Nick Black's F-14 AI conversion project.

[DC1973]

Hi folks,

I'm delighted to say that I'm now working on the conversion of Nick's F-14 for FSX / P3D. It's going to help me learn how AI are done ready to convert my P-61C Black Widow and RF-61C Reporter when they're ready into AI, and I didn't think I'd ever get to work on a Tomcat as there are so many good quality payware and freeware models already out there.

Things are progressing well, with the excellent ModelConverterX allowing me to convert and retain some of the original animations and XML conditions applied to the model. There's some work to be done though as, for instance, the pilots have lost their heads, ( literally! ) and the aircraft won't taxi following the correct paths etc. This may be beyond the model's control though so if any of you know why this is happening please do let me know.

I'll post regular updates here as I go along. Plans for this awesome AI include a couple of minor additions to the model ( ejection seat handles, reflective pilot visors etc ), full use of materials and alpha channel to enhance appearance and take advantage of FSX / P3D graphical functions, ensuring that all existing textures remain functional ( including third-party ones ) and making a few tweaks to the textures myself.

Really enjoying the process, so here's a picture of Miramar's flight line after the first proper conversion attempt.They fly well and swing wings, wheels, undercarriage and most other animations are working fine

[Firebird]

The pilots heads are easily explained.
They use tick18 animation which FSX and later doesn't use. The "Tick18" variable has been depecrated, and replaced by the "Ambient" or "NonRandomAmbient" variables. So you need to alter the animation as described in the FSX SDK.

The taxying thing is a touch worrying as I have never heard of that before. How are they not following the path?

[DC1973]

Cheers Firebird, I suspected that the heads were using Tick18 - it should be quite easy to rectify.

It's the taxiing, as you say, that's odd. They land at Miramar OK, then taxi off the runway fine, but after that they slow right down and just head off in a straight line. I found two of them a couple of miles outside of the base, taxiing happily along through open bush land.

Likewise, they start engines fine, but then taxi very slowly and are reluctant to turn. I wonder if it's a nose wheel steering issue as I don't recall seeing that in MCX when I checked the animations - no c_wheel in the list. I'll keep investigating when I'm next able to and will report back with whatever crops up.

[Firebird]

When I am following a model on an afcad I always load it into AFX as you can see the aircraft you have selected move around the afcad. I am not sure if ADE has a similar function but I would be surprised if it doesn't.

I would check to ensure that it is not picking up a duplicate afcad.

The turning issue would not be anything to do with the conversion, after all a lot of the original FS9 AI models didn't have nose wheel steering at all. The only thing that can affect it following the afcad would be the fde, and most probably the cfg file.

Are you using the original F-14 fde or have you modded it in anyway?

[DC1973]

No changes made to the FDE or .cfg files at all yet, just mdl conversion to FSX standard. Interesting.... The airplanes fly okay as I watched one land.

[Firebird]

OK, either there is some massive difference between the AI engines of FS9 and FSX or there is another factor at play here.
Best to ensure that you only have one afcad for Miramar. We made several over the years so it must be a possibility.

[DC1973]

OK, either there is some massive difference between the AI engines of FS9 and FSX or there is another factor at play here.
Best to ensure that you only have one afcad for Miramar. We made several over the years so it must be a possibility.

The un-altered, FS9 Tomcats operate normally: only the FSX format Tomcats won't taxi correctly. I checked using a search for AFCADS and all files with KNKX in the title, and there are no extras present - the only scenery active is MAIW Miramar. I did have a payware Miramar active previously, but as the FS9 aircraft are fine I don't think this is the issue.

There'll be something behind it all, will keep digging.

[Firebird]

Hmmm. Definitely interested in finding out what is causing this.

[gsnde]

Have you checked and corrected the model bounding box with Arnold's Radiator tool? Not sure if this is fixed by now, but not long ago this value was massively increased during conversion and needed to be resized.

To do so you can check the values of the FS9 model in MCX and then load the converted model in Radiator and apply them there if they differ.

Could very well be the explanation.

From mobile hence short

[DC1973]

Have you checked and corrected the model bounding box with Arnold's Radiator tool? Not sure if this is fixed by now, but not long ago this value was massively increased during conversion and needed to be resized.

Martin, you're a genius! I completely forgot about that! The F-14 had been given a radius of 7,000 meters!! As it's AI of course there's no preview window in FSX so nothing to warn me. Well done mate, I bet that sorts it - will report back about it tomorrow when I get a chance to test it.

Two more shots of work done this evening - pilots now back in place with glossy visors, ejection handles added, cockpit panels updated and the materials in play with subtle reflection and specular shine. All that, and the model is only 120kb larger with nothing else to add

[DC1973]

I couldn't wait and had to check, and can confirm that using the RADItor tool fixed the taxi issue. All the airplanes are now behaving as they should. I need to work on the visibility for certain model parts such as wheel chocks and RBF tags, but it was fun watching the revised Tomcats in the circuit at Miramar. All other animations seem to be working perfectly without any further input from me

[gavinc]

Two more shots of work done this evening - pilots now back in place with glossy visors, ejection handles added, cockpit panels updated and the materials in play with subtle reflection and specular shine. All that, and the model is only 120kb larger with nothing else to add

Hi,
re the size - if you strip out some of the intermediate LODs then you should get the size down significantly. Testing has shown that P3D (and FSX) don't need all the intermediate LODs that FS9 did. What is more critical to minimizing the frame rate hit of AI is minimizing draw calls.

Gavin

[gavinc]

BTW - welcome to the frustrating, addictive world of AI

[DC1973]

BTW - welcome to the frustrating, addictive world of AI

Cheers Gavin!

That was going to be one of my next questions, how many LODs? The F-14 currently has no less than seven, but I've read here that msot folks are creating about three now. Does that sound about right?

Cheers,
Dean

[gavinc]

Hi Dean,
re LODs, 3 sounds about right, I usually take the top LOD (obviously) then the lowest LOD that has 3d shape (quite often the last LOD is just flat fuselage & wings). Then I try to find a one of the middle LODs that has most of the major components but non of the details.

Hope this makes sense - Gavin

[DC1973]

Awesome, sounds like a common sense approach. I'll do that, cheers!

[Victory103]

Thank you, thank you from a Tomcat fan! On the moving crew heads, while airborne with the wings swept back, anyway to have the pilot hand raise with middle finger extended?

[clickclickdoh]

... anyway to have the pilot hand raise with middle finger extended?

That scripting only works while inverted.

[Firebird]

Of course you will also need to add the flash from the camera.

[DC1973]

Thank you, thank you from a Tomcat fan! On the moving crew heads, while airborne with the wings swept back, anyway to have the pilot hand raise with middle finger extended?

That would be awesome! I am considering having helmet visors that raise when the airplane is below 50 KIAS but I don't want to overdo it as the model is great as it is... At the moment I may have to re-build the central fuselage as several visibility items are stuck in place and cannot be separated at present using MCX. It will allow me to correct a tiny flaw in the design though, where the RIO's seat is too far forward.