01 - Writing a physics simulation in 10 minutes

[@GoogleUser-ee8ro]

I can only imagine all the kids on youtube who are learning physics and programming are going to be crazy about this channel channel.

[@mihirmanna1488]

As someone who's been interested in coding physics simulations but never really known where to start, this is truly amazing. Everything was explained so well and I can't wait to keep going! Probably the fastest I've ever subscribed to a channel haha

[@austin-maddison]

This has been such an invaluable resource for me as a student and an aspiring graphics engineer. Thank you so much!

[@bvds2007]

This is a great resource. It is great that you are using straightforward javascript without all sorts of framework-related code which detracts from the main purpose of your code. Thank you.

[@adlbroun]

This is fantastic. When I was first getting into programming over 20 years ago I remember implementing a bouncing ball simulation like this in Excel VBA (best dev environment available on the computers where I was working!). I was blown away when implementing the equations I'd learnt in high school maths that I got such cool results for this and then later a toy planetary motion simulator. I think this is a great way of teaching people how to implement physics simulators, and also to get them interested in programming in general. Looking forward to following along with the rest of the series. :-)

[@sean_vikoren]

Thank you for your generosity!

[@TenMinutePhysics]

My pleasure!

[@voxelltech]

This is awesome! Thank you for all your work on the research! I am really excited to follow along this series!

[@franciscotrigo9656]

I'm surprised how little overhead requires these simulations using HTML and Javascrip. Thank you, Matthias, for opening my eyes!

[@stevenspmd]

Wow, I went back to university to retake some math classes because I'm interested in this stuff! Super cool!

[@cesarmello2006]

Congrats for the great work! Thank you so much for creating this channel.

[@thebirdhasbeencharged]

Just stumbled on your channel, fantastic stuff mate. Thank you.

[@tiago08ci]

So cool! Thanks a lot, you gained a subscriber who recently finished a masters degree in numerical analysis! Cheers!

[@nevil6147]

Mega cool dass dr müeh geh hesch mit video + demo code + sogar no pdf presentation. Echt vorbildliche approach für lehrinhalt

[@nyscersul42]

You've done very well in being clear. :)

[@ekzedex]

Thank you. The video is very helpful!

[@natteraksadawan4772]

Thanks, This is what I am looking for!!!

[@LogicEu]

Hey! Thank you for doing this! It's incredibly educative

[@guilhermejofili]

Wow! Great. Thanks very much. Lots of fun and learning. ❤

[@masterblaster3483]

Hi Matthias, I've been following you for a long time and as always I'm impressed with the quality and distinction of your work. Your openness in sharing your knowledge is admirable. If I may ask, at 10:51 you said that using higher order integration methods would not yeld notable improvements "when collisions occur". Could you please expand on this particular point? Are you suggesting that symplectic Euler is superior to leapfrog integration (2nd order) in case of particle-particle collisions? Is there a real problem in using other integration schemes (e.g. Runge-Kutta et similia) when dealing with collisions apart from the fact that acceleration would never be constant during dt? Thanks in advance for your response.

[@matthimf]

Hi, nice to hear that you like my work. Higher order schemes extrapolate the trajectories into the future. However, extrapolations become invalid when collisions occur.

[@masterblaster3483]

Hi Matthias, thank you for your prompt reply. In the 2018 work "An adaptive timestepping algorithm for particle time integration in coupled CFD-DEM simulations", Hariswaran Sitaraman and Ray Grout, a Velocity Verlet second order integration scheme is used in conjunction with a variable time step approach for local domains. This puzzles me, because the answer you have given me makes perfect logical sense. Nonetheless Verlet integration schemes are often used to integrate Newton's equations of motion. From Wikipedia: "The Verlet integrator provides good numerical stability, as well as other properties that are important in physical systems such as time reversibility and preservation of the symplectic form on phase space, at no significant additional computational cost over the simple Euler method." Your answer, which I consider of utmost relevance, seems to suggest that higher order integration schemes only apply in case of hard-sphere contact model and methodology, typical of molecular dynamics or the Discrete Element Method. Could you please confirm or negate this latter assumption? Best regards, Stefano.

[@444haluk]

My current idol

[@jmauriciou]

This is amazing content

[@sinam1951]

Excellent job! Thank you for making this chanel

[@MsJeffreyF]

I think because of the way we are doing reflection when bouncing off the ground we are losing velocity and the ball will get lower and lower. I think it's because of this because when it hits the ground, it will be less than 0 by a variable amount, but we aren't taking that into account in the position

[@jayanth9661]

Hey , I really liked your channel! Could you please do video on fluid simulation using SPH ?

[@djaccount5458]

Super!

[@JoshyLats]

Thanks!

[@sivabalan_m]

should the delta time dt be equal to the frame interval of the application? But here you are manually giving a value to dt. is this really correct approach?

[@marcosdanieltorres7253]

If im not mistaking is because canvas tries to run at 60fps by default and 1 / 60 is 0.016 which is how much one frame takes. He is assuming it will run at 60fps given that the calculation is simple

[@sivabalan_m]

@@marcosdanieltorres7253 Thanks. I understand.

[@vincenzopanella2705]

What I have found difficult in physics simulation is not integration itself, but dealing with collisions using complex spahes. I'd like to know your opinion on new methods implementing signed distance fields using the gpus, do you think that this might be considered a viable method for future physics engines ?

[@voxelltech]

From my little experience in physics simulation, SDF works really well for static objects.

[@ThankYouESM]

I've been trying to learn Javascript for over 3 years... then completely gave up... then a friend of mine introduced me to python which took me 4 months to learn... whereas I had fun quickly creating a whole lot of fun projects while hoping the processing speed of Python would become more than fast to make AAA 3D games... then after 5 years waiting... decided to give Javascript another chance, but... couldn't seem to progress any further... it's like I need to do 5000 times more work to finally finish a project. Had recently thought of making a project that would convert Python scripts to become purely Javascript much as possible, so... I hope to start this journey from here.

[@Erik_Caballero]

Have you considered using Julia? It reads a lot like Python but it's as fast as JavaScript for most applications.

[@ThankYouESM]

@@Erik_Caballero I do very much thank you for that recommendation... it's just that I seem to have found the ultimate solution because of ChatGPT this December to instantly make awesome Javascript applications by simply converting Tkinter scripts over which it figures out how to optimize. Also... I got it to design what Tkinter version of Blockly, but looks like just a professional text editor until I click near the front of the line that expands it to become the error-resistant template even providing a long list of what it predicts could be what I want it to write. However... because of the time restraint and being new as it gets... I haven't really gotten around to testing it all out... just be gathering scripts galore, then asking the ChatGPT where to place in a skeletal structure as not to overwhelm the network.

[@User-jr7vf]

It would be very helpful if you make a video explaining step by step what is needed to write and run simulations, such as: (1) download Microsoft Visual Studio (2) how to write the actual code (3) how to compile it.

[@JehanSaren]

Stingy

[@zaidalghazzi8699]

you area crazy one cuz you act like no air force at all and that is not real, so all code is useless.

[@DasAntiNaziBroetchen]

Subject to Allah to find the truth. Your existence is useles without the great one!

[@1st_ProCactus]

JavaScript.. that shouldn't even exist. Garbage

[@TenMinutePhysics]

I use c++ for work because its the fastest. Without the fancy stuff JavaScript looks exactly like c++ and the amazing thing is, it's almost as fast look at the demos! The JIT is a masterpiece!

[@captainpissof9632]

why doesn't this dude have like 10mil subscribers?

[@atomictraveller]

it's a question that i personally find easier to answer after having ma5ons crap all over my life. hth.