@@KnakuanaRka There is a "Not interested option" in the drop down menu of each thumbnail in your YT feed. There is no need to click on a video just to downvote it, which is has a negative impact on the video. But maybe you mean that some people are unaware of that option?
Seeing this makes me think that it's not actually that far fetched to think we're living in a simulation. Goddamnit that knitted cloth simulation. Of course, it's still an isolated feature thatl takes fuck tons to render, but imagine this in a couple of decades, hell perhaps even less. This could be used in game engines, expanding from cloth, knit and hair to all types of other materials. Footprints could leave a completely accurate and realistic mark based on the collision between the foot and the dirt, changing based on the pressure applied by the heel or toes, depending on the surface angle etc. No more clipping of clothing and limbs. Hell, it could mean so much for simulated water. This is really awesome.
They run the simulation, find which grains of sand are the ones that end in the position of the "SIGGRAPH" letters, then reverse-lookup to color them starting in the beginning of the simulation. The initial column of sand looks like a turbulent jumble.
@@anbietm4617 Not difficult at all. The difficult part is simulating the sand and cloth. Once you have the simulation, you just go the end of it, color the letters, and rewind the baked data. As you can see in the end, they simply used the final sand simulation, and then colored it. It is the exact same simulation.
This seems possible as long as simulation is deterministic (or gives the same result with same input seed). Though, even if that's not the case, I guess that it's still possible to do it if every intermediate state of simulation is saved and later recolored.
There are so many people coming up with amazing physics simulation "Methods" but since they just sit and wait for someone to buy them there will never be a physics system that collects all the best methods and uses them! C'mon people! These are formulas that CAN be put together on a physics engine or a 3D modeling software even if you're not using a billion particles! The formulas still work! When are these methods going to become available for the small developers? Only Autodesk or Sony have the money to buy all these methods and put them together, but they're very slow and not as connected to the communities to find them and say "oh, let's try this one out" in the first place! And even if they do, they won't release a small and quick experimental product using the super engine, they'll start a new massive product that will take them years to make and won't release any early access. A small developer finding these and making a small interactive simulation is the only hope for non-developer-regular-people to touch these methods before 2025 or something. Yeah, that's another thing, most of these physics methods are from years ago and they're still out of reach from the public. If they're ever used, it's by cgi movie studios, and they'll never release them for public software.
Yes the paper is at www.seas.upenn.edu/~cffjiang/research/cloth/paper.pdf with supplementary technical document at www.seas.upenn.edu/~cffjiang/research/cloth/tech.pdf
@@cffjiang hey, i just recently found bout siggraph, I was wondering what kind of maths approach/field plays role in writing such papers as yours? I'm currently doing by second year BSc and plan to do research in pure maths. So I was thinking if this could be an exciting side job? Thanks!
im building a worksation computer specifically for 3D work like character model, autocad, fire and ocean simulation, or destruction and fluid particle simulation but im not to sure what kind of hardware i would need, mind i ask for some help? im currently planning on buy these 2 main component for my 3D workstation build, please give me come advice and recommendation. can a RTX Quadro 8000 gpu and a AMD threadreapper ryzen 32cores cpu simulate these kind of workflow as shown in this video with same amount of details and polygons? like 1 million or more to 10 million particles or polygons or triangles? my budge is 10k so i really dont wanna risk it, please give me some advice, thank you.
i think they mean to ask, what is simulating these? what is producing the animations? not everyone understands these complex new algorithms n stuff. and i don't either.
Why we rarely see it in actual 3d softwares? And why divelopers prefer to work apart from the rest of 3d world? They could study things and make nice presentations using blender, for example. I understand that divelopers want to sell their work to big companies, but how often this happen? Sometimes it is better to create something for free software and get hired (like Blender cycles creator was hired by Solid Angle)
1. Implementing something complex in Blender is not easy. The point of academic papers is to introduce a new idea, not to spend years implementing it. 2. Look at the end of papers like this and you will find a "limitations" section. SIGGRAPH presentations only show the best results that are obtained. They don't show the cases when the method fails. This particular paper may not be a good example since it does look quite beautiful and robust, but if you just search on RUclips for SIGGRAPH papers you'll realize 90% of them would be worthless to "implement in Blender." 3. If there is money to be made, the authors can file a software patent. If they don't patent it, you can go ahead and implement it in Blender yourself instead. The authors are using the results of their simulation research to launch a startup www.jixiefx.com/ so they *definitely* don't intend to give away this technology for free.
![[SIGGRAPH ASIA 2017] Conformation Constraints for Efficient Viscoelastic Fluid Simulation](http://i.ytimg.com/vi/bRfAeBTfS2g/mqdefault.jpg)
![[SIGGRAPH 2018] Toward Wave-based Sound Synthesis for Computer Animation](http://i.ytimg.com/vi/su6z9snjU-U/mqdefault.jpg)
![[SIGGRAPH 2018] Moving Least Squares MPM with Compatible Particle-in-Cell](/img/1.gif)
That dude jumping like YAY I GOT A PONCHO!!! Serious though this is amazing
a
Who the hell downvotes a video like this??? honestly?!
Sorry sir you are in the wrong platform
The people who created the Lagrangian MPM. Duh.
Because they aren’t personally interested and want it out of their feed?
@@KnakuanaRka There is a "Not interested option" in the drop down menu of each thumbnail in your YT feed. There is no need to click on a video just to downvote it, which is has a negative impact on the video. But maybe you mean that some people are unaware of that option?
i dont understand the physic explanation of these vídeos, but i Love them
hey chef, what's for dinner
cloth
slime
ok
The bestest of obscure memes.
What meal can one make with a cloth or a slime?
@@idot3331 lasagna?
Seeing this makes me think that it's not actually that far fetched to think we're living in a simulation. Goddamnit that knitted cloth simulation. Of course, it's still an isolated feature thatl takes fuck tons to render, but imagine this in a couple of decades, hell perhaps even less. This could be used in game engines, expanding from cloth, knit and hair to all types of other materials. Footprints could leave a completely accurate and realistic mark based on the collision between the foot and the dirt, changing based on the pressure applied by the heel or toes, depending on the surface angle etc. No more clipping of clothing and limbs. Hell, it could mean so much for simulated water. This is really awesome.
Amazing work! Now if only we can make it realtime...
Probably entirely doable if we reduce the number of physics polygons and use plenty of interpolation.
Most impresive thing I've ever seen on physics CG. Congratulations!
I don't know how to write a single line of code but I keep watching these videos
The knit objects are particularly lifelike in behaviour. This is impressive!
3:58
Small wooden man is happy to get poncho.
VR KNITTING SIMULATOR!!!!
i am excited for the next phys game
Fantastic! Awesome work!
Agreed! Beautiful results.
The best siggraph video I've seen so far. I just wish I could have this in my video games today :D
maybe in 30 years, stay healthy
Make it gpu compute and make it an addon for 3d software, and I would pay a lot for this.
Looks incredibly realistic
I can't wait to play with this in VR one day! Incredible work.
I'm more curious about how you simulate the SIGGRAPH thing in the beginning of this vid.
They run the simulation, find which grains of sand are the ones that end in the position of the "SIGGRAPH" letters, then reverse-lookup to color them starting in the beginning of the simulation. The initial column of sand looks like a turbulent jumble.
@@AerysBat that seems difficult
@@anbietm4617 yes
@@anbietm4617 Not difficult at all. The difficult part is simulating the sand and cloth. Once you have the simulation, you just go the end of it, color the letters, and rewind the baked data. As you can see in the end, they simply used the final sand simulation, and then colored it. It is the exact same simulation.
This seems possible as long as simulation is deterministic (or gives the same result with same input seed).
Though, even if that's not the case, I guess that it's still possible to do it if every intermediate state of simulation is saved and later recolored.
Inb4 the "Put it in Blender! Omigosh--start a Kickstarter!" crowd. lol
Fantastic stuff, by the way. Good work, guys.
Very amazing cloth simulations but I'm curious to see what the hair simulations would look like.
There are so many people coming up with amazing physics simulation "Methods" but since they just sit and wait for someone to buy them there will never be a physics system that collects all the best methods and uses them! C'mon people! These are formulas that CAN be put together on a physics engine or a 3D modeling software even if you're not using a billion particles! The formulas still work! When are these methods going to become available for the small developers? Only Autodesk or Sony have the money to buy all these methods and put them together, but they're very slow and not as connected to the communities to find them and say "oh, let's try this one out" in the first place! And even if they do, they won't release a small and quick experimental product using the super engine, they'll start a new massive product that will take them years to make and won't release any early access.
A small developer finding these and making a small interactive simulation is the only hope for non-developer-regular-people to touch these methods before 2025 or something.
Yeah, that's another thing, most of these physics methods are from years ago and they're still out of reach from the public. If they're ever used, it's by cgi movie studios, and they'll never release them for public software.
it is unbelievable, incredible, so realistic.
Wonderful work. Haven't read the papers yet but amazed at the quality of your thin sheet simulations. Are you sure these are MPM only?
Now I believe your saying that "MPM is for everything"
young minchen！
the fibers tearing was so good!
Unbelievable! this is gna change things, now make software so you can make tons of money! I would buy this
This is just .. wow. Awesome work!
Can the effects in the video now be done with Houdini? What do you think of that? And why aren't you updating your video? Gleason loved your video
4:48 876s per frame? u render that 3 days? and minimum 2 times (for 0.1).
Render farms are a thing
Brilliant work guys!
Why these awesome tools do not end up integrating into c4d or other DCC applications? I wonder
I want a knit poncho!
Simply perfection.
imagine seeing this kind of quality in video games
@@ControversialOpinion worth it
How were the yarn cloth models made?
can you simulate the blowing of a balloon ? and when it goes flat ?
do you mean something already in this video? the "cloth and slime" chapter?
Must be amazing to watch all the coding and thinking bear fruits.
Loved the work.
Is there a “computer simulation reaseacher” to English translator?
how can this be done in javascript
The Matrix era is coming sooner than I thought.
It looks so dope
is there any 3d software that have this thing ? :D
How is the starting knit generated? I've always wanted to make knit-level sims!
Cool stuff
computer used?
Wonderful results. GJ.
What’s MPM?
why we dont have clothes liek this on games?
This is the actual proof that science is the truth and can be used and replicated indefinitely.
Djordje Petrovic stop tipping your fedora so hard dude. It's not even _trying_ to emulate correct physics, just more realistic-looking ones.
素晴らしい技術デモ
ビデオゲームの世界に早く反映させて欲しい
*Oddly satisfying*
Beautiful stuff!
Can you include links to documentation or any papers you've written?
Yes the paper is at www.seas.upenn.edu/~cffjiang/research/cloth/paper.pdf with supplementary technical document at www.seas.upenn.edu/~cffjiang/research/cloth/tech.pdf
@@cffjiang hey, i just recently found bout siggraph, I was wondering what kind of maths approach/field plays role in writing such papers as yours? I'm currently doing by second year BSc and plan to do research in pure maths. So I was thinking if this could be an exciting side job? Thanks!
god bless simulation programmers
Sensational stuff!
for some reason if find those half spheres sliding on the ground very cute...
They remind me of Mario 64, those ones that spew fire
Looks fantastic!
What is your software and hardware?
It is made from research code base. It runs smoothly on a standard machine with Intel Xeon CPUs.
I never would have thought cloth to be an elastoplastic
Why not use this in games for ps4, for example ???
Because don't run
how are these rendered? Blender?
im building a worksation computer specifically for 3D work like character model, autocad, fire and ocean simulation, or destruction and fluid particle simulation but im not to sure what kind of hardware i would need, mind i ask for some help? im currently planning on buy these 2 main component for my 3D workstation build, please give me come advice and recommendation.
can a RTX Quadro 8000 gpu and a AMD threadreapper ryzen 32cores cpu simulate these kind of workflow as shown in this video with same amount of details and polygons? like 1 million or more to 10 million particles or polygons or triangles? my budge is 10k so i really dont wanna risk it, please give me some advice, thank you.
what is the software used?
Amazing!
AMAZING!!!
This is something very cool, but the sound is TOO LOW.
wait, there's sound??
Amazing
OMG OMG OMG, I see this, but can't believe this.....
love it!
waht this porogram
3d s max or maya or blendr ?
MS PowerPoint
@@TMZStreetWorkout😂😂
What software do you use to make these simulations?
It's made from research code, not a software.
i think they mean to ask, what is simulating these? what is producing the animations?
not everyone understands these complex new algorithms n stuff.
and i don't either.
Is there any way I can use this in any of my animations? lol
This is the demonstration video of our new research paper. Currently there aren't any existing software packages with the technique integrated.
Chenfanfu Jiang I had a feeling, but man would some of these siggraph things help make animations.
For some of these siggraph things I would pay hundreds lol
doesnt that work in blender already with molecular?
Marienkarpfen Molecular sucks, they need to add the ability to make the particals deform a mesh or something.
Really good job! (y)
this is why i avoided computer science
Houston Helicopter Tours Inc. honestly I feel like it is sometimes this shit seems so difficult
Aida 2.0 framework
Why is this in my recommendations
*Just* 2 minutes per frame.
this is lit
This cant be a software oO
Wow
say hi to nvidia ceo, when he contacts u :) that impressiv shit, must be sold for the right price !!
low volume
>anisotropic elastoplasticity
I vaguely can comprehend what the hell that means.
You can't even type it correctly.
@@VictorNascimentoo wew you caught the typo on a throwaway comment
@@gavindillon1486 hahaha take it easy pal
tutorial? lol
ole toro
this is unreal
Literally.
4:23 awesome for a condom commercial ;)
I can smell a lot of GPU burning together
Why we rarely see it in actual 3d softwares?
And why divelopers prefer to work apart from the rest of 3d world? They could study things and make nice presentations using blender, for example. I understand that divelopers want to sell their work to big companies, but how often this happen? Sometimes it is better to create something for free software and get hired (like Blender cycles creator was hired by Solid Angle)
1. Implementing something complex in Blender is not easy. The point of academic papers is to introduce a new idea, not to spend years implementing it.
2. Look at the end of papers like this and you will find a "limitations" section. SIGGRAPH presentations only show the best results that are obtained. They don't show the cases when the method fails. This particular paper may not be a good example since it does look quite beautiful and robust, but if you just search on RUclips for SIGGRAPH papers you'll realize 90% of them would be worthless to "implement in Blender."
3. If there is money to be made, the authors can file a software patent. If they don't patent it, you can go ahead and implement it in Blender yourself instead. The authors are using the results of their simulation research to launch a startup www.jixiefx.com/ so they *definitely* don't intend to give away this technology for free.
k
Nods... nods...
ахуенно
Lost interest at "2 minutes per frame..." :/
magicstix0r why is that?
Because given infinite time a computer can process anything... Realtime is where the interesting research is...
magicstix0r thanks for the response
With the way computing power progresses it's only a matter of time until these techniques can be used in real time.
Wow first is taking jobs
Amazing!
Amazing!