Predicting shapes of wake effects

[L_Rob]

Hi folks,

I suppose it's way past time for me to start posting on this topic in a more appropriate thread, rather than the general discussion.

Here's a screen clip of the current state of my spreadsheet intended to predict the overall shape of each of the various emitters in wake effects. I've added a few additional parameters other than the ones previously mentioned, like sternpost position that determines where the datum of the effect is located, and some indication of the width of each portion of the effect.

The overlay is scaled based on the known length and beam of one of the boats I'm using as a test case. LOA and beam are among the inputs. I think there are still some errors in the various functions but reading things like =(($C$83*EXP($C$84*G93))*$K$1)-$E$83 will drive you nuts.

 

[Dutcheeseblend]

That is a really nice plot! Sir, I was talking about complex relations, that was based upon an exponential equation. But I didn't realize there should be an angle as well, in the exponent.

Keep this going! This might be useful for other effects too.

 

[L_Rob]

That is a really nice plot! Sir, I was talking about complex relations, that was based upon an exponential equation. But I didn't realize there should be an angle as well, in the exponent.

Keep this going! This might be useful for other effects too.

Thanks!

In working on this I keep finding various oddities, like something I just discovered. One might guess that if you set the sternpost position to zero meters in the sim.cfg of an AI boat, the datum of the wake would be set at the stern of the vessel, probably interpreted as the aft end of the bounding box. But no, that is not the case. For a 150 meter vessel, it appears to be about 10 meters forward of that. I'm trying a percentage of LOA as an offset, but I suppose whether it's really an offset or not depends on one's definition of sternpost, a rather old-fashioned term to use in the first place. As I find and fix these things, the fundamental equation relating position and time seems to be increasingly predictive.

Speaking of angles, the Kelvin wake spreads out at a relatively constant 20 degrees although it can appear to fan out considerably depending on conditions. This is not well approximated by an exponential where you have to start with a scale that will not project beyond the beam and requires a scale rate. But your mention of angles got me thinking: perhaps one could add constant velocity along the x axis to the exponential governed by the scale rate to shape the curve to one's liking, however that would require splitting parts of the effect into port and starboard halves. I'll play around with it. If it works it's as simple as balancing two parameters.

 

[GHD]

The sternpost is where the rudder is attached.

 

[L_Rob]

Right George,

There are several definitions of sternpost; I was referring to my Oxford Companion to Ships and the Sea. The term originated as the aftmost timber in a vessel. Originally the rudder was hung off this post, thus the varying definitions. Maybe the term is still commonly used for powerboats; don't know. By old fashioned, I was thinking that it's not used in the ISAF Racing Rules of Sailing, or the Equipment Rules of Sailing, where standard terms are defined. At least the last time I was involved with the ERS it was not there! I'll admit it; I'm narrow minded.

 

[L_Rob]

George, not that it's particularly pertinent, but here's an amusing thought:

I'm sure you are right; the FSX coders equated sternpost with rudder post, and probably felt that was a good place for the wake effect to start. Of course that leaves open the question of where the rudder post is. Small outboards, large ships, powerboats, sailing vessels of various types all have the rudder or effective rudder in different places.

As an aside, you would not believe the major international disputes over such definitions at sailing world championships, the Olympics and the America's cup. Or maybe you would! Unlike a maneuvering violation, there is no way to exonerate yourself from an after the fact measurement protest (unless it's been added recently). You are OUT.

The amusing thing however is that per the SDK, it appears that the sternpost position/rudder post position only affects the wake. I commonly start with the sternpost position at half the LOA (with the model datum in the center of the boat) and if FSX referenced this as the position of even a hypothetical rudder, it would create no torque and the vessel would not turn!

 

[Anthony31]

I have posted some notes about wake effects in P3D V2 for those who may be interested:

http://www.prepar3d.com/forum-5/?mingleforumaction=viewtopic&t=6589.0

I've found that in P3D wake effects don't render as neatly as they do in FSX. More interesting is that in P3D the water is decided wibbly wobbly now compared to the dead flat water in FSX. This makes use of wake effects in P3D subject to some difficulties (ie the flat wake effects can end up disappearing under severe waves). If have to to resort to particle effects to create wave effects for my two floats equipped aircraft.

 

[L_Rob]

Anthony, I've noticed similar problems, comparable to the screenshots you posted. However the way wakes are treated in P3Dv2 appears inconsistent. Because the datum of the wake effect gets located at the datum of the model (as long as it's specified by wake= ) for models with the datum at the surface, wakes display intermittently, often flickering. For my larger ferries with datums at the LWL, as long as I don't use the Ultra water setting (3D waves) setting the static cg height to 1.3 works fairly well and the wakes look quite similar to the same one in FSX. But my latest small boat suffers from the problem you describe, and elevating the wake via elevating the boat some does not help much. My particle rates are set at 1.00 but as many as 4 seconds can go by before a new texture shows up. When it does show up, it has already grown to the size it should based on my scale rate setting. Same thing for extrusions. So suddenly a quite large tile pops up with a sharp leading edge, looking quite ugly. It's as if P3D is keeping track of what it's supposed to do, but showing the tiles only when if feels like it. Maybe it's related to scenery complexity, or model complexity, or maybe it's related to P3Dv2 offloading effects processing to the GPU. Whatever it is, it's odd that wakes on my large and quite complex models more or less work, but not on my new 20 foot long water taxi. Maybe I can fool P3D by setting the rate to 4 per second.


Edited to add: 4/sec didn't work, oh well. But another oddity-- when the boat alters course, the wake displays fairly well. It appears to restart at each incremental course change, and ends up looking better.

 

[Dutcheeseblend]

I'm currently struggling with the effects, and I don't really like the way they are interpolated. It appears very unnatural, such an exponential shape.