Posted on Leave a comment

Problems – October 2013

These are the major unsolved problems I am working on right now. They are related to both the final update of the FSX version (4.1) and to the planned P3D version (5.0).

Multiplayer

There are a number of problems with Draken in multiplayer, one of them being that animations and fx are not working from the other player’s POV. The reason behind this is that Draken contains custom animations and functions that are compiled into the model. These will not carry over into MP,  you only get the animations that are also present in the model you are flying. So other players flying Draken will see the custom animations, but not players in other types of aircraft. The only way around this is to rewire the custom animations to stock animations. I am working on this but it will take some time.

Another problem is that smaller parts of the model disappear as you move away, regardless of which LOD is currently active. This is due to the somewhat erratic LOD handling in MP, where the SmallPartRejectRadius entry in fsx.cfg takes precedence over the LOD:s. There are numerous posts on the Internet saying that this is deprecated since SP2, but this is not true. If the entry is missing from fsx.cfg in SP2/Acc, the parameter will default to 1.0 (pixels) according to the SDK – although in reality it seems to be more like 2.0. Setting the value to zero will render “everything” but you will also take a minor/major performance hit depending on your hardware, number of players, etc. My advice is to try values between 0.0 and 2.0 and find a good balance for your MP environment.

Afterburner

In Draken the afterburner is activated manually with a key command, but shut off automatically as soon as you move the throttle back. This is achieved by setting the afterburner_throttle_threshold entry in aircraft.cfg to >1.0. This unfortunately also means that the throttle setting will automatically go to the threshold value as you hit the A/B key, which in this case means full throttle.

I discovered that the reason for this is that afterburner_throttle_threshold is not simply a threshold value for triggering afterburners, it’s actually a scalar for the throttle tables in the airfile – but only when afterburners_available is set to a value greater than zero. For example, a value of 0.8 means that you will get maximum fuel flow (and N2) at 80% throttle position. Now a single-stage afterburner (afterburners_available = 1) will ignite at 99% throttle, regardless of the threshold value. With multiple-stage afterburners, the activation threshold is lower. I have not yet figured out exacly why and how these factors work together. But the gist of this is that you cannot separate afterburner from throttle position in FSX. In the real Draken you could adjust the throttle while on afterburner, which could be crucial in formation flying. In FSX this is not possible. With afterburner on, you are stuck with the threshold value. It took me years (!) to discover that this is a limitation in FSX itself.

Another problem is that in multiplayer, other players will not see your afterburner fx since they are tied to a custom function compiled into the model (see above). I therefore tried to rewrite the A/B logic using only the stock functions and animations. This turned out to be a major headache for me, as the criteria for A/B actuation are quite complex in Draken compared to (for instance) the stock F/A-18. I am still working on this but I doubt that I will ever get it completely right…

Radio frequency settings, etc.

People are often complaining (politely) about the fact that the radios in Draken are tied to local variables, which apparently poses a problem when using ATC or when flying in a VA environment such as VATSIM. Now, the radios were among the first gauges I developed for Draken, more than a decade ago, and they were designed for the single-player flying that I was doing at the time. I have since done some changes and simplified the radios so that they respond to FSX key commands, but there is still some work to do. My goal is to have most of the radio and avionics settings using standard variables. Working on this, too.

So, a lot to do, as always.  ;)

Stay tuned  /Tom

Posted on 1 Comment

Not dead yet

Hello again, whoever is listening.

Draken development has been low-priority for a while due to personal and family problems. But the agenda still holds, I am hellbent on moving my simulation of one of the most wonderful jet fighters ever into the the next-generation world of flight simulation. Pretentious? Maybe. But that’s me.

I have now finally moved from modeling in GMax to 3DS Max with the new 64-bit Prepar3D SDK. This means that I have a lot more freedom in modeling, and more powerful tools for texture mapping. Texturing is thus getting a radical overhaul, primarily to allow for higher resolution (and more flexible) repaints.

The virtual cockpit is also being almost completely redone, mainly because VR is getting to be a big thing in simulation games. Lockheed Martin are talking about Oculus Rift compatibility, and there are a lot of users already using different 3D solutions. So I realise the need to come up with a cockpit environment where you can actually move around in three dimensions and still have that “sense of immersion” .

Any screenshots? Not right now. I really need to shape up and get the new texture mapping done first. Then, maybe.

Please stay tuned,

/Tom

Posted on

Airspeed issues solved

One of the more annoying quirks of MSFS is how it (mis)calculates indicated airspeed from true airspeed. Unless you do some tweaking in either the .air file or in Aircraft.cfg, your IAS will be incorrect. The recommended way to deal with this in FSX is by tweaking the airspeed_indicator.n entries in Aircraft.cfg, which will set a scalar value and a base offset value for IAS calculation.

To my horror, I just discovered that in Draken (FS9 and FSX versions) I had set the airspeed_indicator.0 values way wrong, resulting in indicated airspeed in Draken being about 100 kmh too low. This of course negatively affects things like flight plan calculation and formation flying. Furthermore, since I have calibrated the alpha incidence angle (aoa) curves in the FDE to IAS rather than TAS, the whole envelope is slightly off. This is the main reason why I have failed to get realistic performance at high altitude.

Changing airspeed_indicator.0 and cruise_lift_scalar to the following values were successful in my case:

airspeed_indicator.0 = 0.93, 8.0
cruise_lift_scalar   = 1.90

If you feel like experimenting, use these values as a starting point. Please note that other parameters (lift and thrust scalars) may also need some tweaking to get realistic performance.

Thanks to Johan at Novelair for reporting the airspeed error.

Posted on

Progress

work

Finally managed to export the complete 3d model from GMax into 3DS Max, which means that I can start to remodel and retexture my baby the way I always wanted to. Expect more detail, higher resolution textures, and a better overall experience.

Kudos to Tom Faulds, Manfred Jahn et al who made this possible. See this link for more info about the export script.

The new model (5.0) is developed concurrently with the final upgrade of the FSX version (4.1) which is still modeled in GMax.

The target platform for the new model is Prepar3D, although it will be FSX compatible. Repaints that you have made for the 4.x version will not be compatible with version 5.0 for obvious reasons.

Screenshots of the new model will be up as soon as I get some texturing done.

Posted on 3 Comments

Back in business (sort of)

Well, after a two-month break I am back in business… again… ;)

The 3D landing lights are on the back burner for the time being, as they turned out to be a lot more work to implement than I thought. I need to address some higher-priority stuff first (bad texturing, high altitude performance, poly count), then we’ll see. Not a big issue I think.

The high poly count has been a source of complaint from many users who fly in multiplayer, and I recently discovered that the weaponry could have been modeled a lot more efficiently – although weapons are currently only visible in the highest LOD (which in itself is stupid). The missiles, rockets, etc. were modeled at an early stage of development, and were conveniently forgotten when I started optimizing the model for release. Sidewinders and Falcons are around 300 polys a piece, so no big deal as you only have 2 or 3 of them. But the 12 A/A rockets add up to 5200 polys… All in all, a lot of unnecessary polys. As I started to clone the weapons to include in lower LODs I realized that they would have to be optimized or remodeled completely. So, some work to do there.

High-altitude engine performance has always been a problem in both the FS9 and FSX versions of the flight dynamics. The current FSX release (j35bm_fsx_14.air) is a compromise which gives reasonably realistic performance up to about 5000 meters, then gradually starts to lose power until the aircraft stalls at about 12 km altitude. In reality, maximum Mach should be attainable at ~10 km, and the effective ceiling should be about 15 km (clean). The problem lies in the lousy simulation of turbojet engines in the FS flight model, which you need to circumvent by a lot of “cheating” with unrealistic engine speeds, etc. Still working on this, but it is extremely time-consuming work. And boring.

I really have no idea if anyone is actually reading these sporadic blog posts… Regardless, I will continue to report on what is going on in the project, and maybe open up a download section for development versions and related stuff.

Stay tuned   /Tom

Posted on 1 Comment

Status report

The first post of the year, and the first “real” report on the project since 2011. :)

The biggest change so far is that I have added 3 more LODs to the model, which makes for a far better experience in Multiplayer. Still too many polys in the lower LODs, though, but I am working on it… Right now there are 6 LOD levels: 400, 150, 75, 35, 20 and 5.

I spent most of last week rewriting the brake chute function from scratch. It is now working more or less as I intended originally, and is tied directly to the flaps position variable without local vars inbetween. Also, the chute can now only be fired once, the “flaps up” command will no longer reset the animation. If you deployed by mistake (it happens), you can still cheat and reset it by pressing the indicator in the cockpit view. The chute is now also fully animated in LOD 150, with visible cords etc. Looks a lot better in MP.

The drop tanks are another headache, they never worked very well and the fuel system as well as the animation need to be reworked completely. Definitely on the “to-do” list.

At the moment I am working on adding 3D landing/taxi lights (screenshot below). Looking good so far, better than the default lights at least. Thanks to Bill Leaming et al for this.

130130-007130130-005

So, what else? Well, I have finally started dabbling with SimConnect, which I somehow managed to avoid so far. The possibilites are supposedly endless with SimConnect, although I can see some limitations.

Oh, I also (finally) added a cannon effect which you can map to your controller of choice. You can actually pull the trigger and hear/see something happen. Missiles also coming up.

/Tom

 

Posted on 1 Comment

Draken in Prepar3D

I have been experimenting for a while with Prepar3D (P3D), the continuation of the retired FSX/ESP platform, which is published by Lockheed Martin and directed towards professional and academic use.

Since P3D is still basically ESP (the pro version of FSX) with a different UI and some minor functional changes, it is hardly surprising that Draken works more or less out of the box. In my first side-by-side testing Draken actually flew smoother in P3D than in FSX, although this might just be because P3D is a fresh install.

Being meant (and priced) for pro users, and with promises of actual technical support (!), P3D would seem to be the natural choice for cockpit builders and other pro or semi-pro users. Well, we’ll see.

Anyway, I will continue testing P3D and report my results here.

/Tom

Posted on 2 Comments

Welcome!

This blog is basically a revival of my old Draken devblog. The main difference is that I am now using WordPress, which is (hopefully) more secure than the heavily customized Joomla system I used on the old website.

Here I will blog about the continuing development of Draken for FSX/P3D, including the customizations I make on demand for both desktop and cockpit simulators. Feel free to comment in English (preferred) or Swedish.

/Tom