How to Set Up Dynamic Water Physics and Boat Movement in Unity | Ship Buoyancy Tutorial 

Hello! Great video! Can you /anyone help me? My untiy cube is "delaying", mean Its working weirdly.

9.5 / 10

I love how its just above 10 minutes

Hi Tom my Flotter dosn't follow wave movement his y position don't go more then 0.5 and 2.4...., so bettween wave he do levitation otherwise he is submerged like some heightlimiter were set

the probleme was i scaled the plane up.

It's not the _wrong_ way-it's just a starting point that doesn't scale super well, but the theory behind it works and applies to the shader approach as well 🤷‍♂️

Well, the main concept in learning game development is to learn it through trial and error so it was indeed a wise choice since we can try and make it even better, thanks for the great tutorial man, @tomweiland

Thank you :D

Glad you liked it that much :D

Glad you liked it :D

@@tomweiland just for curiosity: have you considered creating the floaters dynamically in the script as points in space instead of manually adding them as gameobjects? Code would be a little more complex but it would be simpler to create new floating objects. You could just find the corners of the gameobject programmatically and add 4 floaters to it, or even do it in a fancier way.

Thank you :D

Thanks, I'm glad you liked it :)

This is awesome to hear, I'm glad you found it helpful! Good luck with the voting phase, assuming that hasn't concluded yet :)

Thanks, I really appreciate that!

Good point-a few people have run into that issue.

Yep this was a problem I ran into just recently as well, my floater was bouncing up and down well below the waves but that fixed it.

Awesome, ty for this comment so much

im having a problem where at this point when I go to play my cube just falls and I get an error saying "NullReferenceException: Object reference not set to an instance of an object Floater.FixedUpdate () (at Assets/Scripts/Water/Floater.cs:14)" which is on this exact line, I've copied everything correctly as far as I can see so I'm not sure whats wrong with it. Any ideas?

Double check that the RigidBody is filled in the editor. For some reason the reference removes itself.@@ethancotton1549

Glad you like it :D

Glad I could help!

Maybe, although you'd need to make sure the floaters know which water instance they're actually floating on. Also, I'm glad you like the speed! I myself hate slow tutorials-there's just something extraordinarily excruciating about sitting through any sort of extra fluff when I just want to know what to do next, but skipping ahead is dangerous because you risk missing something :P

I'm glad you liked it! I don't think I'll cover the shader side of things, mainly because that Catlike Coding article I linked to in the description covers it all. I'd basically just be recreating the article in video form.

@@tomweiland Ok, that's fine! Would it be hard for me to convert the shaders on catlike coding to shader graph? (As HDRP doesn't support custom shaders other than shadergraph) Sorry - forgot to mention that

That's strange...I haven't used HDRP but in URP you can write custom shaders 🤔 I believe shader graph has some kind of vertex manipulation node, so it's really just about plugging in the formula you want to use to calculate the height-it shouldn't be too difficult to convert.

@@tomweiland Ok, thanks!

Glad I could help :)

я тоже хочу, научиться и сделать, что то кроме синусоидальных волн, а именно как в lowpolywater. но не умею, хотелось бы знать как!

Not anymore, I got really sick of it 😂

I love the low poly look as well. Sometimes I feel like I could just gaze at it for hours :P

Glad I could help :)

hehe typical mr.krabs practices :P

You can try :P

Where is your game ?

@@fernandocabal5798 lol

Thanks, glad you liked it :D

Are you talking about water at different heights in the world, like rivers and lakes? If so, you're right that (as is in this tutorial) this solution only works with water that's at the same height. If that's not what you meant, then perhaps using some punctuation would help clarify what you're trying to say 😅

Glad you like it!

Glad you liked the videos-welcome to the community :)

How this game coming along?? I only ask because it sounds a lot like what I’m trying to create.

I'm glad you liked the pace-my first tutorials were definitely much too fast, but just like you, I hate it when a tutorial is really slow and you have to skip ahead :P

Probably not :P There's not _too_ much more I can cover here that isn't already explained in that article I linked in the description. If there was a _huge_ amount of demand, then maybe, but making tutorials isn't exactly the most exciting work...

@@tomweiland I vote for yes, also you have a new suscriber as this is the thing I was literally looking for

@@bustosmusic thanks for subscribing, I'm glad I could help :) I'm assuming you're voting "yes" for turning this into a series, right? What else would you want to see covered (not making any promises, just not really sure what else would need to be explained)?

Glad you like it 😅

Glad you liked it!

Glad you like my content :D

@@tomweiland yea

Thanks, I'm glad you like it :)

Tom Weiland good video man

Right but then you need to set the mass accurately, which then means you have to do a bunch of calculations based on volume, material type, etc. and it just becomes a pain. It's much simpler & more convenient to just have a few multipliers that you can play around with to get something that looks/feels right than something that is completely accurate (which it really never will be because the force would be applied all over the object, not just at one or a few points).

Yes, you are probably right, for a casual gameplay not heavily relied on physics this is a simpler solution.

Glad you like it :)

Thank you :D

Have you tested this? I'll give it a shot, although that doesn't seem right since buoyancy is an acceleration isn't it? Also do you mean to subtract the specific floater's velocity or just he ship's?

@@tomweiland I mean subtract the rigidbody’s velocity at the floater’s position. I haven’t tested it in a buoyancy setup. But your rigidbody already has a velocity, since you’re adding one on top of that to reach your desired position, your ship will overshoot it resulting in a bouncing effect. If not, sorry for wasting your time. Would love to hear the result though. I’m not near a computer so I can’t test it right now.

I'll give it a shot, although on a theoretical level I'm not convinced :P

Glad it was helpful!

Glad it was helpful :)

Thank you :) Thinking about it now, you're probably right. I could have definitely pulled out the drag and buoyancy forces into separate methods-it just didn't feel super complex as I was writing it (especially since it's less than 10 lines). "Big hunting" ?

@@tomweiland lol bug hunting. I love phone swipe keyboards!

Ah, that makes a lot more sense!

@@tomweiland I find that using the right words really helps in comprehension! Didn't bother teaching me THAT in school!

Glad you like it :)

Thanks :D

Glad to hear it :)

Glad to hear it :D

Glad you liked it, and welcome to the community :)

That's an interesting application for this kind of setup-I guess you'd probably raycast downwards at each floater instead of sampling a wave equation, and then apply an upwards force based on how far away the ground is 🤔

@@tomweiland Exactly ;)

@@tomweiland So, I've been messing around with things, if you'd like to see what I came up with, check it out: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-aUgbtHNIo1k.html It needs a lot of work on my end, but your snippet here was essential in getting things going. I moved away from having 4 separate empties with a "hover" script, and instead moved all the logic to the parent, using a foreach on an array of "hover point" empties. I'd like to expand on this later by having the "thruster" destructible, and when destroyed fire off an event, which will then update the list of hover points. Anyway, the prototyping has been fun so far, you've definitely helped spark a new fire under my ass to work on gamedev stuff, and I think I'm gonna have a lot of fun with this.

Thank you :P

@@tomweiland hey man, i know youve got a lot going on. but if you can, could you explain how you adjust the wave manager formula to match whats in the shader? ive tried picking apart the shaders equation and reassembling it in normal script but the variables just dont work.

Glad you liked it :)

🤔

Yep helped me as well. In the video, the water drag is multiplied with fixedDelta time but not the gravity calculations

Good luck :)

I don't know what to do with this information 😳

Glad you liked it :D

@@tomweiland the floater script is working on my cube bit not my model ship with a rigid body any ideas

Glad you liked it :)

Thanks, I'm glad you've found it useful :)

I'm glad you think so!

Glad it was helpful!

Glad I could help!

Good luck :)

Thanks :D

@@tomweiland We have to thank you 🤗.

For sure! I'll be doing that at some point in my own project.

Thank you! To get the wave manager to produce accurate movement for your shader, you just need to make sure that both the wave manager and the shader are using the same formula and offset to calculate the waves :)

@@tomweiland haha yeah that's the tricky part, because shader graph is a visual thing it's hard to work out exactly which combination of nodes and in which order will produce the same movement algorithm. But I'll keep at it, this tutorial has definitely taken me one step further towards achieving it. Thanks again!

@@brandon-leecoskey5942 ah right, I always forget people use shader graph-I still write all my shaders with code :P

Thanks :)

Thanks :D

I'm glad you liked it :D

@@tomweiland Best part of it is that I can load heavy cargo as rigidbodies onto the ship and the ship sinks into the water and tilts according to the weight. Looks really cool :)

I'll be honest, I haven't really dealt with the devops stuff (containers, kubernetes, etc.) so far, so my knowledge in that area is pretty surface level, but what you said sounds okay as far as I can tell 😅

@@tomweiland Nice! I've been trying to implement CI/CD with github Actions using game.ci workflows, but they don't yet support server build just the client. They have a container image with unity that I can use to build the server and upload to my digital ocean droplet, but it's a 9Gb sized image and there's no chance I have to pull it on every commit. I'm trying to find whether is there a chance to cache it for the subsequent commits. I sincerely appretiate the work you've been done so far. Please don't stop posting contents. I loved your implementation of buyoancy. I was trying to do that for a long time and now I think I'll be able to. Thank you so much!

Did you read the article that I linked to in the description? It goes through the process of creating waves in a shader step by step. At the moment I have no plans to make a tutorial on that, but I'm glad you liked the networking videos :)

Thanks :)

Glad you liked it!

You're welcome :)

The displacementAmount is meant to approximate the total amount of water displaced if the object is entirely submerged. We use the displacementMultiplier because the object isn't always entirely submerged. For the drag, you can use the displacement "factor" instead if you like. Honestly this is just an example of how you _could_ mimic buoyancy, so if you think something else might work better, just try it out.

You just have to recalculate the buoyancy the same way the mesh is deformed. This video has a good explanation of how to do it (ignore the GPU readback stuff, what's important is that sampling the wave equation multiple times and adjusting the sample point on each iteration gets you a pretty accurate result): ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-kGEqaX4Y4bQ.html

@@tomweiland Thank you for that very quick and great answer even thpugh this is an older video !:))

I'm glad it's working for you! I was going to suggest increasing your mass (my boat's mass is 500, though it's still somewhat bouncy), but nice work figuring that out yourself :)

It may just be that you need to play around with the floater values more to make it more buoyant. If that doesn't work, join the Discord server and ask about it there :)

I specifically linked the Catlike Coding article to _save_ myself from that, and now here you are, asking for it 😂

@@tomweiland im aware of site for years but examples only go so far

@@wesleycoomans1433 "examples only go so far" but you want me to provide an example...?

Most formulas you might use to generate waves will repeat at some point (an untransformed sine wave has a period of 2π, which is how often it repeats), so you could use some math to figure out the period of your wave equation and reset the offset every time you reach that value. That should result in a seamless transition (assuming you do the math right 😆).

@@tomweiland yea noticed that too lol to be smooth it has to reset at the right time thanks for the response. very cool system

I don't plan on making a tutorial about shaders. I really don't see how it would be different if it's in video form, and I don't want to spend time turning a perfectly good article into a video. If you're struggling to understand the article (although you really just have to copy + paste, and I believe you can download the project files as well), then you should learn some shader fundamentals first, and there's tons of tutorials out there for that.

Glad you liked it :)

You'll probably want to use raycasts or trigger colliders to ensure the player is in the correct area, and then just check if they press the interact button.

@@tomweiland thx so much

Glad to hear that!

The video has been out for one minute-how is that enough time to make a verdict 😂

@@tomweiland it was just what was i looking for xD

@@kouni.i that's awesome to hear! Did you find it in search or was it suggested to you?

@@tomweiland Notification popped out, because im subbed to you since Dani's video

@@kouni.i sounds like a happy coincidence then!

😅

@@tomweiland I really enjoyed the tutorial tho lol... Keep on dev logging too! You're doing great!

Just because you simulate the waves in a shader instead of on the CPU doesn't mean you need to change the formula, but there are much better formulas to generate waves than using a Sine function (look into Gerstner waves). I just happened to mention the two at the same time, but you can do one without the other.

@@tomweiland Ty for replying. Gonna Try your tutorial with GPU shader from the asset store. No nothing about shaders and wave formulas:( . Hope it works. Even if it isnt 100% accurate. Need just the feel.

It may look that way if you've never done any shader programming, but it _is_ a step by step guide so you don't really need prior experience.

You just need to make sure that you're using the same formula and offset in the shader as you are in the height calculation method and it'll give you the same value.

I left plenty of time to pause between steps, which in my experience is what you end up doing anyways when watching a tutorial. This way you at least get to see the next step in full quicker instead of having to skip ahead because it's SOOO slow :P

I did the same thing. 50%

You just calculate the height on the CPU (aka via C#) using the same formula.

I only simulate the physics for the exact point that I need it. You can use the same formula (or an approximation if you're using Gerstner waves) on the CPU to get the height of the water at any give point, regardless of whether you actually have a water plane there or not.

@@tomweiland Thanks for the reply! I feel I've been able to get things working well with your advice, but I had another question regarding how you organize large areas of water and coordinate the waves across planes? I am completely lost on how to approach this

I did it that way to address situations where one side of an object is submerged while the other side is not, but now that I think about it you may be right about it being unnecessary-I wonder if not "splitting" the gravity among the floaters would make any noticeable difference on how it looks 🤔

Thank you, I will!

Awesome :D

1. The low poly look is achieved by having the entire surface of each triangle use the same normal. If a triangle shares vertices with neighbouring triangles, the normals of those vertices will be an average of all the surrounding face normals (I hope that makes sense). My low poly shader is sort of my "secret sauce", but I'll point you in the right direction-in order to have a universal face normal, each of the triangle's vertices need to have the same normal value, and what's most commonly done is to simply split vertices, meaning each triangle has its own unique set of vertices (since you can only have one normal per vertex). Alternatively, you can ignore the normals stored in the mesh data and recalculate them in a shader. This can be done using geometry shaders (which allow you to perform calculations on a per-triangle basis rather than per-vertex/per-pixel), or you can use the screen space derivatives in the pixel/fragment shader function. Over the course of several months, I probably read every article/question/forum post out there relating to low poly rendering trying to figure this out :P 2. I set the offset for both the WaveHeight method and the shader manually.

​@@tomweiland I managed to hack something together with a flat-cell shader I found by incorporating the vertex displacements... Much studying to be done on my part to understand it beyond that, though. I'll try not to make something too similar and step on your creative toes!

@@richardstevenson371 that sounds like an approach that I'm not sure I've heard of. If you don't mind sharing I'd love to know how it works-maybe we can talk on Discord?

@@tomweiland Whoops! I wouldn't mind, but I can't really offer any expertise myself. I simply implemented the wave functions into a shader I found online: hextantstudios.com/unity-flat-low-poly-shader/

:)

Thanks :)

👍

You can use that article I linked to in the description or you can follow along with one of the many video tutorials that are out there.

I half fixed it. Adding a collider seems to have fixed the problem, for some reason, but the ship is still floating really weirdly. It looks like the gravity aspect of the floaters arent quite working as expected there (The bow and stern remain roughly aligned with each-other, even when they really shouldn't be), so the ships gundeck, and sometimes even the upperworks, regularly get submerged. This isnt a gamebreaking issue, unlike the one that the collider solved, but it is annoying and quite distracting. This is also a problem on the cube i used to debug the prevous issue (I had to scale it significantly for it to actually be noticeable though). I personally suspect I must have gotten something wrong in the floater script somewhere but I wonder if anyone else had this issue?

What did you attach a collider to? The water plane shouldn't have a collider attached. How many floaters are you using? If your boat is substantially bigger than that test cube and you're using the same amount of floaters, that's probably the problem. The water height is only check at the position of each floater, so for example if you only have a floater at each corner of the cube/boat, a higher water height halfway between those floaters (aka around the middle of the boat) won't in any way affect the boat's movement. Based on what you're describing, it sounds like you probably need to add more floaters at different positions.

​@@tomweiland Thanks for the response! The water has no collider - the mother gameobject of the ship (which also has the rigidbody) has a box collider. beneath this object are the floaters and the ships models. I had 7 floaters, which i thought was an alright amount. I now have 14 and the issue persists. The test cube also seems to have this problem, so I dont think floater counts are the issue. I did some testing I also tried giving the floaters colliders - no effect - and rigidbodies - completely broke everything - I finally reenables gravity on the Rigidbody - I dont think that helped but it may be an avenue of investigation. I did realise something, though, my ship tends to list (lean) towards random sides a bit arbitrarily. I didnt think much of this at first since, well, I expected that behaviour, but I did just realise - it might be gravity that is doing that, and if so, then gravity seems to work on the sides, its just the front and back that remain roughly stationary. If this is indeed the case, then could it be a transform issue? Here is my floater script, if it helps. public class Floater : MonoBehaviour { public Rigidbody rigidBody; public float depthBeforeSubmerged = 1f; public float displacementAmount = 3f; public int floaterCount; public float waterDrag = 0.99f; public float waterAngularDrag = 0.5f; private void FixedUpdate() { rigidBody.AddForceAtPosition(Physics.gravity / floaterCount, transform.position, ForceMode.Acceleration); float waveHeight = WaveManager.instance.GetWaveHeight(transform.position.x); if (transform.position.y < waveHeight) { float displacementMultiplier = Mathf.Clamp01((waveHeight - transform.position.y) / depthBeforeSubmerged) * displacementAmount; rigidBody.AddForceAtPosition(new Vector3(0f, Mathf.Abs(Physics.gravity.y) * displacementMultiplier, 0f),transform.position, ForceMode.Acceleration); rigidBody.AddForce(displacementMultiplier * -rigidBody.velocity * waterDrag * Time.fixedDeltaTime, ForceMode.VelocityChange); rigidBody.AddTorque(displacementMultiplier * -rigidBody.angularVelocity * waterAngularDrag * Time.fixedDeltaTime, ForceMode.VelocityChange); } } }

Does your water plane have a scale of 1? If not, that'll cause some major problems

@@tomweiland the water plane does have a scale of 1. At this point I'm considering scrapping the gravity part of the code and figuring out my own solution to create the same general effect, since I'm completely lost as too what I've done wrong. (I do have ideas on how I could do that)

Yes, I'm using URP. Shader Graph is the only way to automatically use URP's lighting (that I'm aware of), and I think it's intended as the alternative to surface shaders. Since Shader Graph still doesn't allow you to modify certain things that I need to modify, I'm basically just piggy-backing off of URP's shaders. I copy + pasted the code files for the lit shader and then added my extra code, but the drawback is that any time Unity updates or changes something about the lighting calculations, I need to manually update my shaders with that same code for it to take effect.

@@tomweiland thanks for the info as always. I just created a new lit shader graph and copied its generated code, seems to work fine!

You're welcome!

I started learning when I was 12 (I'm 19 now), so about 7 years, although up until 2-3 years ago it was all super basic stuff.

@@tomweiland i really respect people like you that teaches us new programmers new stuff everyday, keep up the good work

You get the height of the waves from the WaveManager script (meaning it _is_ still needed) regardless of whether or not you manipulate the mesh in a shader or on the CPU. You just need to make sure that the shader uses the same formula as the WaveManager script.

This gets pretty tricky with a bunch of stacked Gerstner waves. Still debugging : )

@@richardstevenson371 if you make a method that returns the height for a single Gerstner wave, then all you have to do is call it multiple times and sum the results.

@@tomweiland I've been trying this but the results arn't what I'm expecting. Does something need to change with the offset or speed variable from the more basic waves?

Thank you, glad I could help :)