At the current point, I am 100% certain Blender can't match Houdini in simulations. However, the fact we're even considering it, is a massive leap from the early days. A film studio could realistically use Blender now. Maybe in 10 years, Blender is the Wikipedia to Houdini's Encarta.
Film studios are already using Blender now. It's a great tool for the beginning of the pipeline (concept design + asset production). It totally sucks for anything that is at the end of the pipeline and it'll take another 5-10 years for it to reach the place where others are now and that's why it will never dominate the end of the pipeline like the competition does. Houdini especially is a fast moving target that is trailblazing hard and it already has an considerable advantage that doesn't get smaller, it gets bigger with every release.
@@RyoMassaki the problem with your assessment, is that nobody really believed Blender would come so far, not even the creators. So it's perfectly possible for it to be at feature parity with future Houdini in 10 years. It's a long time.
@@4.0.4 It's "perfectly" possible if SideFX would stop developing H right now for whatever reasons and if the Blender Foundation suddenly hired the best developers in the world and throws Millions of dollar at the problem. So in short, the chances are close to ZERO.
@@RyoMassaki don't get me wrong, I don't think the chances are high; I just wouldn't discount it entirely. The whole simulation workflow would need to be reinvented from scratch. The chances with ZBrush are a lot better I think.
@@4.0.4 I don't want to come off as too hostile or dismissive, I actually use and love Blender, but I have come to really detest the parts of the Blender community that dream of supremacy over all and other ego-driven nonsense. Which is completely delusional and based on (willful) ignorance. Blender will become better and it already has a solid place in various industries. But when it comes to simulations I am not very optimistic, especially the more experience I gain with Houdini. Even with my knowledge (and pessimism), the amount of complexity that is between those 2 seems to me much bigger than assumed. I don't think the BF has the necessary resources, and even if they had, I think they would be better spend elsewhere (animation for example).
That's a pretty good summary. Workflow is great, until you are buried in trying to edit one specific feature, shuffling the deck, like some 2 bit hustler, lost in a casino…
Houdini is also slowly taking over every single department of the production. It used to be FX / Procedural modeling only, now it's FX, Groom, CFX, Lookdev, Assembly and Rendering in so many places, including ILM / Weta / Framestore (in varying amounts depending on location). Give or take 10 years and there won't be anything but Houdini + Standard modeling DCC + Zbrush.
Totally agreed. Sure it used to be an Fx tool, but so many big and medium size studios are switching more and more to Houdini, give it a few years and maya will be entirely replaced . Can't say autodesk did much to avoid this from happening...
Just FYI, C++ is not a scripting language. It's a compiled systems programming language. There's a reason that game engines and stuff like the core of Houdini are written in C++ and C#, not in Python. They are very different beasts.
@@saricubra2867 I don't like Python and I think it makes Javascript look like a well-designed masterpiece, but it's used a lot as the glue between C++ libraries.
I was at uni with a guy who could read assembly and tell you what it did. He refused to program in it because he said compiled languages were more efficient.
@@KeithNeilson Its not as hard to read as you think, it can take a bit of effort and rereading to figure out more complicated parts of it, but the instructions themselves are easy to read and it just takes time. He is right that it is much more efficent to program in compiled languages, if you need the speed and control use C or C++. That said, it can be fun to program in for small personal projects (although some people think it is masochism). you can also use it for small parts of larger projects that need to be super fast.
not only all of the other softwares not even trying to catch Houdini in simulation.. but also Houdini slowly taking over there specialized area ( Copernicus for example vs substance) ... so as someone said " you guys make stuff quicker than i manage to learn them".
Great video! You’re absolutely right! Houdini is King of VFX simulations. Now when you combine Houdini, Blender, and render a you get a powerhouse system that is an amazing. I believe that Houdini when combined with other tools makes production pipelines better, faster, stronger, more flexible, and more robust.
Yep, I totally agree with you: Houdini is the master of simulations!! The beginning with it can be difficult (and sometimes frustrating) 🤯🤯, but when you start to understand the logic, it becomes so powerful!! 😅🤩 I already worked with some other softwares during my career (even Blender), but I never regretted my choice to work in totality with Houdini! 😆😆😉
Node based workflows are logically and mathematically superior to workflows that are not node based, it really is this simple. Also the reason why Nuke pawns After Effects
It's not that simple at all. Linear workflows like organic modelling, or keyframed animation, are *exceptionally* badly suited to node-based workflows.
@@mylesprospero8105 well, geo nodes is a very rudementry node system compared to Houdini and really only for stacking topology ontop of other topology. Most of all it is not using any simulation of forces like gravity, smoke or liquids. The whole physics simulation part is not node-base so far (Blender 4.2). It is super un-intuitive and tedious to even get a simple water simulation setup. Without additional addons it is close to unusable. But the biggest problem is it's performance and stability. The fact that Blender just crashes without any warning or chance to save your work, when it overreaches your VRAM, is what makes the difference between industry standard and a free swiss army knife. But who knows what the future brings. The best thing about Blender's developement is, that it is super open to community input and change. I don't know any software that has that much room for customizability. And it's free.
3dsMax + plugin like Tyflow/TP/Rayfire/Phoenix/Ornatrix/etc is a high end solution too, and it's more artist-friendly in the traditional sense for those that don't like working with nodes. Plus Max also has Bifrost now. I'm not saying it's better, but it's an option that's up there with Houdini.
How well do these separate plugins connect to each other? Can you combine them in a meaningful and efficient way? One of the reasons Houdini is so efficient - you can combine everything with everything and drive anything with anything while keeping all of it procedural and non-destructive.
I am 100% convinced in specific areas there are better options than houdini. Tyflow multifracture is better and faster than houdini booleans (by 100x). Rigid bodies is way faster in tyflow. Rigid bodies voxel based has no rival in thinking particles. If you add phoenix+tyflow+ max its still cheaper than houdini. But houdini power is data structure and unification, despite slower and less artist friendly tools.
I would argue that JangaFX may be a potential future rival for Houdini Yeah, their software's such as EmberGen, GeoGen and LiquiGen don't get absolute most realistic simulations out there, but the mere fact that those softwares are all real time, and with that level of quality for simulation. Definitely puts it up there as a potential rival
I have to agree with what the video said at the end, sure some would say X and Y can compete. I'm sorry but you can't genuinely try Houdini and still thinking its not unrivalled that after it
Houdini is certainly a very good piece of software, but it's probably not suitable for all profiles of 3d artist. It's software that will suit the most technical profiles.
Growing up thru the 90's, It really felt the future of gaming was going to be awesome. Gaming wasn't just hyper focused on graphics. It was focused on physics effects, rag doll effects, an assortment of art styles, having a range of gameplay modes, creative and fun gameplay mechanics, satisfying damage effects, in depth environmental interaction, environmental damage effects, engaging CPU AI to play against in single player modes, CPU bots in offline modes. Current video games focus on graphical nitpicking. It seems they forgot how much players enjoyed physics effects & other aspects that make games fun. Quirky graphics can work, if a creative art style is used. Add a selection of interesting gameplay modes, the list goes on.. I seriously don't know why gaming had to fall into this rut that we are in..? Also: {We have seen that even if companies delay games, it isn't improving the quality of games being released. Which is a sign there is a MUCH DEEPER issue going on with modern games. No excuses, devs used to make great games with much less and they would have given anything to have certain capabilities to utilize that are available today. It just takes actual passion, focus & talent to make great games. Sidenote: It's so crazy how many people are still supporting super mega rich companies putting out unacceptable products.. i don't understand certain people. I don't understand how certain companies can be messing up so badly. It's frustrating that other passionate people are not allowed to make a version of a type of game that we used to have but no longer get anymore.. Yet these huge companies get full control over if an unused game IP gets made & they do everything they can so no other versions of a game idea will ever come out. Even if there's a huge fan base around certain ideas and there's a ton of passionate devs out there dying to make a version of a certain video game idea.. that frustrates me beyond belief}
I have a feeling that software like Houdini are going to become truly unlocked through AI -not as a form of generative substitution to human ingenuity, but rather as a language interface. The node based system only works if you understand exactly how the different nodes interact with each other, the limitations, the possibilities, etc. This makes a very steep learning curve, and it also makes simple tasks very complicated (tho complicated tasks easier, as said by another commenter). If AI is introduced at the language interface level, instead of working with nodes directly, you could instead communicate in regular speech with an AI model that then builds the node system to your specifications. This can allow you to generate very complex simulations in a very short amount of time because the tedious labor is taken care of by the AI following your instructions -and at any time if you wish to fine-tune and tweak a specific aspect of the node system the AI has built, you can simply go into manual override and change the things you want, after the AI has built the main blueprint for it. When we think of how AI can assist in workflow, this is the kind of stuff I'm most excited for. Not the generic generative social media type stuff everyone gets mad about. Not the stuff that replaces jobs. Rather, the stuff that creates jobs, by minimizing the tedious aspects of things like simulation physics -removes a large part of the learning curve, adapts the software to our benefit rather than us having to spend a ton of effort to work on the software's terms. The famous saying of "most people don't want to understand computer language. They want computers to understand human language", this is where AI truly has its main benefits in this industry.
@@Tayosteph It's probably just a matter of time. We already see in raw code development how AI is used to create the foundation for many systems to help developers speed up the tediousness of manually writing everything word by word, checking errors, etc. Node systems in 3D and simulation software isn't far from that in how we interact with it. Glad to see you (and hopefully soon others too) understand this aspect and see the value in it.
I've used Modo, Blender, Maya, ZBrush, and now Houdini. Even from a modeling perspective, Houdini gives you the most control and is superior to the other packages, IMO. Once you understand how to use attributes and occasionally a little bit of VEX, alongside it's procedural workflow, no other modeling DCC comes close. It also comes with HoudiniEngine, which is invaluable in game dev pipelines and last I checked none of the other software have a bridging application as advanced as HoudiniEngine.
@@hound_of_justice Not true, maybe for viewport interactions but that's about it. Houdini can do everything those other programs can for general-purpose modeling. It's not like you are forced to be completely procedural 24/7. Even if Houdini doesn't come with a native tool that another DCC has, you can create it with ease. Just drop nodes into a subnet and create an HDA. It doesn't get easier than that.
@@caden6305 You are wrong. Houdini is a horror to model with for most stuff with exceptions like procedural modeling. There are reasons why pretty much nobody uses it as a main modeling tool and replacer of DCC like 3ds Max, Maya and Blender. It should focus on stuff its actually at least devent and good at instead of fighting a impossible fight with other DCC
@@hound_of_justice It isn't a "horror" to model with though...lol, maybe if you are a beginner in 3D I'd agree. Houdini should not be someone's first 3D program. I've clocked in 100+ hours on each of those previously mentioned DCC's and like I said, Houdini gives you the most control over constructing meshes out of all of those programs.
@@caden6305 Its not just me, its also the industries professionals. 3ds Max, Maya and Blender for example are much better for general 3D modeling than Houdini. The control that Houdini offers is a) unnecessary for the most part in this case and b) control comes with a huge caveat to it and that is learning curve and time consuming workaround for what? Its...just really not the best software to model with in general.
I don’t get it , is Houidini this good ? Or is it just needing a lot of knowledge and being an artist god to master it therefore it looks good ? I really want to understand this from people using Houdini cause I’d like to invest time learning it ( I come from 3ds max-UE-blender and substance background) I only watched some Vex videos and would like to know if learning Houdini is worth it if you know those previous package ..
@@zatlanibrahim5438 Other simulation packages don't require you to learn how to program code line by line, such as the plugins mentioned in this video for Maya and 3ds Max. Chaos Pheonix FD Autodesk Bifrost Thinking Particles Realflow Turbulence FD ZivaVFX ...and many more. There are separate applications like Embergen and Liquigen and Realflow that are offered, amongst many others. You can get fantastic results without ever opening Houdini and you don't need to have a programmer mindset to use them either. Houdini really only makes sense if you need a project with totally nondestructive workflow as that allows you to independently change things without negatively impacting other things up and down the chain of a desired outcome. Technical artists that actually need to change thousands of primitives in a scene by changing one variable, would definitely be less of a hassle in Houdini, compared to using other software that isn't as non-destructive in workflow; for anything else where that isn't required, the other software suffices and can produce very similar or better quality results in similar or less timeframe without any programmer or technical know-how, which is what we should be pushing in the industry anyway. Not everyone wants to program, nor is it necessary, outside of those who are making the apps. The artist just wants to make stuff at the end of the day.
It's good but the lack of guide on the internet is bad, the hardest part of Houdini is basically figuring out a problem that you can't google search, which happens quite often for me, Houdini community is quite small compared to other 3d programs so if you have a problem you have to solve it on your own or paid expensive course....
You do genuinely get better results out of the box in most areas within Houdini. Not only does it look better, but it can handle a massive amount of geometry, you can combine solvers to make pretty much any type of simulations interact with each other. Houdini gives the ultimate control and the ultimate set of tools to do literally anything. This amount of control comes with a crazy learning curve, and it can feel very difficult to do simple tasks, and sometimes very simple to do very challenging tasks. The biggest selling point is that everything is procedural. You can go back and tweak settings at the start of your project, without having to start again from scratch. Imagine a client needed a change, this can be very helpful.
Houdini: 2000$+ Blender: Free! I mean, how would one even compare these two? It's like comparing a F1 car with a container ship. First one can only drive in circles on a track, no trunk and only one seat. Second one can haul anything across the ocean - but just in bulk containers. Still, if you would needed it, you could make a aircraft carrier out of it. It will be super slow but gets the job done somehow.
I remember the time i was on trying to get into Houdini, i was trying to make a plate with texture with its own UV in diffusion, but with other texture that is for displacement with its own UVmap for. And i can't find the f-n solution even on forums. Of cause there is the solution, but the f-n fact i cant find it quick made me give up. I know it's a beautiful piece of engineering but... the Blender only took the place in my heart after groundbreaking 2.8 version with its interface changing.
Perfect description of Blender: “It cannot compete with any of the major packages in their specialty areas” Not even with addons A blender Character with Auto rig pro and a faceit rig is good. But Still can’t compete with built-in Maya advanced skeleton characters. And the decrepid blender NLA will thankfully be killed off when “Baclava” arrives with a new animation layer system ..in 2025. And there are no Blender cloth or hair addons that can compete with Maya’s Ncloth or Xgen hair Blender with My Flip Fluids(or RDB labs) addons, still can’t compete with Houdini’s procedural workflows or even Bifrost & Aero in Maya because Blender CORE is took weak to handle the data flows such as cached rig evaluation for real time viewport performance which is also why geometry nodes will hit a hard limit that even high specced hardware can’t overcome.
Bifrost sucks. No point in building any tool if the UX is going to be garbage. See the amount of cool stuff being made using Blenders geometry nodes. Now see how hardly anyone in comparison uses bifrost.
And the same thing can be said for every single software out there. Blenders speciality is modelling, and it does it arguably better than anything out there. Maya or houdini or whatever cant compete with blender when it comes to realtime rendering, stylized and 2d workflows, sculpting, or even traditional modelling workflow, where maya feels so outdated.
@@vladimirtomin8223 Yet you can’t name any major 2D Animated film/TV productions series/movies like “Demon slayer” or “one piece” that are using Blender grease pencil. Grease pencil is only popular inside the Blender communities. And sorry but .. no Blender cannot compete with dedicated 2D animation suites Like TV paint or toon Boom harmony.
@@askeladden450 Yes geonodes does “cool stuff” but can’t produce Feature film quality VFX for fluids, smoke & Pyro. while Bifrost is used in Hollywood film productions for these very effects. And even the 3D motion graphics of Geonodes have yet to surpass the mograph features in my old seat of C4D R26.
Using Houdini and other 3d Software as hobbyist is different from using it professional. They are several factors coming into the game, if you work for money. Houdini is a vfx specialist, fully procedural. In many other departments, it has some really annoying weaknesses, when compared to, for instance, blender. While Blender is some kind of swiss army knife, Houdini fx does a couple of things great, and i love how consistent the software is, logical and consistent. When you next time scroll and click through Blenders cluttered menus and tabs, you will remember the last sentence :-) I guess my point is, professionals go not for Software Names, this vs that, they use tools, the best tools that helps the project succeed in the timeframe asked. When i am able to succeed with my Project easier with Blender, why should i switch to hard mode, when it's not necessary and vice versa. So far, my Blender was never at the same spot in the pipeline as houdini fx.
@@sino1449 as I said, render is very fast and smooth. but as someone working in VFX industry, the quality is not there (yet). Houdini is on another level. Embergen can be used for mobile games and light quick vfx shots but that's it.
Idk but EmberGen also game changer in nowadays. Can you beat Houdini? Yes, but within the next 3-5 years or more, to even beat hoodini really had to have more people in the team at JangaFX to create real-time software... Even LiquiGen is coming to public next month maybe, it could more gam changer for complex liquids simulation only in RTX, without high RAM gigs.
in my opinion Blender's modelling beats other software, i don't think it's fair to say it doesn't beat any other software.... i accept houdini and zbursh are better at what they specialize in though... and use the adobe substance which i also feel no other software matches
Houdini rent/subscription pisses me off, I would have loved to buy software and own it (even if it more expensive than subscription). I did not knew blender had RDBLabs addon, seems really good (for indie stuff)
I love Houdini and its result and process, and I love( Max + tyflow + phoenix) too, but houdini as a steep learning curve, whereas Tyflow and pheonix fd is a lot more straight forward. I feel Tyflow is a lot more faster in its sims than Houdini.
@@sino1449 sorry, reply your question. Pyro does have option to use gpu to boost the sim, there is another gpu solver like Axiom, but if you looking for realtime sim, I think embergen is a good one.
The yearly subscription to Houdini might seem like a lot... but it could be broken down to about $40 per month... if they only would follow that method. I bet they also would get a lot more new paid clients.
I think time has come for me to learn houdini after it's been 2 years i started to learn 3d. It's been couple of days since i started. I've tried it to learn when i was beginner but it has very high learning curve. This software much more suitable for intermadiate and advanced artists. I think the only downside of houdini is modeling. I prefer to model in 3dsMax because it's much easier. However, if you model repetitive models, in that case it can come in handy. Other than that, best 3d package of all time.
Embergem and liquidgem will replace Houdinis pyro and liquid simulation in the future. Time is money and there isn’t a lot of time left these days to watch paint dry.
The thing is if you have the knowledge it all translates and can be done in blender. Though it's more a shortcut. I think your a bit biased towards payed software, their biggest competitor is Blender. Why cause it's opensource c++ and python modules. Has a huge user base that keep polishing it to their demands, there is no commercial software as well designed as blender, to me and i use it commercial at work only if we cannot do it in blender we hire but we hire less and less and instead expand in the blender region started as Indy and now becoming large. With our policy we keep silent not even discussing our company we work with a set of contracts that we decide with who we want to expand. Gives us more artistic freedom as were a group of artist...
It is painful to watch complex software interfaces made for non-programmers. "It is procedural [....] and _node-based_". No, it's just a dang diagram because people are reluctant to learn a few programming instructions. It is just imperative programming as everything else (speaking of the user interface). When it comes to the numerical side of things on the other hand, which I am an expert in, we have to acknowledge that Houdini could not possibly do realistic simulations, as this involves, e.g. for a simple turbulent layer, several core millennia, hence why super computers even exist.
ANSYS & StarCCM+ are Engineering simulation package, quite different compared to the simulations for CG. Also keep harping on Bifrost. it's supposed superiority was not enough to gain much market share. No one is using it.
@@UlfricStromcloak Polyphony Digital uses Bifrost to create all of the procedural buildings in Gran Turismo, but sure, "nobody is using it." Do you realize how ridiculous you sound? When Bifrost's predecessor Naiad was created back in 2008-2009 for the water simulation effects in James Cameron's avatar, they didn't use Houdini as far as I can tell. So, even a decade and a half ago, Bifrost's predecessor was good enough for the most premiere blockbuster of 2009, yet its successor Bifrost in 2024 isn't good enough in the right hands? Make up your minds, at least I don't need to evangelize Houdini or Blender when there is clearly competition out there that's better in key specific areas and doesn't need programmatic skill to utilize effectively.
@@NUCLEARARMAMENT Have you ever heard of the concept of hyperbole. Yes, there are people and studios using it- does it matter in the grand scheme of things? No, Bifrost has no real meaningful market-share, Houdini is dominating. The ratio of both is 97 to 3. I am being very nice here, I pulled these numbers out of my arse (the truth is probably much worse). Also Bifrost is NOT Naiad, doesn't matter if it was technically its predecessor, they are not comparable. Autodesk fucked up (to nobodies surprise), crippled Naiad while not providing a proper replacement, development went at snails pace for years. All while Houdini was trailblazing fast and consistently. Naiad was great back then, Bifrost is still meh. I've tested it myself (superficially I admit). What stood out was the lack of performance while showing an incredible high demand for hardware resources. Houdini simply runs circles around Bifrost - effortlessly. Lack of educational resources doesn't really help the adoption rate. There is a reason for why people and studios avoid it and it has a lot to do with Autodesk's reputation. Slow, too little, too late.
lol all of this going to get thrown into the bin 🚮 once ai video becomes mainstream. sora can already generate photorealistic visuals. imagine a year later how good these models will get.
@@businessmanager7670 You AI-bros don't really think that through to its logical conclusion. If AI-software will eventually outperform and destroy these traditional tools (which isn't guaranteed), what do you think will happen? Trained and experienced artists will use these AI tools and they will outperform and destroy non-artists. Creative people with experience in creating art themself will ALWAYS create better results than people that can't express themselves artistically and are dependent on an external tool to create something. If the tools are making everyone equal the one with more knowledge and experience will have an advantage over those who don't. Always has been the case, always will be.
@@RyoMassaki my point of refutation here is, ai will replace these tools. that's all i said. you said a bunch of stuff which is not relevant to what i said.
At the current point, I am 100% certain Blender can't match Houdini in simulations. However, the fact we're even considering it, is a massive leap from the early days. A film studio could realistically use Blender now. Maybe in 10 years, Blender is the Wikipedia to Houdini's Encarta.
Film studios are already using Blender now. It's a great tool for the beginning of the pipeline (concept design + asset production). It totally sucks for anything that is at the end of the pipeline and it'll take another 5-10 years for it to reach the place where others are now and that's why it will never dominate the end of the pipeline like the competition does.
Houdini especially is a fast moving target that is trailblazing hard and it already has an considerable advantage that doesn't get smaller, it gets bigger with every release.
@@RyoMassaki the problem with your assessment, is that nobody really believed Blender would come so far, not even the creators. So it's perfectly possible for it to be at feature parity with future Houdini in 10 years. It's a long time.
@@4.0.4 It's "perfectly" possible if SideFX would stop developing H right now for whatever reasons and if the Blender Foundation suddenly hired the best developers in the world and throws Millions of dollar at the problem.
So in short, the chances are close to ZERO.
@@RyoMassaki don't get me wrong, I don't think the chances are high; I just wouldn't discount it entirely. The whole simulation workflow would need to be reinvented from scratch.
The chances with ZBrush are a lot better I think.
@@4.0.4 I don't want to come off as too hostile or dismissive, I actually use and love Blender, but I have come to really detest the parts of the Blender community that dream of supremacy over all and other ego-driven nonsense. Which is completely delusional and based on (willful) ignorance.
Blender will become better and it already has a solid place in various industries.
But when it comes to simulations I am not very optimistic, especially the more experience I gain with Houdini. Even with my knowledge (and pessimism), the amount of complexity that is between those 2 seems to me much bigger than assumed.
I don't think the BF has the necessary resources, and even if they had, I think they would be better spend elsewhere (animation for example).
I tell people it makes really complicated tasks easy and really easy tasks complicated.
Yes, intuitively, that's very easy to imagine. It's one of those programs I would be very happy to let other people work with.
@@richardconway6425are you my boss?
@@abdou.the.heretic are you my fortnite?
That's a pretty good summary. Workflow is great, until you are buried in trying to edit one specific feature, shuffling the deck, like some 2 bit hustler, lost in a casino…
Houdini is also slowly taking over every single department of the production. It used to be FX / Procedural modeling only, now it's FX, Groom, CFX, Lookdev, Assembly and Rendering in so many places, including ILM / Weta / Framestore (in varying amounts depending on location). Give or take 10 years and there won't be anything but Houdini + Standard modeling DCC + Zbrush.
Autodesk had it coming.
Totally agreed. Sure it used to be an Fx tool, but so many big and medium size studios are switching more and more to Houdini, give it a few years and maya will be entirely replaced . Can't say autodesk did much to avoid this from happening...
Maya is coming with MAYA ASSIST artificial intelligence@@3dwithanaccent695
its slowly going to get Thrown out by ai video lol. houdini will survive for another 1 to 2 years before ai video crushes it and can do everything
@@businessmanager7670I see someone has fallen for the marketing.
Blender and Houdini seem to be the only 3D programs whose developers seem to actually care about, much respect to them
Just FYI, C++ is not a scripting language. It's a compiled systems programming language.
There's a reason that game engines and stuff like the core of Houdini are written in C++ and C#, not in Python. They are very different beasts.
Python is a waste of bits. Assembly is 100% the opposite but only a god can program on it.
@@saricubra2867 I don't like Python and I think it makes Javascript look like a well-designed masterpiece, but it's used a lot as the glue between C++ libraries.
@@saricubra2867gods and the guy who made rollercoaster tycoon
I was at uni with a guy who could read assembly and tell you what it did. He refused to program in it because he said compiled languages were more efficient.
@@KeithNeilson Its not as hard to read as you think, it can take a bit of effort and rereading to figure out more complicated parts of it, but the instructions themselves are easy to read and it just takes time. He is right that it is much more efficent to program in compiled languages, if you need the speed and control use C or C++. That said, it can be fun to program in for small personal projects (although some people think it is masochism). you can also use it for small parts of larger projects that need to be super fast.
not only all of the other softwares not even trying to catch Houdini in simulation.. but also Houdini slowly taking over there specialized area ( Copernicus for example vs substance) ... so as someone said " you guys make stuff quicker than i manage to learn them".
Its not taking over just because they introduced alternative to Substance Designer.
@@hound_of_justice it was just a simple example for those who didn't tried it
Great video! You’re absolutely right! Houdini is King of VFX simulations. Now when you combine Houdini, Blender, and render a you get a powerhouse system that is an amazing. I believe that Houdini when combined with other tools makes production pipelines better, faster, stronger, more flexible, and more robust.
the new MPM solver looks absolutely incredible i hope we have it in hand soon
Yep, I totally agree with you: Houdini is the master of simulations!! The beginning with it can be difficult (and sometimes frustrating) 🤯🤯, but when you start to understand the logic, it becomes so powerful!! 😅🤩 I already worked with some other softwares during my career (even Blender), but I never regretted my choice to work in totality with Houdini! 😆😆😉
Thanks! I love those helicopter view videos about the industry from you!
Node based workflows are logically and mathematically superior to workflows that are not node based, it really is this simple. Also the reason why Nuke pawns After Effects
Its not that simple. Node based workflow isnt better in all cases and After Effects excels at motion graphics more than at composition.
It's not that simple at all. Linear workflows like organic modelling, or keyframed animation, are *exceptionally* badly suited to node-based workflows.
but blender has geometry nodes now, why is houdini still better?
@@mylesprospero8105 well, geo nodes is a very rudementry node system compared to Houdini and really only for stacking topology ontop of other topology.
Most of all it is not using any simulation of forces like gravity, smoke or liquids. The whole physics simulation part is not node-base so far (Blender 4.2). It is super un-intuitive and tedious to even get a simple water simulation setup. Without additional addons it is close to unusable.
But the biggest problem is it's performance and stability. The fact that Blender just crashes without any warning or chance to save your work, when it overreaches your VRAM, is what makes the difference between industry standard and a free swiss army knife.
But who knows what the future brings. The best thing about Blender's developement is, that it is super open to community input and change. I don't know any software that has that much room for customizability. And it's free.
In the real-time simulation field, EmberGen is emerging as a strong challenger. It's easier to use.The Maya Bifrost has long been abandoned.
3dsMax + plugin like Tyflow/TP/Rayfire/Phoenix/Ornatrix/etc is a high end solution too, and it's more artist-friendly in the traditional sense for those that don't like working with nodes. Plus Max also has Bifrost now. I'm not saying it's better, but it's an option that's up there with Houdini.
How well do these separate plugins connect to each other? Can you combine them in a meaningful and efficient way? One of the reasons Houdini is so efficient - you can combine everything with everything and drive anything with anything while keeping all of it procedural and non-destructive.
Nah
Price of all this vs houdini?……
I am 100% convinced in specific areas there are better options than houdini. Tyflow multifracture is better and faster than houdini booleans (by 100x). Rigid bodies is way faster in tyflow. Rigid bodies voxel based has no rival in thinking particles. If you add phoenix+tyflow+ max its still cheaper than houdini. But houdini power is data structure and unification, despite slower and less artist friendly tools.
@@andvfx269$ per year for houdini indie which has everything,
…..max 235$ a month….then add phoenix and tyflow
I would argue that JangaFX may be a potential future rival for Houdini
Yeah, their software's such as EmberGen, GeoGen and LiquiGen don't get absolute most realistic simulations out there, but the mere fact that those softwares are all real time, and with that level of quality for simulation. Definitely puts it up there as a potential rival
Besides FX houdini does other stuff, which still blows my mind to this date after 8 years of using it.
If only it wasn't on a yearly subscription basis...
You can buy perpetual license with yearly maintenance
@@changleon7441 Yea, I'm too poor for that. lmao
I have to agree with what the video said at the end, sure some would say X and Y can compete. I'm sorry but you can't genuinely try Houdini and still thinking its not unrivalled that after it
great overview, been slowly learning the software and this was great insight into its capabilities
MPM solver was actually used by Bifrost in Maya.
Houdini is certainly a very good piece of software, but it's probably not suitable for all profiles of 3d artist. It's software that will suit the most technical profiles.
Growing up thru the 90's, It really felt the future of gaming was going to be awesome. Gaming wasn't just hyper focused on graphics. It was focused on physics effects, rag doll effects, an assortment of art styles, having a range of gameplay modes, creative and fun gameplay mechanics, satisfying damage effects, in depth environmental interaction, environmental damage effects, engaging CPU AI to play against in single player modes, CPU bots in offline modes. Current video games focus on graphical nitpicking. It seems they forgot how much players enjoyed physics effects & other aspects that make games fun. Quirky graphics can work, if a creative art style is used. Add a selection of interesting gameplay modes, the list goes on.. I seriously don't know why gaming had to fall into this rut that we are in..?
Also:
{We have seen that even if companies delay games, it isn't improving the quality of games being released. Which is a sign there is a MUCH DEEPER issue going on with modern games. No excuses, devs used to make great games with much less and they would have given anything to have certain capabilities to utilize that are available today. It just takes actual passion, focus & talent to make great games. Sidenote: It's so crazy how many people are still supporting super mega rich companies putting out unacceptable products.. i don't understand certain people. I don't understand how certain companies can be messing up so badly. It's frustrating that other passionate people are not allowed to make a version of a type of game that we used to have but no longer get anymore.. Yet these huge companies get full control over if an unused game IP gets made & they do everything they can so no other versions of a game idea will ever come out. Even if there's a huge fan base around certain ideas and there's a ton of passionate devs out there dying to make a version of a certain video game idea.. that frustrates me beyond belief}
My promotion hinges on me mastering Houdini.. Wish me luck
Mastering? See you in 5 years :P.
I have a feeling that software like Houdini are going to become truly unlocked through AI -not as a form of generative substitution to human ingenuity, but rather as a language interface.
The node based system only works if you understand exactly how the different nodes interact with each other, the limitations, the possibilities, etc. This makes a very steep learning curve, and it also makes simple tasks very complicated (tho complicated tasks easier, as said by another commenter).
If AI is introduced at the language interface level, instead of working with nodes directly, you could instead communicate in regular speech with an AI model that then builds the node system to your specifications. This can allow you to generate very complex simulations in a very short amount of time because the tedious labor is taken care of by the AI following your instructions -and at any time if you wish to fine-tune and tweak a specific aspect of the node system the AI has built, you can simply go into manual override and change the things you want, after the AI has built the main blueprint for it.
When we think of how AI can assist in workflow, this is the kind of stuff I'm most excited for. Not the generic generative social media type stuff everyone gets mad about. Not the stuff that replaces jobs. Rather, the stuff that creates jobs, by minimizing the tedious aspects of things like simulation physics -removes a large part of the learning curve, adapts the software to our benefit rather than us having to spend a ton of effort to work on the software's terms.
The famous saying of "most people don't want to understand computer language. They want computers to understand human language", this is where AI truly has its main benefits in this industry.
This is so true.
I feel you're the only one in these comments that really sees how powerful simulations could become with the help of AI
@@Tayosteph It's probably just a matter of time. We already see in raw code development how AI is used to create the foundation for many systems to help developers speed up the tediousness of manually writing everything word by word, checking errors, etc. Node systems in 3D and simulation software isn't far from that in how we interact with it.
Glad to see you (and hopefully soon others too) understand this aspect and see the value in it.
I've used Modo, Blender, Maya, ZBrush, and now Houdini. Even from a modeling perspective, Houdini gives you the most control and is superior to the other packages, IMO. Once you understand how to use attributes and occasionally a little bit of VEX, alongside it's procedural workflow, no other modeling DCC comes close. It also comes with HoudiniEngine, which is invaluable in game dev pipelines and last I checked none of the other software have a bridging application as advanced as HoudiniEngine.
For procedural modeling, for general modeling all of the DCC blow Houdini out of the water. It just isnt made to be the DCC replacer.
@@hound_of_justice Not true, maybe for viewport interactions but that's about it.
Houdini can do everything those other programs can for general-purpose modeling. It's not like you are forced to be completely procedural 24/7.
Even if Houdini doesn't come with a native tool that another DCC has, you can create it with ease. Just drop nodes into a subnet and create an HDA. It doesn't get easier than that.
@@caden6305 You are wrong. Houdini is a horror to model with for most stuff with exceptions like procedural modeling. There are reasons why pretty much nobody uses it as a main modeling tool and replacer of DCC like 3ds Max, Maya and Blender. It should focus on stuff its actually at least devent and good at instead of fighting a impossible fight with other DCC
@@hound_of_justice It isn't a "horror" to model with though...lol, maybe if you are a beginner in 3D I'd agree. Houdini should not be someone's first 3D program.
I've clocked in 100+ hours on each of those previously mentioned DCC's and like I said, Houdini gives you the most control over constructing meshes out of all of those programs.
@@caden6305 Its not just me, its also the industries professionals. 3ds Max, Maya and Blender for example are much better for general 3D modeling than Houdini. The control that Houdini offers is a) unnecessary for the most part in this case and b) control comes with a huge caveat to it and that is learning curve and time consuming workaround for what? Its...just really not the best software to model with in general.
I hallucinated, saw the explosions and misread it as Houthis.
I don’t get it , is Houidini this good ? Or is it just needing a lot of knowledge and being an artist god to master it therefore it looks good ? I really want to understand this from people using Houdini cause I’d like to invest time learning it ( I come from 3ds max-UE-blender and substance background) I only watched some Vex videos and would like to know if learning Houdini is worth it if you know those previous package ..
@@zatlanibrahim5438 Other simulation packages don't require you to learn how to program code line by line, such as the plugins mentioned in this video for Maya and 3ds Max.
Chaos Pheonix FD
Autodesk Bifrost
Thinking Particles
Realflow
Turbulence FD
ZivaVFX
...and many more.
There are separate applications like Embergen and Liquigen and Realflow that are offered, amongst many others. You can get fantastic results without ever opening Houdini and you don't need to have a programmer mindset to use them either.
Houdini really only makes sense if you need a project with totally nondestructive workflow as that allows you to independently change things without negatively impacting other things up and down the chain of a desired outcome. Technical artists that actually need to change thousands of primitives in a scene by changing one variable, would definitely be less of a hassle in Houdini, compared to using other software that isn't as non-destructive in workflow; for anything else where that isn't required, the other software suffices and can produce very similar or better quality results in similar or less timeframe without any programmer or technical know-how, which is what we should be pushing in the industry anyway. Not everyone wants to program, nor is it necessary, outside of those who are making the apps. The artist just wants to make stuff at the end of the day.
It's good but the lack of guide on the internet is bad, the hardest part of Houdini is basically figuring out a problem that you can't google search, which happens quite often for me, Houdini community is quite small compared to other 3d programs so if you have a problem you have to solve it on your own or paid expensive course....
You do genuinely get better results out of the box in most areas within Houdini.
Not only does it look better, but it can handle a massive amount of geometry, you can combine solvers to make pretty much any type of simulations interact with each other.
Houdini gives the ultimate control and the ultimate set of tools to do literally anything. This amount of control comes with a crazy learning curve, and it can feel very difficult to do simple tasks, and sometimes very simple to do very challenging tasks.
The biggest selling point is that everything is procedural. You can go back and tweak settings at the start of your project, without having to start again from scratch. Imagine a client needed a change, this can be very helpful.
@@zurasaur why do I see ads in comments?
@@zedeon6299 not sure pal
Houdini: 2000$+
Blender: Free!
I mean, how would one even compare these two? It's like comparing a F1 car with a container ship. First one can only drive in circles on a track, no trunk and only one seat. Second one can haul anything across the ocean - but just in bulk containers. Still, if you would needed it, you could make a aircraft carrier out of it. It will be super slow but gets the job done somehow.
Houdini is the best software out there.
for procedural workflow and high complex VFX.
@@hound_of_justice will be for everything in the future
I remember the time i was on trying to get into Houdini, i was trying to make a plate with texture with its own UV in diffusion, but with other texture that is for displacement with its own UVmap for. And i can't find the f-n solution even on forums. Of cause there is the solution, but the f-n fact i cant find it quick made me give up.
I know it's a beautiful piece of engineering but... the Blender only took the place in my heart after groundbreaking 2.8 version with its interface changing.
Try Bing next time
Imagine being so influential that the most famous and prestigious VFX software is named after you
what about embergen? not as wide used as houdini but its pretty good for real time VFX
Perfect description of Blender:
“It cannot compete with any of the major packages
in their specialty areas”
Not even with addons
A blender Character with Auto rig pro and a faceit rig
is good.
But Still can’t compete with built-in Maya advanced skeleton characters.
And the decrepid blender NLA will thankfully be killed off when “Baclava” arrives with a new animation layer system
..in 2025.
And there are no Blender cloth or hair addons that can compete with Maya’s Ncloth or Xgen hair
Blender with My Flip Fluids(or RDB labs) addons, still can’t compete with Houdini’s procedural workflows or even Bifrost & Aero in Maya because Blender CORE is took weak to handle the data flows such as cached rig evaluation for real time viewport performance
which is also why geometry nodes will hit a hard limit that even high specced hardware can’t overcome.
Bifrost sucks. No point in building any tool if the UX is going to be garbage. See the amount of cool stuff being made using Blenders geometry nodes. Now see how hardly anyone in comparison uses bifrost.
And the same thing can be said for every single software out there. Blenders speciality is modelling, and it does it arguably better than anything out there. Maya or houdini or whatever cant compete with blender when it comes to realtime rendering, stylized and 2d workflows, sculpting, or even traditional modelling workflow, where maya feels so outdated.
Also Blender can beat most 2D frame by frame animation software.
@@vladimirtomin8223 Yet you can’t name any major 2D Animated film/TV productions series/movies like “Demon slayer” or “one piece” that are using Blender grease pencil.
Grease pencil is only popular inside the Blender communities.
And sorry but .. no Blender cannot compete with dedicated 2D animation suites Like TV paint or toon Boom harmony.
@@askeladden450 Yes geonodes does “cool stuff” but can’t produce Feature film quality VFX for fluids, smoke & Pyro.
while Bifrost is used in
Hollywood film productions for these very effects.
And even the 3D motion graphics of Geonodes have yet to surpass the mograph features in my old seat of C4D R26.
what about embergen and liquigen? heard they are starting to get pretty close to houdini's levels of detail
I'll enjoy the results! LOL. Looks pretty complex for a casual hobbyist.
Using Houdini and other 3d Software as hobbyist is different from using it professional. They are several factors coming into the game, if you work for money. Houdini is a vfx specialist, fully procedural. In many other departments, it has some really annoying weaknesses, when compared to, for instance, blender. While Blender is some kind of swiss army knife, Houdini fx does a couple of things great, and i love how consistent the software is, logical and consistent. When you next time scroll and click through Blenders cluttered menus and tabs, you will remember the last sentence :-)
I guess my point is, professionals go not for Software Names, this vs that, they use tools, the best tools that helps the project succeed in the timeframe asked. When i am able to succeed with my Project easier with Blender, why should i switch to hard mode, when it's not necessary and vice versa. So far, my Blender was never at the same spot in the pipeline as houdini fx.
..and by far the best gui
Would love to try houdini but think I'll stick to embergen for now.
Ohhh so Houdini is also probably one of the engines Blizzard uses for their crazy WoW cinematics
Houdini is not an engine.
Since i was a child i always wanned to play with those destruction examples showed here xd
I thought embergen was a cutting edge simulator
Yes, and then coming LiquiGen soon!
It cuts way to many edges. That's why VFX industry is using Houdini.
Do they use bullet physics?
Hi, please compare between Houdini vs embergen
embergen is fast but nowhere near houdinis pyro if you want serious quality.
@@ibotpl is it smooth or heavy renderer,
@@sino1449 as I said, render is very fast and smooth. but as someone working in VFX industry, the quality is not there (yet). Houdini is on another level. Embergen can be used for mobile games and light quick vfx shots but that's it.
@@ibotpl thank you bro
@@ibotpl What do you think of realflow vs Houdini's?
Now add this to games in real time, and you will revolutionise gaming industry.
Idk but EmberGen also game changer in nowadays. Can you beat Houdini? Yes, but within the next 3-5 years or more, to even beat hoodini really had to have more people in the team at JangaFX to create real-time software... Even LiquiGen is coming to public next month maybe, it could more gam changer for complex liquids simulation only in RTX, without high RAM gigs.
Well let's hope houdini will never improve in that time period.
@@Capumaraca Yes!
in my opinion Blender's modelling beats other software, i don't think it's fair to say it doesn't beat any other software.... i accept houdini and zbursh are better at what they specialize in though... and use the adobe substance which i also feel no other software matches
Spectacular...
Who knew Christopher Walken was a VFX geek?
Houdini rent/subscription pisses me off, I would have loved to buy software and own it (even if it more expensive than subscription).
I did not knew blender had RDBLabs addon, seems really good (for indie stuff)
they do have a perpetual license for houdinifx (if you want to own it), it cost $6,995. but I would rather pay the $270 per year for indie.
you can buy it, but you won't get updates, daily bug fixes. the bigger license you get technical support when you have a question for a fx shot.
Unreal engine has been getting interesting lately...
5:11 you say dynamic operator but that could mean a lot of things. Hard to make it to 10:14 I guess lol
I love Houdini and its result and process, and I love( Max + tyflow + phoenix) too, but houdini as a steep learning curve, whereas Tyflow and pheonix fd is a lot more straight forward. I feel Tyflow is a lot more faster in its sims than Houdini.
Is Houdini real time???
😂😂😂Jeez
in viewport can be realtime. yes. LOL
@@xyzDist79 is it fast as embergen?
@@sino1449 sorry, reply your question. Pyro does have option to use gpu to boost the sim, there is another gpu solver like Axiom, but if you looking for realtime sim, I think embergen is a good one.
@@sino1449 embergen is about realtime. houdini is about flexibility, quality and amount of data. they are built and optimized for different things
The yearly subscription to Houdini might seem like a lot... but it could be broken down to about $40 per month... if they only would follow that method. I bet they also would get a lot more new paid clients.
Houdini, your rival is here, Unreal Engine 5😊
UE5 Clan reporting in!
You do know Epic is an investor for SideFX right? 😅
Unreal is still a gaming/ender engine right now.
Unity particle system
I think time has come for me to learn houdini after it's been 2 years i started to learn 3d. It's been couple of days since i started. I've tried it to learn when i was beginner but it has very high learning curve. This software much more suitable for intermadiate and advanced artists. I think the only downside of houdini is modeling. I prefer to model in 3dsMax because it's much easier. However, if you model repetitive models, in that case it can come in handy. Other than that, best 3d package of all time.
0:07 that didn’t look that great IMO, car looked like it was made of butter
what is wating u if want to learn houdini sop dop rop pdg solaris kinefx apex vex apex script python and houdini expressions... did i forgot something
yes: vellum
So THATS what dua lipa was talking about.. cleverr
Houdini is more capable than blender for modeling and almost any thing , blender is 100 year light behind Houdini
4:43 is how R6 Siege should have been
Embergem and liquidgem will replace Houdinis pyro and liquid simulation in the future. Time is money and there isn’t a lot of time left these days to watch paint dry.
Can you show any production ready liquidgen setup?
Enserio? jaj, JAngaFX entro al grupo.
It cooked my 12700k
"side a fax" and "prison" either these companies have very weird names, or you have a very weird accent
The thing is if you have the knowledge it all translates and can be done in blender. Though it's more a shortcut. I think your a bit biased towards payed software, their biggest competitor is Blender. Why cause it's opensource c++ and python modules. Has a huge user base that keep polishing it to their demands, there is no commercial software as well designed as blender, to me and i use it commercial at work only if we cannot do it in blender we hire but we hire less and less and instead expand in the blender region started as Indy and now becoming large. With our policy we keep silent not even discussing our company we work with a set of contracts that we decide with who we want to expand. Gives us more artistic freedom as were a group of artist...
Nodes. I fucking hate nodes.
Utinni
lets be honest you only came here to see cool simulations
Master vfx first then go master modelling and rendering
But blender in my opinion
I only like to watch content about Blemder. Please make videos about Blemder because it is FREEEEEEE
Yeah but the interface itself is ASS so i dont care.
Blender might take over in every aspect in the future
Houdini is the Unreal Engine of simulation... Wait
ue5, blender, maya… all these are better than houdini
SO UNFAIR
however there is blender
It is painful to watch complex software interfaces made for non-programmers. "It is procedural [....] and _node-based_". No, it's just a dang diagram because people are reluctant to learn a few programming instructions. It is just imperative programming as everything else (speaking of the user interface). When it comes to the numerical side of things on the other hand, which I am an expert in, we have to acknowledge that Houdini could not possibly do realistic simulations, as this involves, e.g. for a simple turbulent layer, several core millennia, hence why super computers even exist.
unbearable voice... see a surgeon
Still cant make a good UI
The UI is exactly what it should be for what you need to do and how you need to do it. But tell us what is not good about it?
I'd take a software that is fully functional to my needs than something that looks fancy yet cant do shit as to what I want to project
There is competition. It's called Maya Bifrost. There is also ANSYS and StarCCM+. Keep being evangelists about Houdini's supposed supremacy, though.
ANSYS & StarCCM+ are Engineering simulation package, quite different compared to the simulations for CG.
Also keep harping on Bifrost. it's supposed superiority was not enough to gain much market share. No one is using it.
there's always an autodesk defender
@@UlfricStromcloak Polyphony Digital uses Bifrost to create all of the procedural buildings in Gran Turismo, but sure, "nobody is using it." Do you realize how ridiculous you sound?
When Bifrost's predecessor Naiad was created back in 2008-2009 for the water simulation effects in James Cameron's avatar, they didn't use Houdini as far as I can tell. So, even a decade and a half ago, Bifrost's predecessor was good enough for the most premiere blockbuster of 2009, yet its successor Bifrost in 2024 isn't good enough in the right hands?
Make up your minds, at least I don't need to evangelize Houdini or Blender when there is clearly competition out there that's better in key specific areas and doesn't need programmatic skill to utilize effectively.
@@NUCLEARARMAMENT Have you ever heard of the concept of hyperbole. Yes, there are people and studios using it- does it matter in the grand scheme of things?
No, Bifrost has no real meaningful market-share, Houdini is dominating. The ratio of both is 97 to 3. I am being very nice here, I pulled these numbers out of my arse (the truth is probably much worse).
Also Bifrost is NOT Naiad, doesn't matter if it was technically its predecessor, they are not comparable. Autodesk fucked up (to nobodies surprise), crippled Naiad while not providing a proper replacement, development went at snails pace for years. All while Houdini was trailblazing fast and consistently.
Naiad was great back then, Bifrost is still meh.
I've tested it myself (superficially I admit). What stood out was the lack of performance while showing an incredible high demand for hardware resources.
Houdini simply runs circles around Bifrost - effortlessly. Lack of educational resources doesn't really help the adoption rate.
There is a reason for why people and studios avoid it and it has a lot to do with Autodesk's reputation. Slow, too little, too late.
@@NUCLEARARMAMENTbitfrost a competition to Houdini in 2024? 😂
lol all of this going to get thrown into the bin 🚮 once ai video becomes mainstream.
sora can already generate photorealistic visuals.
imagine a year later how good these models will get.
Houdini already control and run sora, stable diffusion and etc.
@@restwiththetablet how do you know? what's your source to support that claim
@@businessmanager7670 because i work with sora and stable diffusion inside houdini with PDG context. for automation big group contents
@@businessmanager7670 You AI-bros don't really think that through to its logical conclusion. If AI-software will eventually outperform and destroy these traditional tools (which isn't guaranteed), what do you think will happen?
Trained and experienced artists will use these AI tools and they will outperform and destroy non-artists.
Creative people with experience in creating art themself will ALWAYS create better results than people that can't express themselves artistically and are dependent on an external tool to create something.
If the tools are making everyone equal the one with more knowledge and experience will have an advantage over those who don't. Always has been the case, always will be.
@@RyoMassaki my point of refutation here is, ai will replace these tools.
that's all i said. you said a bunch of stuff which is not relevant to what i said.