Just flight BAe 146

[jon b 2,041]

7 hours ago, dbw1 said:

Ukraine?

No, most traffic seems to be avoiding  overflying the Ukraine at the moment. There’s often GPS jamming over the Caspian Sea for example, and the Middle East as Jan says. It’s not too much of a problem as the aircraft just continues with IRS data, however it takes out the ADS-B output so you’ll disappear from Flightradar24 🙂 but In a JF146 using the default FMC to navigate you’d have problems if you didn’t know how to use a VOR😂

 

3 hours ago, Janov said:

(I wish it was magenta, still)

😂 Your little digs at the Airbus always make me laugh when I read them.

Yes those single engine ferry guys must be mad.

I once worked as relay for one guy as we were about to go oceanic off the north coast of Scotland. He made a mayday  a couple of hundred miles out low over the Atlantic with his engine spluttering, but Scottish ATC couldn’t hear him.

ATC got him pointed towards Stornaway and scrambled the SAR helicopter, they asked me to ask him if there was anything else he needed, to which he replied “ yes, a new engine !’ 
 

I checked with ATC on the way home a couple of days later and was relieved to hear he’d made it to Stornaway and dry land ok.

 

 

[Murmur 3,941]

4 hours ago, jon b said:

There’s often GPS jamming over the Caspian Sea for example, and the Middle East as Jan says.

I wonder, do they also jam Galileo and GLONASS? And are current airliner nav systems able to use Galileo and GLONASS, anyway?

[jon b 2,041]

Very good question, the manuals I have for the 787, 747, and the Honeywell FMC only make reference to “ GPS” wether that’s being used as a generic term for satellite navigation or specifically the American GPS system isn’t made clear.  The 787 however is equipped for GLS approaches ( GPS landing system)  information on GLS mentions  the “global navigation satellite system “ and all 3 separate satellite navigation systems. 
 

So I’m not 100% certain but if I were to guess I’d say the 787 at least ,is possibly future proofed and can pick up all 3 systems. but as I say that’s just a guess, I stand ready to be corrected, I’ve not had any training  on GLS as yet.

To your first question, if someone has the intent for what ever reason to disrupt satellite based navigation my guess is they’d be doing it for all the systems available just to make sure.

 

[jcomm 5,888]

https://aerowinx.com/board/index.php?topic=6378.0

[jon b 2,041]

Found this reference in one of the 787 manuals about FANS:

Navigation - Satellite navigation utilising the Global Navigation Satellite System (GNSS). This is a combination of the US Department of Defense GPS and the Russian Glonass system (also a military-derived system).

So it would appear the answer is YES

edit to add after digging around...

The Honeywell GNSSU  ( global navigation system sensor unit ) which are a common fit tin Boeing, Airbus and everything in between can receive .....”GPS, GLONASS, Galileo, Beidou and other regional systems.“

Edited May 4, 2021 by jon b

[Bjoern 1,153]

Navigation talk aside: Does anybody already own it and how are the framerates?

[jon b 2,041]

Yes,  frame rates don’t seem to be an issue,  it works fine for me in VR with a 1080ti.

[AJZip 250]

Just flown from EGBB to EGNX in the 100 version in VR (HP Reverb Gen 1; Ryzen 5800X; nVidia 2070 Super) under ATC direction and including a successful ILS landing using the LOC and GSL capture...

...and it's simply Magnificent!

It took a while to familiarise myself with the cockpit and avionics - and I did make sure I read the manual (there are some quite important things you HAVE to know to be able to achieve certain things). 

I also experimented with my graphics settings.  I run both my CPU and GPU overclocked and can usually run X-plane 11 in VR at Medium Visual Effects (my rig really doesn't like HDR at all) and Maximum World Objects.  Well, for the JF BAe146 to be comfortable, and to ensure everything inside the cockpit and outside was crisp, I popped the World Objects down a notch.  With that adjustment the cockpit was a joy and, in flight, I was achieving an average 30-40fps, completely stutter free and with both GPU and CPU working hard but well within their limits.

It was, probably, the most immersive VR flight I've had yet - a nice blend of old-school and modern in the aircraft model itself and, I think, the modelling captures that superbly well.  The 146 model is one that I have flown in as a passenger many times and it had the feel and sounds that felt immediately familiar. 

Noted a couple of quirky things with the controllers (not able to transport within the passenger cabin was one and unable to manipulate the cockpit ipad) but that might be me.  I'll have a closer look at the manual and flag them to JF if needs be.  But nothing negative in anything that matters.  

Great job, Just Flight!  🙂 

[Bjoern 1,153]

I should have been a bit more precise: By how many percent does your framerate change when using the 146 compared to the default 172 (or any other default aircraft)?

It may sound like apples and oranges, but I figure a percentage is more representative than absolute figures without comparison.

 

 

By there way, there are some video tutorials from Joseph, one of the devs:

https://www.youtube.com/watch?v=SN6boYpAeVY&list=PLD5GlL3ffqlD47Bn1oybTJIaTj4Z5dcGy

[AJZip 250]

22 hours ago, Bjoern said:

Navigation talk aside: Does anybody already own it and how are the framerates?

Can't get back on it today, but tomorrow, I'll do an A/B comparison for you.

[AJZip 250]

 

19 hours ago, Bjoern said:

By how many percent does your framerate change when using the 146 compared to the default 172 (or any other default aircraft)?

It may sound like apples and oranges, but I figure a percentage is more representative than absolute figures without comparison.

Here are some representative snapshots of the fpsVR output for the two models, both on the runway (Orbx EGNX) and flying over the same spot of a small city at the same height (Orbx True Earth Scenery) at same time and same clear weather.

These are WMR captures of the view through the HP Reverb Gen1, 5800x CPU overclocked to 4.8GHz and nvidia 2070Super also moderately overclocked.  X-plane 11 Visual Effects at Medium (1 notch below HDR) and World Objects at High (1 notch below Max).

There's a lot of info on the fpsVR screens, but the most relevant are: the left hand FPS - while this is a snapshot, I've taken a representative one in each case; the graph and value of the GPU frametimes on the left; the graph and value of the CPU frametimes on the right; the GPU and CPU temperatures; the CPU average core and highest core loading numbers and individual core loading bar graph below the CPU output speed graphs on the right.

On the ground, end of runway, engine running...

Stock Cessna 172

JF BAe146 100

In both cases, this is after a few seconds 'settling down'.  For the settling down period, the loading on the CPU on the 146 is much higher but, once in a stable condition relatively close at 37fps as opposed to 41fps for the Cessna (c. 9% drop).

Average FPS on the right, by the way, is a bit misleading on a short flight like this - it depends how long you've been sitting at the less taxing runway.  So for this comparison, a very short flight, my choice of the snapshots to use showing the momentary FPS on the left is more representative.  

 

In the air, the 146 is now doing many more calculations, so as expected, it is the CPU that gets hit harder.  Same spot over same city at same height...

Stock Cessna:

 

JF BAe146 100:

 

Now you can see a bigger fps hit.  Again, while this is a snapshot, it is a representative one.  Cessna at 45fps and BAe146 at 29fps (c. 35% drop).

The GPU framespeed graph is very similar across the two, but the CPU framespeed is 'in the red' - which denotes that it is running below 30fps, dragging the total framespeed down.  You can also see the CPU has an average core loading of 67% (clearly this varied considerably on a snapshot basis) with the highest core here running at 81% loading.  So it is more a latency issue than pure CPU grunt.

 

But - when you consider just what the 146 is doing...that's a pretty darned good result.  

I say it again - great job Just Flight! 🙂

  

 

[danklaue 65]

That's interesting to see!  Just to provide a bit more insight into this, of course, you've already pointed out that the BAe146 is throwing a lot more at the entire system than the default Cessna.  Here are some data points that illustrate this:

 

• While the C172 has a total pixel count for albedo, night, and normal/PBR textures of around 250 million pixels, compared to the BAe's  800 million pixels.
• While the C172 has less than 100 click spots, the BAe146 has almost 500 click spots.
• While the C172 has a total mesh size of around 18 MB, the BAe146 has meshes totalling about 173 MB.
• While the C172 has 6 Lua scripts, averaging about  40-50 lines each, the BAe146 has almost 80 logic modules, totalling roughly 80,000 lines of code.
• While the C172 has a sound bank that's around 7 MB in size, with around 28 discrete sound events, the BAe146 has 170 MB worth of sounds, with about 160 discrete sound events.

 

Phew! This was an interesting exercise! I mean, I was also curious to look into, for instance, how many animations, keyframes, 3D lights, etc. were used in each plane, but some of these would be very difficult to tally up, and I've got updates to work on. 🙂 The file size may tell part of the story, but really, the amount of stuff we're throwing at our computers these days, expecting them to just crunch through those at even just 40 FPS! (I get 80 FPS inside, and around 120 outside with the BAe146, both on my iMac and my PC), I'd say we are getting a lot out of our machines these days.

 

This whole project had as one of its main focuses, optimizing stuff, and these optimizations we try to pass along as options that are configurable by end-users, so that they can have some control over which aspects of the plane can be neglected in favour of some performance gains... for instance, almost every logic module, registered in the "Manifest.json" file (which is configurable in a text editor), has a "Skip_Frames" variable, where you can determine which module should be running at lower frame rates, so that unimportant or FPS-heavy modules can be set to use less resources.  Editing a single module's "Skip_Frames" value may not make a huge difference, but it can add up, if you can do this for over 40 modules.

Or, for instance, unloading parts from memory that are not in view, dynamically.  When you are sitting in the pilot's seat, looking forward, the wings, engines, fans, stabilizer, etc. aren't drawn (or are replaced with lightweight meshes that only serve to cast shadows on the ground), but as soon as you turn your head or move around a bit, these parts are commanded to load up.  It's almost Quantum mechanical! 

It's these initiatives, along with many other per-module optimization initiatives that contribute to this plane's FPS performance.  

[Bjoern 1,153]

Wow, that's a whole lot more than I expected. Thanks, @AJZip and @danklaue.

 

Does anybody know if JF is planning to release a demo?

[danklaue 65]

Yes.

[AJZip 250]

2 hours ago, danklaue said:

The file size may tell part of the story, but really, the amount of stuff we're throwing at our computers these days, expecting them to just crunch through those at even just 40 FPS!

Indeed.

By the way, I should emphasise that this:

is while flying low over a city in the pretty realistic Orbx True Earth scenery and through VR at 2160x2160 per eye - and, even though the CPU is working hard, was as smooth as silk.  No stutters, no judders. 

One of the most immersive (once I'd moved the fpsVR pop up out of the way 😉 ) and enjoyable X-plane 11 VR flights yet.