Water

[Destined]

Hi all,

 

I tried to add a floating boat to my WIP, but as soon as I tested, how it looked in-game, it immediately sank. After checking the properties, I noticed, that the boat has a density of 0.5, while the water has a density of 0.0015. I checked the def files and found this comment:

 

 

// density of water is 998kg/m³ at 20° C, however, our moveables usually
// have realistic masses coupled to a collision model that is slightly
// too big (since they are very inaccurate), so we drop the density slightly
// to make them properly sink (the default is 0.001043):
"density" "0.0015"

 

Frankly, I don't understand this number for the density. If the density is around 1000 kg/m^3, why is it smaller by a factor of 1000000 compared to kg/m^3 and even a factor 1000 compared to 1 kg/L in TDM? The other point I wanted to get at is: It should be stated somewhere which units are used to describe the density for players that want to work with buoyancy. I suppose the creator of the boat thought something like "Water has a density of 1 kg/L, so lets make the boat 0.5, so it floats." and made it far too dense by accident. The Wiki describes how to create water and visibility etc, but does not describe density anywhere...

I will check the density values of other entities (and materials) when I come home this evening and will check if there are other entities that should float, but won't, because a too high value. Maybe I can then give a list, where this should be changed (if there actually are more entities that use density).

Edited March 11, 2016 by Destined

[Destined]

Ok, I did some testing and found out, that all moveables have a density spawnarg of 0.5 and if so they float on water (at least the crates, the barrel and the plank I found). I increased the value for one crate: 0.8: it still floats; 1 it sinks. This seems to be the threshold (for crates, at least). The boat also has a density of 0.5, but does not float. I thought, maybe it has to do with its mass (500), but reducing that to 50 also had no effect. What I still don't get: When the density of the liquid is 0.0015, why do objects wiht a density of 0.5 float? Maybe I am missing something here. Can anyone help? I would be fine with the boat floating on the water, actually.

[Moonbo]

Why not just use a func_bobbing or func_mover?

[Destined]

Yes, this would be a good workaround, but I actually want to understand what is going wrong here...

[Moonbo]

In general Doom 3's physics just arent as complete or integrated into gameplay as something like Half Life 2. There are sadly a lot of things that just don't work outside of the most basic simulations :/.

[demagogue]

Revamped physics was 4th on my pipedream list, after stand-alone, softshadows, and multiplayer. Well we got one of them and another in the pipe, so anything's possible.

 

I'm guessing density for objects and water have to be different units to explain something like that. And on top of that no doubt id4 has some shortcut method. The actual physics of floating is technically the weight of water displaced relative to the object's weight, which density relates to depending on what it's actually measuring.

[Destined]

As stated in my first quote, the density of moveables is apparently calculated by dividing the defined mass by the volume of the collision model. I think I read that the density spawnarg is supposed to override this calculation. But it might also be, that this is a factor by which the calculated density is multiplied. This would at least explain, why the crate sank with density = 1. The question would be: Which mass is defined for func_liquids and is the mass actually depending on the brush size used? Getting behind all this would require more testing than I am willing to do.

 

The actual physics do definitely not apply, because a floating object stays floating even if another object with greater mass (i.e. the player) jumps on them. During my testing on Fryday I jumped on a small plank that did not sink, when I jumped on it. So I believe, the engine just compares densities and from them defines, whether an objects floats or doesn't. I still have not found an explanation for the sinking boat though... Maybe I will really just take Moonbo's advice and create a func_bobbing. Still it bugs me

[SteveL]

At a guess, I'd say that the tiny density numbers are for doom units cubed, not m^3.

 

There are competing mechanisms here -- spawnargs overriding very rough calculations. I put in a fix for the mass of certain moveables in 2.04, and we'd better make sure that it doesn't make everything float in water during beta!

[Destined]

At a guess, I'd say that the tiny density numbers are for doom units cubed, not m^3.

 

That's what I also thought. But for some reason, objects with higher density (0.5) float on these liquids. And these should have the same units.

[AluminumHaste]

Yeah the physics engine isn't that robust. I do salute you for trying though.

[Destined]

Just today I found out that the boat works perfectly fine. The problem seems to be another one. I placed the boat around the height of the water surface but on map start it is not there but on the bottom of my water entity, which is why I assumed that it sank. However, when I dived down there and touched the boat it resurfaced and floated on the water. So the problem is not theat the boat does not float but that the boat is put on the bottom (either during dmap or at map start) and seems to get stuck there somehow. Maybe it is possible to prevent this "getting stuck" and finally have the boat float as it should...

[Dragofer]

Interesting - maybe it should start in the air and drop down into the water?

[Destined]

I already tried that. My suspicion is, that during dmap or map start TDM checks for moveable objects, that are in mid-air and moves them along the z-axis until they touch static object. Apparently the water is not recognized as solid and thus the boat is moved to the bottom. The only thing I don't understand is why it gets stuck. I would have expected that as soon as the boat is registered as a movable object in the water the physics engine would kick in. Maybe I should try the other way around and put it on the bottom to see if it will resurface. I will try that this evening.

[Oldjim]

If all else fails would it be possible to put a support under the boat and then after the map starts remove it

[grayman]

I assume the boat is a moveable. Place it where you want it, then give it this spawnarg:

 

nodrop 1

 

That keeps moveables from settling on the floor beneath them at spawn time.

[Destined]

Yep, nodrop did the trick. The boat is not floating initially and still has to be touched by the player for the physics engine to kick in, but this could simply be done by placing the map start on top of it. Thanks a lot!

Ironically, I am currently thinking about sinking the boat, as it makes more sense for the story. The boat was used to get to the target location and should raise suspicion as soon as a guard sees it. The exit is planned from the start to be a different one, so the thief would be better off, if he had sunk the boat. Will think bout it...