F/A 18C Hornet

[stu]

Next up is the smooth transition between the twin engine housings at the rear of the aircraft. Because it's a smooth transition I can't simply intersect two pieces of geometry, so there are a few steps to creating organized geometry that will allow me to add a beveled edge.

I started with an extruded curve that had been resampled to match the number of spans in that part of the aircraft:

Next, using the Cookie SOP, I was able to intersect the extruded curve with one of the engine housings (obviously I only have to do one side because the aircraft is symmetrical). It's worth mentioning that I could have gotten the same result with the Ray SOP by projecting the curve onto the engine:

This line describes where the two surfaces will eventually meet. Next, I deleted all of the polygons that were going to eventually get replaced by the new geometry:

Now I needed to create the geometry that will occupy that hole but will be more appropriately positioned to accept the bridge surface. Two of my previous steps come in handy here:

1. The engine housing originally existed as a nurbs surface, so I can use a Creep SOP to create new geometry that will properly adhere to the original surface.

2. I went through the trouble of making sure that the nurbs spans on the engine housing were evenly spaced, so any Creep SOP operations will be evenly spread out without any stretching or bunching up.

First, I created a grid with the Grid SOP that had the same number of polygonal spans as the hole that I had created in the engine housing and I crept it into position:

I created two curves that will eventually be skinned together and allow for the bevel. Referring to the curve that I had previously created with the Cookie SOP, I used an Edit SOP to push the points around while opening up a preview window for the Creep SOP so I could see the crept changes in realtime:

Finally, I combined the skinned curves with the grid, and deleted the polygons that I didn't need any more:

Now I have geometry that lines up with the rest of the engine housing and has a convenient polygonal path that will allow me to bevel the edge with a Polybevel SOP after I combine it with the flat piece that acts as a bridge between the two engines:

Merge the newly Polybeveled geometry with the engine housing, mirror the combined geometry, delete the centerline, make a couple of small tweaks, throw in a Facet SOP, and I think it's in good shape:

I may make a few tiny changes to the organization of the transitional geometry as it flattens out because the main fuselage has to make the same sort of smooth blend into the engine housings - we'll see how it goes.

[stu]

Here's a website that is providing some of my inspiration:

http://www.warbird-photos.com/walk-f18dhornet/index.html

Who knew that a jet could be so complicated?

[LEO-oo-]

Thanks again for the details - very impressive!

[Marc]

yeah very nice, I'm learning a lot from this thread.

[stu]

Me too.

[FrankFirsching]

I'm not sure I understood everything in your last step, but I have to say that your stuff is really enlightening.

Looking forward to what comes next ;-)

[stu]

If you have any questions, please ask! I try not to skip any major steps in my explanations, but, you know, I'm lazy.

[stu]

I thought I'd do some work on the "turkey feathers" - the metal plates that expand and contract around the engines:

I started by drawing a simple curve with the Curve SOP:

...extruded it:

...and then used the Polybevel SOP to create the round corners:

That makes up the basic shape for one of the metal "feathers". If you look at the photograph that I posted above, you'll notice that the "feathers" all have a curve in two axis that allow them to circle the engine nicely as well as overlap each other smoothly when they contract and expand. So, to accommodate the bends, I need some interior geometry:

Yup - the Divide SOP again. It works here better than it did in the elevator fin because of the nature of the shape of the "feather" - it's largely rectangular, so the Divide SOP didn't leave me with a bunch of orphaned triangles. I'm also not making a change to the "feather's" depth (although bent, it remains flat), so I'm okay with this solution in this case.

I decided to make the bends in each direction with two separate Creep SOPs. The reason I split the bends into two different operations was because it gave me better control over each bend after I assembled all of the "feathers" in order to make minor adjustments if there were any intersection issues. Laying thinly bent animating geometry so closely together can be a little tricky and can require some tweaking.

Creep 1:

Creep 2:

Now the "feather" bends in each direction.

Next, using a Transform SOP, I placed the "feather" into position behind the aircraft. I took special care to make sure that the pivot point was centered and at the base of the "feather" where it would attach to the aircraft. This is important because I'm going to use this Transform SOP's rotations to animate the expansion and contraction, and if the pivot is in the wrong place, the "feathers" will intersect:

And finally, using a Copy SOP, I made 11 more copies around the center of the engine:

Now I can animate the "feathers" by simply using the Transform SOP that I set up earlier to move everything all at once:

feather_animation.mov

Easy. There are a few additional elements to be included with the "turkey feathers" (other moving interior plates, etc.), but this is the basis upon which everything else attached to the "feathers" will operate.

[FrankFirsching]

If you have any questions, please ask! I try not to skip any major steps in my explanations, but, you know, I'm lazy.

I was wondering why you do all the work with the creep-SOP and building the curve on the crept grid? Wouldn't it be easier to offset the flat piece, either along the y-axis or using the peak-SOP. Then with the cookie-SOP in "Crease" mode get the curve on the engine. The original curve of the flat piece I would ray-SOP onto the engine. Now I have 2 curves, that i can polyLoft. Afterwards I would have to dissolve the extra edges, cause the polyLoft-SOP creates triangles. This would result in something that would be similar to what you have in this image, but with all the extra points needed to attach the grid like in this image. A second cookie-SOP that is used to subtract the area below the offset surface of flat part would leave the part of the engine, that is needed. The flat part is rebuild using the ray'ed version of the original curve and the original curve. Now I would have 3 parts I could merge and fuse together.

I hope this description was not too confusing. I just would like to know, why you chose the way you did it and if there are advantages of this method I didn't see.

[stu]

That's a good question.

The reason I chose to use the Creep SOP the way I did was because I wanted the two curves that would eventually be skinned together to be largely parallel to each other as they existed on the surface of the engine housing. Offsetting the flat piece and performing a Cookie SOP operation wouldn't work in this way because the engine housing surface that it would have to intersect with makes a 90 degree curve from the side to the top, and the two intersected lines would converge in either the Z or the Y axis depending on which way you offset the flat piece.

I liked that I could control exactly how far apart those two lines were without have to worry about maintaining the original surface (the Creep SOP guaranteed that all of the points lived on the original surface) and that I got exactly the shape I was modeling in the flat grid with the two curves after it had been applied with the Creep SOP.

The Cookie SOP idea would have definitely worked if the engine housing surface didn't bend 90 degrees from the side to the top.

Thanks for the question!

[stu]

Step by step coming shortly:

[maharzan]

This looks great and the details are very nice.. I will follow along soon..

Thanks STU for putting this up..

[MatrixNAN]

Hey Stu,

Great work so far you are really nailing it down. I am sure when you finish it will be perfect like all your other models. Your a great modeler. I am loving your attention to detail. Keep up the great work. Can't wait to see this finished

Cheers,

Nate Nesler

[Jordan]

hey stu,

very good houdini modelling tutorial from you.

Have you thought of making a tutorial video huh? esp. for sucker in modeling like me.

looking forward to see the finished piece.

Jordan.

[stu]

Have you thought of making a tutorial video huh? esp. for sucker in modeling like me.

It's crossed my mind. Most of the houdini modelling tutorials (although useful) deal with simplified subjects - I wanted to make something that would apply to real production work.

Thanks all for the kind words, much appreciated - update coming soon.

[Jordan]

It's crossed my mind. Most of the houdini modelling tutorials (although useful) deal with simplified subjects - I wanted to make something that would apply to real production work.

Thanks all for the kind words, much appreciated - update coming soon.

Yes, please. A houdini advanced poly/nurbs modeling DVD tutorial.

I will be glad if u make it free and post it to SESI tutorial page.

[JDenker]

i really like the progress so far. thanks for the explicit explantations!

thats great.

Cheers

-J

[Guest Swann]

i really like the progress so far. thanks for the explicit explantations!

thats great.

Cheers

-J

Do you plan some advanced shading for this model ? You know, ADVANCED shading not just advanced . It will be nice to see it.

[schyzomaniac]

Very nice work and very kind of you to take the time to show your steps in this thread.

Thank you - I learned a lot from this.

Daniel

[stu]

Thanks again everyone, much appreciated!

I haven't thought much about shading, but I could certainly go into that. The only downside might be that, as a hardbodied subject, the shading won't be all that complex (mostly paint with some plexiglass, metal, rubber, and fiberglass here and there). I think that there are some cool rigging opportunities - not only with the landing gear and shocks, but with deforming tires and maybe automatically animating the control surfaces (flaps, elevators, rudders, etc.) based on the motion of the craft.

Update coming soon, I promise.