### Re: Current Projects

Posted:

**Sat Apr 09, 2016 8:55 pm**I think the skittles might more look like eccy pills tbh.

The best English YSFlight fan community since 2010.

https://forum.ysfhq.com/

Page **1** of **1**

Posted: **Sat Apr 09, 2016 8:55 pm**

I think the skittles might more look like eccy pills tbh.

Posted: **Mon Apr 11, 2016 8:18 pm**

Taking a break from Albatrosses 24/7. I'm painting up my beauty machine for AirFest, one last time. Trying a couple of things out with decals

Posted: **Tue Apr 12, 2016 7:28 pm**

Decided to changed up the coloration style on one of the wings of the Super Pitts:

Posted: **Tue Apr 12, 2016 7:37 pm**

Checkout my current project here viewtopic.php?p=97536#p97536

Posted: **Wed Apr 13, 2016 10:41 pm**

Another l39! W.I.P. Got to do the crew and correct a mistake I've made with the intake color. This plus a re-paint of the YSF F86 in as close as I can get to the 'Rampant Horses' scheme. They were an Italian team in 1956 - 57, according to a book on aerobatic teams I have. Only things wrong are the yellow spots, and part of the tailplane on the underside, as they should be the dark blue. But due to how the AC was colored originaly, It comes out like that. As normal all done in NOTEPAD.

Posted: **Tue Apr 19, 2016 6:21 am**

First tests of Аn-26 with C-130 dat file.

(mesh will be smoothed, fixed and animated soon)

(mesh will be smoothed, fixed and animated soon)

Posted: **Tue Apr 19, 2016 6:45 am**

Looks great! hope the cockpit also looks nice later.....

Posted: **Tue Apr 19, 2016 6:36 pm**

Hercules XV209 in 1970s desert camoflague. Have to say the early RAF schemes were mighty fine.

Posted: **Tue Apr 19, 2016 8:22 pm**

wow! this looks amazing...well done neil!Neil² wrote:

Hercules XV209 in 1970s desert camoflague. Have to say the early RAF schemes were mighty fine.

Posted: **Thu Apr 21, 2016 3:19 am**

Same here. Those were the days of free expression,Neil² wrote:

Hercules XV209 in 1970s desert camoflague. Have to say the early RAF schemes were mighty fine.

Posted: **Sat Apr 23, 2016 10:40 pm**

For a future map release

Posted: **Sun Apr 24, 2016 12:26 am**

Looking gooooooood buddy, I hope you add some water and trees and hills.

Posted: **Sun Apr 24, 2016 12:43 am**

Out standing.May I suggest Patrick's PNMV current ground pack for the details Kujo.

Posted: **Sun Apr 24, 2016 2:19 pm**

@Gunny wrote:At some point I will be releasing a pack of what I call raw airports from the UK and Ireland.No ground object involved.

Posted: **Sun Apr 24, 2016 3:20 pm**

@u2fly This is my hobby not my job.And I take my own sweet time in doing it.I will release it when I am satisfied that it is ready for release.Besides I do have a real life to attend to and health issues to address that takes up most of my time.So when you see the post that is when it will be released.Thank you for your cooperation.

Posted: **Sun Apr 24, 2016 3:30 pm**

@Gunny. OK, I only ask.

We all there for fun, not for job

We all there for fun, not for job

Posted: **Mon Apr 25, 2016 4:33 am**

Hovileta Airport

Maggiore Airport

Del Rosario International Airport

*no image*

Kansho Air Base (Descent sized cargo ramp, home of some F-22s & F-35s, Two intersecting runways & a great place to fly around)

*no image*

Vermilion Airport (Grass Runway, Small Ramp, Made for Cessnas and those types of aircraft)

There will be two versions of this map, one with and one without ground objects surrounding Hovileta Airport for the airshow. I'm loving it so far, but it needs way more mountains and roads

Posted: **Mon Apr 25, 2016 4:51 am**

I know that plane!Kujo wrote:

Posted: **Mon Apr 25, 2016 7:22 am**

My current project of the day is sitting down with pen and paper and drawing some polygon simplification and rendering algorithmns.

Those of you who have followed my work in the past may remember "Flake Engine" my little c++ spin off designed to render a SRF file in Direct3D.

The problem with YSF SRF files is that polygons can be much larger than 3 verticies that OpenGL/Direct3D etc. draw in.

So how to simplify these polygons and draw correctly?

"Just trace from vertex 0, to 1, to 2... then do 2 to 3 to 0..." This is the FAN technique, which works for an ordered list.

Problem is what if you get a complex polygon where the edges intercept each other and draw holes?

So now it gets much harder to render a polygon.

I have read of the "ear lobe" technique and also the "odds/even" technique. I've written my own solution, which is similar to Soji's YSClass's YSSword function as well.

The algorithm is simple actually: Simplify the edges, then cut off the lob ears until no polygons can be drawn.

I will now apply this algorithm to this complex polygon:

https://i.imgur.com/oYUMaYZ.png

===

**Step One: Simplify the edges.**

https://i.imgur.com/utVp7Pj.png

Start by checking if any edges intersect each other, as viewed from the polygons normal axis. All the intersections will need to have a vertex added.

https://i.imgur.com/yKu2XC4.png

The vertex is added and the polygon definition is updated to put the vertex in between the two affected edges. This will increase the polygons vertex count.

https://i.imgur.com/HbdiqIU.png

The edges are simplified until there are no more intersections.

**Step Two: Start Ear Lobing.**

https://i.imgur.com/bC6i3jz.png

Trace from the current vertex root (starts at 0) to it's next linked vertex (vertex 1), and then to the one after that (vertex 2). If any edge, especially the end vertex to the root intersects another edge, the polygon is no good (overflow) and iteration will restart from the next highest vertex root (vertex 1). Once a successful preliminary match is found, Find the center of the polygon that could be drawn by these 3 vertices and make that the polygon center for testing.

https://i.imgur.com/ksfgVPd.png

Draw an infinite ray from the calculated polygon center and count how many edges it intersects with. If the number of edges is ODD, it is INSIDE the final polygon, and will need to be drawn. If it is EVEN, then the polygon would be OUTSIDE the final polygon and should not be drawn. In this case, the root number of the vertices should be increased by one, and Step Two will re-iterate, testing and drawing new polygons based off the defined order.

**Step Three: Breakdown of Ear Lobing**

https://i.imgur.com/WuvRuxI.png

As each successful polygon is calculated, it is added to a completed list, and the polygon definition is updated.

https://i.imgur.com/OzTlxyc.png

Step Two is resumed on the new polygon definition, finding Ear Lobes and testing them to see if they are in the polygon and do not cross any other edges.

**Step Four: Repeat**

The process is repeated until no more edge conflicts are found, no more polygons can be created, or the working polygon definition (the one we keep breaking down as we break off ear lobes) is reduced to 3 vetecies or less.

https://i.imgur.com/RduynFT.png

The complex polygon is now drawn.

Some examples rendered with this method:

https://i.imgur.com/3LYMiUS.png

https://i.imgur.com/TKvEtfU.png

https://i.imgur.com/rWhIlTx.png

This method works for all polygons I can think of in my mind at the moment. Problem is that it is not very efficient.

This style of solving triangulation has existed for many years (at least 10?) now, but it's a slow approach. It'll get the polygons right every time, but it WILL take time to calculate. In fact the exact same lag problem exists in the real YSFlight executables, in YSClass (Soji explained this to me, he has a few methods, and he tries them all from fastest to slowest based on some binary tests).

So if I go back to my rendering engine code and re-write it, I should be able to FINALLY draw polygons completely correctly.

I will use this algorithmn to load SRF and eventually DNM models into my own programs when I finally get that far in coding.

Those of you who have followed my work in the past may remember "Flake Engine" my little c++ spin off designed to render a SRF file in Direct3D.

The problem with YSF SRF files is that polygons can be much larger than 3 verticies that OpenGL/Direct3D etc. draw in.

So how to simplify these polygons and draw correctly?

"Just trace from vertex 0, to 1, to 2... then do 2 to 3 to 0..." This is the FAN technique, which works for an ordered list.

Problem is what if you get a complex polygon where the edges intercept each other and draw holes?

So now it gets much harder to render a polygon.

I have read of the "ear lobe" technique and also the "odds/even" technique. I've written my own solution, which is similar to Soji's YSClass's YSSword function as well.

The algorithm is simple actually: Simplify the edges, then cut off the lob ears until no polygons can be drawn.

I will now apply this algorithm to this complex polygon:

https://i.imgur.com/oYUMaYZ.png

===

https://i.imgur.com/utVp7Pj.png

Start by checking if any edges intersect each other, as viewed from the polygons normal axis. All the intersections will need to have a vertex added.

https://i.imgur.com/yKu2XC4.png

The vertex is added and the polygon definition is updated to put the vertex in between the two affected edges. This will increase the polygons vertex count.

https://i.imgur.com/HbdiqIU.png

The edges are simplified until there are no more intersections.

https://i.imgur.com/bC6i3jz.png

Trace from the current vertex root (starts at 0) to it's next linked vertex (vertex 1), and then to the one after that (vertex 2). If any edge, especially the end vertex to the root intersects another edge, the polygon is no good (overflow) and iteration will restart from the next highest vertex root (vertex 1). Once a successful preliminary match is found, Find the center of the polygon that could be drawn by these 3 vertices and make that the polygon center for testing.

https://i.imgur.com/ksfgVPd.png

Draw an infinite ray from the calculated polygon center and count how many edges it intersects with. If the number of edges is ODD, it is INSIDE the final polygon, and will need to be drawn. If it is EVEN, then the polygon would be OUTSIDE the final polygon and should not be drawn. In this case, the root number of the vertices should be increased by one, and Step Two will re-iterate, testing and drawing new polygons based off the defined order.

https://i.imgur.com/WuvRuxI.png

As each successful polygon is calculated, it is added to a completed list, and the polygon definition is updated.

https://i.imgur.com/OzTlxyc.png

Step Two is resumed on the new polygon definition, finding Ear Lobes and testing them to see if they are in the polygon and do not cross any other edges.

The process is repeated until no more edge conflicts are found, no more polygons can be created, or the working polygon definition (the one we keep breaking down as we break off ear lobes) is reduced to 3 vetecies or less.

https://i.imgur.com/RduynFT.png

The complex polygon is now drawn.

Some examples rendered with this method:

https://i.imgur.com/3LYMiUS.png

https://i.imgur.com/TKvEtfU.png

https://i.imgur.com/rWhIlTx.png

This method works for all polygons I can think of in my mind at the moment. Problem is that it is not very efficient.

This style of solving triangulation has existed for many years (at least 10?) now, but it's a slow approach. It'll get the polygons right every time, but it WILL take time to calculate. In fact the exact same lag problem exists in the real YSFlight executables, in YSClass (Soji explained this to me, he has a few methods, and he tries them all from fastest to slowest based on some binary tests).

So if I go back to my rendering engine code and re-write it, I should be able to FINALLY draw polygons completely correctly.

I will use this algorithmn to load SRF and eventually DNM models into my own programs when I finally get that far in coding.

Posted: **Mon Apr 25, 2016 7:57 pm**

This looks amazing! Just be wary of overdoing mountains. Try to make sure they make sense and don't just rise up everywhere like I do. It gets kinda weird. Just remember Vuelo Beach and its 2200m mountains and flat ground everywhere else lolKujo wrote:There will be two versions of this map, one with and one without ground objects surrounding Hovileta Airport for the airshow. I'm loving it so far, but it needs way more mountains and roads

Can't wait to see what other airports you've made!