Well in another video, you could always do how to do explosions in Blender Vs Houdini Vs Cinema 4D couldn't you? (shrugs) mix and match videos and keep them interesting with different ideas, i'd say! Cheers for the vid anyways man!

@@MGSBigBoss77 Thank you! I've been thinking of a comparison like that for a while. I'd need to learn Houdini and C4d for that, which are software I already wish to learn, so it shouldn't be much of a problem there. I guess the problem then would be being able to tell when I know enough about how to use them, to have a fair comparison.

@@IagoMotaVFX Hey?, i'm constantly studying them! And have several magazines and software to keep on going in that endeavor. Never give up, as there is no end to learning new things dude!

Question. How long would such frame to render in HD res and can you project a mov video file onto the smoke?

@@claudianreyn4529 The render speeds depend entirely on your PC specs, the complexity of the shading, the quality of the render settings, and how much screen area the smoke covers. In my case, it can take anywhere between 10 and 60 minutes to render a single frame. And yes, to project an image onto the smoke, you could use the more common way to do this; which is with a spotlight with an image sequence texture. Though, you can also try, in the material node tree of the domain, plugging the image sequence texture into the color input of an emission shader, and mixing that with the rest of the smoke shader. (this would probably work best for a thin layer of smoke, instead of a whole plume)

Don't know why the smoke still processed faster than my rx580, even my Vegas.

just get embergen it's realtime and looks better

@@brysonadams8510 Yes and no. Other than VFX embergen is great and when you use it just from time to time i wouldnt spend the money on it.. I had great results in mantaflow.

Me: *cries with Intel HD Graphics* :(((

@@brysonadams8510 Just download Houdini apprentice and import the vdbs.

why he does these things? Im curious

That's the power of Houdini! (and experience, too)

Huge thanks, Ciaran. And good luck on the simulations!

And I'm back to this video once again. You make the best smoke simulations I've seen. They look so real!

Thank you, Noud!

That sounds interesting. I should try that out.

Yep. There's also the advantage of making the explosions very detailed in a matter of hours (depending on the computer, of course). But, on the other hand, 2d does have one pretty big advantage: the artist has full control over how it looks and moves. So they can design the explosion to look very interesting without having to "fight against" the software.

Thank you! And great results! You definitely got that mushroom cloud movement right.

thats easy though. add cube, quick explode, play with the settings. add material and voila

Thank you! Yeah, from my experience, particles tend to work best when their directions are more specific, as opposed to random. But, even when they are moving *evenly* in all directions, there can be problems. Apparently, when you have a small sphere emitting thousands of particles outwards, it's practically impossible to get the smoke to continue in that movement after the particles have disappeared. But yeah, particle emitters tend to produce very messy results. Hopefully, there will be ways to improve this in the future.

@@IagoMotaVFX About the impossible smoke to continue in that movement, not so sure about that..I thought I could get that to continue, but then I haven´t tried with 1000 particles. Will have to take a look at it again. But as I said, they could use a force grid for the particles instead of just particle fluid emission, will have to set something up with lightwave and see how the vdb export from lightwave works for that, and perhaps a demonstration of me plugging the nodes in to this kind of force channel. The problem with Lightwaves vdb gas solver fluids, is that it´s in a very early stage and crude, in terms of visual display and speed of the simulation, currently I have choosen to not invest in turbulenceFD for various reasons. I did buy in to the 2019 lightwave upgrade just because I found the VDB tools so interesting, but the 2020 release wasn´t very exciting to see..some improvements there perhaps..but not enough for my liking, so I need to focus more on blender and let Lightwave rest a bit, save some money while Lightwave has to grow a couple or years, or stagnate more. Mantaflow I just haven´t liked so far in terms of workflow and feedback. The older fluid system I like better in terms of using, but then again it´s slow to simulate as well. TurbulencFD is probably much faster when dealing with larger scale sims...since it uses the GPU, and it´s kind of easier to get cleaner look in the simulation, on the other hand..you have to use the older Lightwave volumetric system and either render in passes, or convert to VDB for rendering with the newer PBR volumetrics in lightwave. So ..nothing really satisfy me with all this, but Embergen seems way cool in terms of quality and simulation speed, I would need to verify the export speed and how long time that takes, and it would need more than static object emission, but otherwise really sweet simulations with it.

@@PrometheusPhamarus I really hope it's possible, since expansion is what makes an explosion an explosion. The thing is that, even in the most perfect conditions, I never managed to get the smoke to expand in a spherical shape. It always becomes very irregular. I might be missing something very crucial, though, so I'm not sure what to do about it for now. Mantaflow now has smoke guides. But unfortunately, it doesn't support particles. Unless, if you were to use the liquid particles of a fluid domain, but that's pretty different. I've never done any testing with guides, which is unfortunate. There might be a lot of fantastic ways it could be used in, so I should try it out. Embergen is very surprising in quality and speed. There are several cases where the smoke doesn't behave accurately, but I'm looking forward to seeing more done with it.

@@IagoMotaVFX Not sure which Embergen you are trying if you already purchased it, or only used trials? I just got a notification today that they are releasing the Betaversion now..with 14 days of fully working export for the vdb. I got to try that soon, but I still have the latest Blender build to install..got to get going not, interesting times with the fluid tools now and vdbs.

@@PrometheusPhamarus Oh, I haven't used Embergen, I've just seen some simulation demos on KZfaq and Twitter. Personally, I'm more interested in Houdini now that I've been shown that the apprentice version can actually export VDB.

I usually just add a point light to give the smoke some general lighting (the simulation has to be baked again for this change to be visible). Other than that, I frequently adjust the thickness in the domain's Viewport Display panel (in the smoke physics settings).

Ideally, after it starts to look good, I would use the live preview after every bake. But, since I'm lazy, I don't tend to do that as often as I should.

Thank you! I haven't looked into importing Houdini vdb into Blender yet. Since Houdini Apprentice (the version I'm planning to get) can't export to other software, I doubt it'd export vdb, so I've never dome much googling of that subject. But, I'm sure it might be possible with the full version of Houdini and the upcoming version of Blender (2.83). That blender version will have vdb support. I've seen plenty of Embergen fluids imported into Blender through that mechanism.

@@IagoMotaVFX Actually the free version can. Look up Jonathan Kron, he has a very good explanation on how to export the vdb files. You don't need Houdini indie, you can use apprentice. Shocking I know! I just wanted to make sure you knew about it since you said mantaflow is being a bit glitchy, but now I'm absolutely glad I said something.

@@brightgarinson3099 Oh, that's awesome! I really thought that was too good to be true. Well, now I have no excuse not to start learning Houdini asap. Huge thanks!

@@IagoMotaVFX I can't believe I helped my favorite KZfaqr, what a great start of the day.

Thank you! Well, first you need to open the files in Blender 2.83. It doesn't seem to properly load domain settings in previous versions. After that, you need to select the domain, and under "Cache", you need to set the cache directory (pretty much any directory will do, just make sure the hard drive has a lot of space). Then, you press "Bake Data" and wait for Blender to complete the simulation. By default, the project files here don't come with all the cached smoke data, since that would make the downloadable file weigh hundreds of gigabytes.

With ground shockwaves, I'd avoid relying too much on force-fields. It's helpful to consider that the shochwave we see on the ground isn't the shockwave at all, but just the dust kicked-up by it. So, while the shockwave expands at the speed of sound, the dust doesn't really do much other than rise and move out just a bit as it's pushed by the air. So, I usually just make a quickly expanding circle (with an emission texture for variety). If there's a turbulence force-field, parts of it will rise nicely from the ground, but make sure to decrease its intensity over time, to allow the dust to settle. I usually enable "Initial Velocity" in the emitter settings so that the smoke immediately follows the expansion (just make sure to keep the strength low). An interesting detail about shockwaves is that they tend to leave behind an area of low air pressure. So after the shockwave agitated the ground, it's actually pretty common to see a sort of "implosion" where the dust begins moving toward the explosion, as outside air begins to rush back in. Which is why, instead of using a force field to push the smoke out, I use one to pull it in. This way, the smoke begins by immediately expanding outwards, but is then accelerated back in. All these movements are somewhat subtle, by the way. To make the shockwave look like it has a gradual fall-off, I just add a sphere gradient texture and use it to control the shader density. This way the dust gets thinner the further away it is from the explosion.

In most cases, grenade explosions don't really produce much of a cloud themselves. Most of the smoke is dust risen by the blast, and a little bit from the combustion. For that, I suppose a quick particle explosion could serve as a decent emitter.

also, do you have instagram account? cant find you

Thank you! Yes, I'm planning on doing nuke stuff this year. Can't guarantee a tutorial, since I have to know if I'm even able to begin creating that kind of an explosion. I don't have an IG account at the moment.

@@IagoMotaVFX fuck yes!! Cant wait ti see your results!

That depends a lot on what is going on in the explosion. If it's a simple plume with the least amount of turbulent movement, then it should take somewhere around four hours to bake in my pc. If it's a complex simulation, it could take over 10 hours. Render times depend on even more factors: camera distance to explosion, complexity of the node tree, sample size, etc. In my pc, it's pretty common they take somewhere between 30 minutes to 1.5 hours per frame. In this video, I tried to keep the explosions simple, so most of the stuff didn't take longer than 50 minutes/frame.

Blender 2.82. Newer experimental builds have some updates to Mantaflow, so it's good to check them out as well.

At 3:15 I duplicated the sphere and moved it to a new layer, just so I would have a working backup copy. At 3:19 I duplicated all the vertices of the sphere in edit mode, so that I'd have one object comprised of multiple spheres.

​@@IagoMotaVFX Much useful tips! Last thing: so you duplicated the vertices, but I see that you can you select the duplicated with just one click: did you create vertex groups? (in case I want to modify the duplicated vertex)

@@RaffoVFX I hovered my mouse over it and pressed "L" to select all the linked vertices. Makes it a lot easier to select separate chunks of a mesh.

@@IagoMotaVFX thank you!!!

@@RaffoVFX No problem!

Thank you! I'm not sure if I understand the question, though. Do you mean the finer noise in the simulation or the general shape? I know I could've certainly worked more on the general shape, but I didn't want to spend too much time on it, for the sake of the video. Simulating a plume that doesn't get "shredded" over time is really difficult, unless the simulation just has a really low vorticity. And, unfortunately, there isn't a direct way of controlling that. Usually, I experiment with a bunch of settings, like the ones related to buoyancy. But those don't tend to do much. This is something I'm really trying to figure out; getting simulations that look "compact" (for the lack of a better word, I guess) is really important in many cases, just very difficult. And, about some of the finer noise, there is a way of controlling the size of it in the Noise settings. But wavelet turbulence in Mantaflow (Noise) doesn't seem to behave accurately yet. From the looks of it, it doesn't follow the smoke properly, so I'm hoping to see some issues around it fixed.

@@IagoMotaVFX cheers, yeah I think that's what I was seeing, is small plumes that don't get shredded... they look fine for the first second or two, but then I guess I subconsciously expect them to turn into smoother smoke and they remain tight little balls. I think I need a little more horsepower to play around with this stuff.

The dragonite is mine now.

Thank you! I'm currently doing some RnD for the next video. I'm not yet sure of when it'll be complete, but it's a "quick tip" video, so it shouldn't take too long to make.

@@IagoMotaVFX Okay! I can't wait!

Obrigado! Acho que comecei usar o Blender uns oito anos atrás. Porém, foi só recentemente que comecei a considerá-lo uma opção de carreira.

@@IagoMotaVFX Tenho pensado nessa ideia ultimamente, mas não chego a ter nem 1 ano de Blender. Ainda me parece muito distante e complicado de se alcançar, mas tenho muito interesse. Parabéns novamente pelo trabalho, me inspirou bastante a continuar!

Yeah, I've tried it, but unfortunately, it makes the simulation far too slow for it to be workable in my PC, so I usually just keep it turned off.

@@IagoMotaVFX cool cool. I have a 1070 and a ryzen5 and it’s acceptable. But I don’t mind a 2 day bake. Great vid though explains what I was saying other day to someone that particles are ok but not brilliant. I couldn’t say for why except blocky and resolution dependant acting as planar emitters. Hands down best smoke work I seen in blender and not at 512 resolutions.

@@ghettomining7580 Thank you!

Usually takes me about 8 or so hours to render the whole animations. I tend to keep a lot of the render settings at a pretty low resolution to make it easier.

Yeah, Mantaflow has changed quite a lot since this video was made. For instance, the way gravity and force-fields affect the smoke has been changed. I think that was to make the scale of the simulation look more accurate (Though, I'm not sure if I remember that correctly). So, different settings would have to be used today, like maybe much higher temperature and initial velocity? It's still hard to tell how to compensate for a lot of these changes, unfortunately. I guess we'll have to wait for everything to become a bit more stable.

@@IagoMotaVFX will some more experimentation I managed to fix the error, so that the object was actually emitting smoke. But it never had that cool rising effect of basically any explosion ever. The smoke just lingered there and rose super slowly, not at all the effect I imagined. I’ll try turning up the temperature but there are so many values to change I don’t even know where to start XD

Hello! Sorry for the really late response. Unfortunately, I haven't planned on making anything similar to that. So, I'm really sorry about that. I'll say though, the way I would approach that would be by starting with some "previs" to break down all the elements, plan the color scheme and build depth into the effect. After that, I'd make the different parts individually, preferably in separate .blend files to keep things organized and efficient. In the demo, it looks like the main part of the effect is a pretty simple (at least in concept) ring of flowing smoke. However, there the smoke itself isn't rendered and, instead, the particles are moving in its velocity, which is something that could be achieved with the Smoke Flow force field. The rest of the elements look much simpler, particles moving in a vortex motion with some turbulence added to them. And the outermost element there looks like it could be achieved by rendering a single frame of a crazy number of particles, and using that rendered image as a rotating texture on a circle with surface displacement, making it look like the particles have multiple levels of movement. I suggest that for some of the slower "flowing" movement you try using the "Force Field Settings" panel in the particle settings. It basically turns all of your particles into force fields (so, make sure to use that on a separate particle system, with much less particles). Using the "Vortex" force field and messing with the particle settings can sometimes give almost smoke-like movements. Anyway, I hope this wasn't boring to read. Just remember, this is how I'd first try to approach an effect like this. It's likely that this is much more complicated than it seems to me now. But hopefully this helped at least in some way.

man, that's not fair

For that case, animating the Time Scale slider would probably be the most effective way to do it.

That's something im currently trying to figure out. Im guessing you take render image sequences of each angle of the explosion and use in a particle system in unity (with flipbooks)...

That would definitely be more realistic. The reason I tend to avoid explosions like that, though, is because they're incredibly difficult to get right. In those kinds of explosions, since the smoke is really dark, I won't be able to get away with leaving the fire inside the smoke to have it just scatter the light. The fire will have to be completely exposed, and getting the movement of fire like that to look good is tough. Before Mantaflow, this could have been achieved quite reasonably. I used to separate the domains into "Smoke Domain" and "Fire Domain". (I still simulated smoke and fire in both to match their movements, I just rendered the fire and smoke elements separately) The Fire Domain had a considerably lower resolution and vorticity than the Smoke Domain. Knowing how turbulent and detailed fire is, this might seem unreasonable at first. But, lowering the base resolution allowed me to strongly increase the High-Resolution samples. The quality of the fire then depended entirely on the wavelet turbulence. Wavelet turbulence is a pretty accurate way of simulating details, so it usually did a great job whenever I wanted to get realistic large-scale fire. Right now in Mantaflow, though, I've been experiencing several issues with wavelet turbulence. It doesn't seem to be following the velocity of the smoke properly, making it look "noisy". So, currently, something like this would be a lot more difficult to get right.

@@IagoMotaVFX Okay, first of all thanks for your really detailed answer, I don't expected that! Maybe it would be possible to examine that a little bit more in the future? I'm very interested in fluid simulations in genreal (and I've done a couple dozens of them in the past), so that would be awesome. That with the Wavelet Turbulence seems indeed a little bit odd after I tried it out, do you think that will get fixed in one of the next updates? Or aren't the developers aware of this problem? If I may ask: do you informed them about this maybe already? For your future content I think it could be interesting if you would do some sort of a lifestream tutorial. That way you can teach a lot more of your Emitter workflow. The Emitters are my biggest problem, until now I don't have the awareness in the right amount of those things. Maybe you could adress these things a little bit more in future videos, kind of like in the latest "doing random stuff" video? Either way, thank you for your great tutorials , stay healthy and safe in these times and I can only say that I'm very hyped for your next videos :)

@@friedrichrengah4747 Yeah, there are many things I want to try out in the future when Mantaflow is more stable to work with. Fire of different scales is one of the things I'm really considering. And, usually, if I want to get fire and smoke to both look right, I have to separate the domains, so it's very likely I'll tackle that in the future. I've been thinking of making a bug report on the wavelet turbulence issue. The tricky thing is to make a very clear demonstration to really show the difference between Mantaflow and 2.81. I'm keeping a close eye on the bug reports. Gottfried Hofmann's reply on developer.blender.org/T71391 shows an issue that I think might be related to or a side effect of the problem we're talking about. So, I'm also waiting for developer's responses there, to know whether a new bug report is going to make a difference. I've been thinking of livestreaming for years now! I just haven't gotten around doing it, as I'm not good at coming up with things to say on the spot while trying to also focus on the simulation. Maybe after a few more "doing random stuff" videos, I might gain the skillz and the confidence. But yeah, thank you for the regards! I hope you're staying safe as well

Iago Mota sorry for bothering you once more, but I‘ve encountered a big problem lately. I‘ve rebuilt your fire shader for a simulation I did and in the rendered view it seems alright. But every time I try to render, everything is rendered as smoke, no light is emitted. I checked everything possible in my mind what may cause this problem. Checked the Light Paths, the sampling, the volume interpolation and so on. I then tried other blenderversions. I did the sim originally in the 2.83 beta, then I tried it in 2.82 and 2.81. Both of these didn‘t work. Did you may once encountered a similar problem? Is there a small option that I missed? I searched the Internet for several hours but didn‘t found a thing relatable to my problem. Would love to get an answer, maybe something that helps... If anyone of you had a similar problem and got it fixed please write it down. Thank you :)

@@friedrichrengah4747 I've only come across this issue once and never really understood its origins. In fact, the simulation at 12:58 is that one time it happened. Both domains were supposed to have fire (one for the detailed external fire and the other for just the scattered light inside the smoke), but only the former is visible in the render. For me, it renders correctly in the viewport but not in the final render. It might be related to this bug, but I'm not sure: developer.blender.org/T73766

After setting the new folder directory and baking again, it just reverts back to the C drive? Like, no matter what, it just doesn't bake to any other drive? That's actually a problem I haven't seen before. Do you know what version of Blender this is happening in?

@@IagoMotaVFX The problem is that I select the OpenVDL format before baking. But you just want to fry it on drive C. If you set another drive as the baking path or the Blender file is on another drive, it will no longer do anything. It only creates empty folders but does not include baking information files. On drive C, yes, but there's no room to bake there. v2.83.4 version

@@imoteep4984 Oh, in that case, does it work is it's baked in the Uni Cache format? If that's the case, I think it might be a problem with the drive itself. But, unfortunately, since I haven't seen this problem before, so I can't be certain. I really apologize for my lack of knowledge there!

@@IagoMotaVFX If set to Uni Chace or Raw Cache, it will run on any drive. But OpenVDB unfortunately only wants to bake on the C drive and there is not enough storage space there.

@@imoteep4984 Yeah, sorry, I really don't know what might be going on there. Maybe it might've been a bug that was fixed in newer versions, like 2.9, but I can't be sure, unfortunately.

Go to View -> Viewport Render Animation. That'll quickly render a series of images in your output directory, that you can then import to your editing software and play in real time.

@@IagoMotaVFX Unfortunately, even viewport render runs hellishly slow for me

@@Sh-hg8kf Ouch. In my case, when the smoke resolution is high, it can take somewhere around 30 minutes to render the whole animation through the viewport. It is pretty annoying, but it usually pays off.

@@IagoMotaVFX 30 minutes on the viewport?! Yikes