WIP Steam Gunboat

[Pieter Boelen]

I'll try it without that line altogether then and if that does something weird, just keep it in for now and fiddle with the speed modifier. I do like that idea of having an extra attribute; seems more logical as it gives us a baseline.

Ok so I'll just leave inertia alone for now then.

You might want to copy the numbers from a ship that is more similarly sized to your own.
The steam frigate is, of course, substantially larger.

 

[Robert Nutter]

Are there any steam ships which are smaller then? As I was under the impression Astrea's model was our only benchmark!

 

[Pieter Boelen]

There are not.
But there are plenty ships that are smaller.
You just need to copy the steam related part from the steam frigate.
Anything else you can take from any other ship.

 

[Hylie Pistof]

Ship weight works with other values for things like the ship rising up out of the water as cargo is taken out as well as how it turns as well as how fast or slow it accelerates/decelerates. Look to naval vessels for guidance there.

 

[Robert Nutter]

Thanks Hylie! I worked out she'd be around 250-400 imperial tons based upon the heavy beam engine machinery and robust construction of the historical Charlotte Dundas, so I just hope the engine agrees that an amount like that is good for it in terms of acceleration/decel etc

In-game she's about the size of the war tartane, though perhaps a little wider on the beam with a deeper draft.

 

[Pieter Boelen]

Ship weight works with other values for things like the ship rising up out of the water as cargo is taken out as well as how it turns as well as how fast or slow it accelerates/decelerates.

In combination with these lines you mean, right?

     refShip.SpeedDependWeight = 0.0;
     refShip.SubSeaDependWeight = 0.0;
     refShip.TurnDependWeight = 0.5; //KNB was 0.0 but all were then each set 0.8

In-game she's about the size of the war tartane, though perhaps a little wider on the beam with a deeper draft.

Have a look at the War Tartane stats then and see if yours sort-of fit into the same range.

 

[Armada]

I followed this stage exactly, but two sided lighting does nothing; in fact, checking it on only makes non-shadowed parts of the deck slightly brighter. I don't know if we have an issue here anymore though, as she looks fine in-game and all textures/faces are responding as expected under all the differing lighting conditions. I did also turn 'Hardware Textures' on, mostly on a whim, which then made a HUGE difference as I could of course see what i'd applied without the renderer constantly crashing (which it'd do with more than one texture applied). Turning it off again also eliminated the black...weird, huh? Could this also be why she works fine in-game despite the blackness?

I don't know... something still doesn't sound right. If you can upload the file you imported in the first place, I'll take a look and see if there are any issues.
Chances are it might be fine, but I'd like to be sure so I can update the tutorial if necessary.

After seeing her in-game it doesn't really bother me *that* much anymore, as most of the time you're seeing it from an angle anyway. But would you say it's something that *needs* fixing? No problem either way.

Up to you; it's your ship.

Is it bad that I was dreading doing this, even for such a simple model? I think because I can't get my head around how it stays put just by using one locator. So I simply use the same locator method I did with the whole model but on the mast, then export selection? It's adding the new one to the hull I get caught up on. If I add that in Maya, TOOL will see it and I'll be able to add it to the table? Then the mast will be anchored to that locator's origin and the engine will make sense of that? I couldn't find where Maya shows position info though...just rotation, scale and translate (which I tried entering co-ordinates into but definitely doesn't control position!!)

You can use the Channel Box or the Attribute Editor to show transform coordinates. I tend to prefer the Channel Box for quick reference and directly changing the values.

As for locators... technically, you can export them all straight from Maya and not have to make any changes in TOOL.
Let's say you choose to set up the flagpole as a mast. Create a locator called "mast1" and parent the mast model to that locator. The locator's position doesn't matter too much, so keep it to (0,0,0) if you like. Make sure you parent this locator to the main locator along with the rest of the model.

Now, select both locators in the scene (the main parent locator and the mast locator) and use Export Selection as usual. This will export two GM files and automatically create a Geometry locator on the hull for the mast. You can create more locators in Maya and export them with the model in the same way, but I haven't yet explained this in full in the tutorial.

 

[Hylie Pistof]

Yeah, that too. Find a similar size warship that is not one of the originals and then cut its cargo capacity in half. That would be a good start.

 

[Robert Nutter]

I don't know... something still doesn't sound right. If you can upload the file you imported in the first place, I'll take a look and see if there are any issues.
Chances are it might be fine, but I'd like to be sure so I can update the tutorial if necessary.

Attached is the .OBJ I exported from 3DS to make this work. I'm sure it's nothing to do with your tutorial and more to do with me using C4D. Some quirk in the system.

Up to you; it's your ship.

I'll probably fix it using Pieter's advice. Seems a shame not to.

You can use the Channel Box or the Attribute Editor to show transform coordinates. I tend to prefer the Channel Box for quick reference and directly changing the values.

As for locators... technically, you can export them all straight from Maya and not have to make any changes in TOOL.
Let's say you choose to set up the flagpole as a mast. Create a locator called "mast1" and parent the mast model to that locator. The locator's position doesn't matter too much, so keep it to (0,0,0) if you like. Make sure you parent this locator to the main locator along with the rest of the model.

Now, select both locators in the scene (the main parent locator and the mast locator) and use Export Selection as usual. This will export two GM files and automatically create a Geometry locator on the hull for the mast. You can create more locators in Maya and export them with the model in the same way, but I haven't yet explained this in full in the tutorial.

I didn't realise it worked like that straight from Maya, was under the impression it had to go through TOOL or a worksheet to work. I feel a little better about doing it now! Does that work with the gun locator?

 

[Pieter Boelen]

I didn't realise it worked like that straight from Maya, was under the impression it had to go through TOOL or a worksheet to work. I feel a little better about doing it now! Does that work with the gun locator?

Let's put it this way.... the original game developers didn't HAVE the TOOL, because that was made by someone here at PA!
They did everything through Maya.

 

[Robert Nutter]

So TOOL is more for tinkering with existing models than setting them up in the first place then? As from what Armada says, it does sound more straightforward doing it from Maya (famous last words). Though I don't pretend to understand how I'd go about doing gun locators in it!

 

[Robert Nutter]

While we're on the subject of weight, I wouldn't mind some ideas on how to get around what it's doing when I sail against the wind using the default engine speed:

DROPBOX LINK: VIDEO

Dropbox - ENGINE 2016-02-15 23-58-04-94.mp4

Of course It's because the engine wasn't designed to handle such small ships going against the wind so fast, that and the fact historically the experimental Charlotte Dundas would probably have done exactly this on the open sea due to her size, hence never leaving canals and rivers. So I don't think it's a modelling fault on my half or possibly even the engine doing anything wrong; just history proving itself in the game . But that's one hell of a nose dive she's doing into those waves; lord help me how it'll behave once I add a load more speed on, as I'm planning on doing. I'm a bit concerned it'll break player immersion if they see this bumping about in the background :/ Is there a way perhaps I can make her heavier (or even lighter) to 'trick' the engine into thinking she can 'surf' a bit more? Or does it not work like that? The only other thing I can think of is to make her sit higher in the water, but I'd rather not as she's perfect how she is and it'd look wrong the rest of the time. Last resort option is to make her bigger physically, which isn't great for obvious reasons.

Could this have anything to do with center of gravity being affected by the centrally placed locator? As she does behave a bit like a speedboat at the moment...

In other news, it's great fun running away from Bermuda sloops and Xebecs with her

 

[Pieter Boelen]

Cannot watch the video right now, but what speed do you have her at?

The location of the (0,0,0) punt could be related.
Where do you have that now?

 

[Robert Nutter]

The default speed, so whatever Astrea's engine is set to :/

That's still at 0,0,0.

 

[Pieter Boelen]

The default speed, so whatever Astrea's engine is set to :/

She's set to 15.0 knots, right? Might be a bit much for such a small engine-propelled ship.
But anyway, with the engine she should currently do only 10% of that.
So does that mean she was really doing only 1.5 knots and moving like mad?

That's still at 0,0,0.

But where IS (0,0,0)? That might make a difference.
You can show that using GM Viewer with the "3D grid" option.

 

[Robert Nutter]

To be honest even sitting still she dunks her nose under the waves...it's just what happens with this shape of hull and the size of the swell in POTC, I think. I'm just concerned that if at 1.5 kn she's this bad, imagine what 13 or 14 will be like...

Will check that out! As that could potentially be an issue. Though there is still the issue that she's buffing the waves with one hell of a force.

 

[Pieter Boelen]

To be honest even sitting still she dunks her nose under the waves...it's just what happens with this shape of hull and the size of the swell in POTC, I think. I'm just concerned that if at 1.5 kn she's this bad, imagine what 13 or 14 will be like...

The game doesn't know anything about the shape of the hull.

Is she really similar in size to the War Tartane? Does the War Tartane experience the same effect?
If the War Tartane does not, then try to replicate the stats and (0,0,0) point of that one; see if that improves things.

 

[Hylie Pistof]

Overall the ships seem to rock a little late in the waves causing their bow and stern to dip lower than they should. This can be seen when some ships, especially frigates, sink so low in the stern that their windows are under water.

Then there is the size of the hull. Medium size hulls ride the swells pretty well while the smaller ships can have a rough time.

In WWII the US Navy learned a hard lesson. During a typhoon the smaller destroyers would ride up and down the waves and survived ok. The battleships and aircraft carriers rode three waves and survived ok. But the cruisers rode two waves and the constant stress from being supported at the ends and then in the middle broke them in half.

 

[Robert Nutter]

Ok, so this probably doesn't help:



Don't get why it's wrong though, as I set it exactly central in Maya Ah well.

@Hylie Pistof I can see why that would've been a problem. They'd get high centered constantly, so couldn't ride it out. I wonder if any wooden ships ever suffered that fate...I'm sure on countless occasions! Either way, even with the locator centered I think I'll still have something of an issue with this; probably with little or no way completely fixing it

 

[Pieter Boelen]

Ok, so this probably doesn't help:

Actually, yes it does! As you can see there, the (0,0,0) point is quite far near the stern. I think that explains her weird pitching behaviour.
You could try to shift the entire ship model until the (0,0,0) point is close to the actual middle of the model.
See here:

It may not completely solve the problem, but I reckon that should at least improve things.