saying I don’t believe this, but surely it would take some processing power. Are you saying there’s absolutely no impact on performance?

With AIG TC at 100% at Heathrow (more than 200 aircraft in the area), there’s an FPS of about 30-40.

There is impact on performance, it's however significantly less than with the other injectors.

29 minutes ago, Cpt_Piett said:

Not saying I don’t believe this, but surely it would take some processing power. Are you saying there’s absolutely no impact on performance?

With AIG TC at 100% at Heathrow (more than 200 aircraft in the area), there’s an FPS of about 30-40.

I can't explain it, but the results so far I am not seeing FPS hits.  I am using AIG, FSLTL and FS Traffic models too.   I am using Lossless scaling, but even without it my FPS is not taking a hit. 

I am flying at capped 60fps...now I have top end system but I also have tons of addons.

Edited September 8, 20241 yr by knich

On 9/7/2024 at 5:08 PM, Speedbird 217 said:

On the ground in Hong Kong I explored via drone mode after parking up at my gate and spotted that there was a complete gridlock on runway 07R (used for departures as in real life). Long queue of aircraft trying to depart with one on the runway but just stuck there and aircraft waiting to cross the runway from both sides, even some facing each other from opposing ends of the same taxiway. I also spotted some aircraft stuck during taxi on the taxiways.

I can confirm that I saw similar behaviour in Dubai with a big conga line waiting for 12R with a Saudia aircraft just sitting there while the queue behind him grew longer. I cheated and just took the parallel taxiway to take the next intersection. It’s a pity as it was all very immersive up until then. 

1 hour ago, bennyboy75 said:

I can confirm that I saw similar behaviour in Dubai with a big conga line waiting for 12R with a Saudia aircraft just sitting there while the queue behind him grew longer. I cheated and just took the parallel taxiway to take the next intersection. It’s a pity as it was all very immersive up until then. 

I would bet on this being an AFCAD issue causing the problem. Something I wish that developers would take more care with.

How does the AI traffic flow without BATC? Same or similar problems?

3 hours ago, Cpt_Piett said:

Not saying I don’t believe this, but surely it would take some processing power. Are you saying there’s absolutely no impact on performance?

With AIG TC at 100% at Heathrow (more than 200 aircraft in the area), there’s an FPS of about 30-40.

There is some impact, especially using the slider set to max but considering the amount of traffic it injects I'm still amazed but how little performance it costs. Just did a quick comparison between no AI traffic (AIG models) vs PSXT Parked Only vs BATC slider at 10 and moving traffic at EGLL and facing central London in the background. I use Nvidia DLSS + LSFG, so the frames I get while flying are 4x of what's displayed in the FPS counter in the sim, capped at my screen refresh rate of 165.

EGLL no AI Traffic - 52 FPS

PSXT Parked Traffic only at 30%, 88 static aircraft displayed as per counter in PSXT - 32 FPS (-20 FPS)

BATC Live Traffic, slider set to 10, almost every gate occupied and manually counted 110 aircraft parked at the airport with some already starting to move (first few were down by T2 so not much action yet at T5) - 35 FPS (-17 FPS)

Those 35 FPS translate into 140 FPS using DLSS + LSFG. Pretty impressive considering there's 110 AI aircraft at the airport with a bunch of them moving around, taxiing, approaching and taking off. I haven't tested this with AIG Traffic Injector or FSLTL at those traffic levels but would expect FPS to be far worse than with any of the options shown here. Traffic with BATC was nicely spread across the airport too - I even had some BA aircraft parked by the engineering base and there were aircraft parked at T4 plus 2 DHL freighters in the horseshoe. As BATC uses FR24 for parking assignments it thankfully doesn't rely on accurate airport parking assignments in the scenery.

 

**Minor Update 1.2.40.Experimental**

 

```- Fixed some low wind ops values that could lead to arriving and departing in opposite directions.

- Fixed runway assignment direction issue if all runways were either low_crosswind or crosswind status.

- Fixed traffic not respawning after changing the traffic slider value or turning traffic off and then back on.

- Change the balance of parked traffic that has arrived versus due to depart, more traffic is now pending departure.

- Prevent traffic flights of the same registration being injected parked on the initial load of traffic, this was causing an aircraft that had arrived and departing again with a quick turnaround being present twice. In this scenario only the instance due to depart will now be injected.

-Fixed fatal error caused by traffic landing at a non-player airport.```
 

Did (only) two flights with the experimental version.

Having traffic now is great, very immersive! However, there isn't much traffic if you don't take off from the larger airports. Took off from ENHD yesterday and there was no traffic at all. But well, perhaps that airport is closed on sunday. Flying into EDDF on the other flight with the traffic slider at 10 gave me plenty of traffic though, haha!

However, the first problem I ran into was that when it gets busy it also gets hard (to sometimes impossible) to talk to ATC (using button control in my case). Just like with default ATC you can only 'talk' when the AI-traffic is silent so you have to wait (and wait and wait) until the right moment and then quickly 'talk' before someone else is talking again. Specially when it's important to contact ATC quickly this can become quite frustrating. I do hope BATC will change this to how FSHud does it: with hFSHud the list with possible things to say is always active and you can select an option even when ATC and AI are talking: you are simply put on hold and in line and ATC will react ASAP.

The even bigger problem though is that the basis stuff still doesn't work well, at least not during my two flights (as well as most of my pre-experimental flights).

In flight one BATC (as usual) didn't tell me to descend but it at least it gave me the STAR so I knew when to descend myself. But when I got close to the destination ATC led me 40 nm away from it... and then it told me to fly a given heading towards the IAF but that heading was totally wrong. It also told me to maintain 2000 ft (at 40 nm...!) which (no surprise) had me flying straight into a mountain. Luckily I paid attention and ignored most of what ATC said so I landed well regardless.

In flight two ATC didn't gave me the STAR at all and when I finally called for descend (because I was getting closer and closer to my destination) it was way too late to make it down in time and follow the given STAR. It also gave me the approach while I was almost at the airport already and the directions it gave me were useless.

So yes, it's nice to have traffic but what's the use if the basics aren't working properly yet... BATC is pretty good from start to cruise but from there on it's still very unreliable. And yes, I've only done two flights with this version but I did tons of flights with the previous versions and the second part of the flight always has been very unreliable. BATC clearly still is very EA and very experimental indeed. It has its great points and can be quite immersive but if you want reliable ATC you are better off using another ATC solution for now.

Question re the performance with AI traffic from BATC and other injectors: what does BATC do in such big cities like London? If you fly from EGLL, will the traffic on Stansted, Gatwick and City Airports be simulated as well? If not, you have the explanation why BATC has superior performance in such scenarios. With AIG (and FSLTL), you have every airport in your bubble filled up with traffic equally to the airport you depart from. And also if you fly over them en route. What does BATC do in this situation, e.g. if you fly from London to Rome and cross Paris? Will you then see an empty LFPG underneath you or will it simulate the traffic on all airports enroute as well?

13 minutes ago, AnkH said:

If not, you have the explanation why BATC has superior performance in such scenarios.

FWIW, I didn't notice any performance issues with KLAX and traffic. I did with KPHX, but I think it's because that airport is laggy anyway.

Where BATC is getting the performance boost is handing all the injection within the app. For instance, if you use FSLTL and FSHUD, you'll get a performance hit from FSLTL injecting the live traffic from an external source.

BATC, instead, uses a historical timeline of a week's average for your route. I actually like this idea. If I fly on the back end of the clock in New Zealand, but set my sim time to daylight I'll get more traffic than one based off what is current for the area.

5 hours ago, mistolip said:

 

So yes, it's nice to have traffic but what's the use if the basics aren't working properly yet... BATC is pretty good from start to cruise but from there on it's still very unreliable. And yes, I've only done two flights with this version but I did tons of flights with the previous versions and the second part of the flight always has been very unreliable. BATC clearly still is very EA and very experimental indeed. It has its great points and can be quite immersive but if you want reliable ATC you are better off using another ATC solution for now.

I found the arrival portion of the flight unreliable, so I just started ignoring everything from the start of the descent to landing. After a while, I just stopped using BATC all together. If they can fix the approach and landing, I probably will go back to it, because some of the features are nice but unusable for me in its current state. 

33 minutes ago, AnkH said:

If you fly from EGLL, will the traffic on Stansted, Gatwick and City Airports be simulated as well? If not, you have the explanation why BATC has superior performance in such scenarios.

First point: not with the current alpha. It only loads traffic for your departure and destination.

Second point: I flew into EDDF yesterday and it's been years since I saw my performance plummet down to below 20 or so. Don't know the exact number (I don't have an fps counter) but it was really stuttering all over the place. A lot worse than it does with FSHud. I had the traffic slider fully right though at 10 so I was getting all the traffic there could be: FSHud doesn't inject that much traffic. But you won't hear me say BATC has no impact on performance.

Performance on ENHD was absolutely perfect and awesome but that was because I encountered only one plane (during cruise) LOL

  

5 minutes ago, Bobsk8 said:

I found the arrival portion of the flight unreliable, so I just started ignoring everything from the start of the descent to landing. After a while, I just stopped using BATC all together. If they can fix the approach and landing, I probably will go back to it, because some of the features are nice but unusable for me in its current state. 

Yep, that's my view on it too right now.

Edited September 9, 20241 yr by mistolip

15 hours ago, Speedbird 217 said:

 Just did a quick comparison between no AI traffic (AIG models) vs PSXT Parked Only vs BATC slider at 10 and moving traffic at EGLL and facing central London in the background. I use Nvidia DLSS + LSFG, so the frames I get while flying are 4x of what's displayed in the FPS counter in the sim, capped at my screen refresh rate of 165.

EGLL no AI Traffic - 52 FPS
PSXT Parked Traffic only at 30%, 88 static aircraft displayed as per counter in PSXT - 32 FPS (-20 FPS)
BATC Live Traffic, slider set to 10, almost every gate occupied and manually counted 110 aircraft parked at the airport with some already starting to move (first few were down by T2 so not much action yet at T5) - 35 FPS (-17 FPS)

Aha that's quite different from your previous posts about no performance hit at all! 🙂  

The difference between BATC and PSXT is minimal:  -17 fps vs -20 fps! And don't forget you can get 100 live aircraft easily in PSXT on top of that without much extra costs.  
The number 160/165 you mentioned in previous posts and/or here is - apparently- due to other mechanisms. That was what confused me, one can of course use that with PSXT too...

BTW: I would have been surprised if BATC would do the injecting job much better then PSXT because both programs use the same method of injecting. We both have to use the SimConnect interface, that's all there is (unless they, like Pete Dowson in the past with PSUIPC, have "hacked" MSFS/SimConnect).
 

 

Edited September 9, 20241 yr by kiek
typo's