eikonoklastes

The C4D peeps are having a bunch of fun with this, so why not join the party? (Well, looking in from the outside, I guess . . .) If you Google "chladni pattern formula", you can pick up the formula from the first result, and just directly plonk it down into a Point Wrangle. Generate a couple of sliders, and Bob's your uncle. I also added a locator to alter the center point, for more variety. Scene file attached below. houdini_VkTEEjYWSH.mp4 Chladni Patterns.hiplc

It seems more broad, from what I understand from that excerpt. They (effectively) say "you cannot use the app to train any AI system in order to simulate or perform a function that is similar to any function in the app". So it's not just restricting you from generating images, it's restricting you from doing anything that is similar to any function in C4D. Which is a pretty remarkably broad stroke in itself. I wanted to say "ludicrously", but settled for "remarkably".

What are they worried about? That someone will use AI to create 3D tools that match or surpass those in C4D? Isn't that already what's happened, and without AI?

Another fun setup, from a C4D thread (Houdini scene file attached below): houdini_6Y6BwBfK7E.mp4 This scales really well, too. Here it is with 12 thousand arrows, playing in real time in the viewport, on my 7-year old machine. Not sure what that would ever be useful for, but still pretty cool to see in action (the FPS is displayed in the bottom right - green means real time): houdini_pk5I6C4dpk.mp4 This is the C4D thread: Arrow wave modeling and rendering translucent planes - Cinema 4D - Core 4D Community Arrow Waves.hiplc

Hi, I have attached my scene file in the original post, you are welcome to take a look at it. I do use a particle system to emit the circles to get a pulse emission, rather than a random emission. There are other ways to do this, but I just used whatever was the easiest for me to pull off. The rest of the setup is also very easy - the only complicated bit is the loop to get the variation on each copy. While complicated, it's actually easy to execute, it's just done in a rather unique Houdini way.

This is a turbo-charged version of the previous setup, inspired by a recent post in the C4D thread doing the same thing. Here's everything happening in this one: Circles are emitted and scale up in concentric fashion They distort as they grow, but only on the outer circles - the inner circles remain intact. They have a colour gradient from inner to outer They are duplicated and projected onto a sphere Each copy has its own noise variation - specifically each copy has a different noise offset and frequency applied. I can precisely define what these values are for each one. There is a mask (falloff) that drives both scale and colour of these copies. There is a scene file for this. I also just threw a noise on top of the whole thing, for shits and giggles, and added a basic render setup in Karma. houdini_SJEtYGNQHo.mp4 This is the post from the C4D thread: Concentric Rings on a Sphere Scene File.zip

You can add an extra 1 TB for $10 per month. I don't believe that that is limited to any amount, as long as you keep adding $10 for every TB. However, if your size requirements are that large, you'll want to look into their business plan offerings, which work out cheaper at larger capacities, and for more users.

I use OneDrive myself and am quite happy with it. I have the Office family sub, that gives each member 1 TB of storage, which is a pretty good deal. I haven't faced any issues with speed or sharing limits, it has all worked fairly well for me so far.

Here's a souped-up 3-ball version of the above. This isn't included in the scene file above; I'll leave that to you guys to figure out. There's obviously some heavy trickery afoot to keep them all on the plate though... Untitled_H.264.mp4

Cleaning up some files, and came across this fun one that I did a few months ago. Houdini very helpfully allows you to feedback a simulation's output back into that same simulation, which makes this possible. There is a noise force being applied to the ball that pushes it around, while the plate tries to keep it from falling off. There are no keyframes, it's entirely simulated. I have included the scene file below. It was inspired by a similar animation I saw on the Houdini Reddit, but I can't link that right now unfortunately as that subreddit has gone private. houdini_IrqnsPIect.mp4 Self-balancing Plate and Ball v2.hiplc

Taking some inspiration from the C4D forum, and thought this'd be a nice, fun and simple exercise. There's mask control, with a ramp, and you can specify the colouring and distortion based either on the age of the particles, or the mask. I've included the scene file here. This is the C4D thread: Emit Circles and Distort Them as They Move Outwards.hiplc

I'd like to add an extremely important factor here — that only SideFX seems to provide — and that is world-class customer support. We are so used to what Autodesk and Maxon and Adobe call "customer support", that what SideFX do here seems almost like fantasy, but is actually real. For starters, they put out daily builds that contain bug fixes and even new features on occasion. Anyone can download these daily builds (even the free Apprentice users). If you submit a showstopping, high-priority bug report, they will address it with urgency and put out a bug fix at astonishing speed. You can send them your scene file to analyse, and they will actually analyse it and get back to you with a surprisingly quick turnaround. Their customer support staff are well-trained and actually competent, and you will frequently get direct responses from the developers, even for minor reports. The forums are also frequented by developers, and by helpful industry veterans, so you will almost always get a useful, helpful response to even tricky questions. I am not a SideFX shill, and will criticise them where applicable, but I can tell you wholeheartedly - the support you get from them is on an entirely other level from everyone else out there, and that adds a tremendous amount of value to a production.

It's the opposite though, as they have, thus far, been extremely careful to use only their own licensed material for training, and have pledged to compensate contributors to their stock library (that is being used for training their AI models). As with all things Adobe, we'll have to see how that pledge actually plays out, but so far all indicators are that they want to play fair. I suppose we will have to define what "significant human intervention" means. There already exist many apps that can produce compelling imagery at the click of a button (that don't use any AI). This is a much broader topic, and likely exceeds the scope of a single forum post, but the last 40 years of hardware and software progress have made it easier and easier to produce better and better imagery, with less human intervention. Rendering is literally a process whereby you press a button to relinquish human control to a computer to produce the final output. Yes, there has to be initial prep before you press that button, but all that prep is basically translating a human thought into a state for the computer to output. My stance is that it's the same with AI — except that it just makes all that prep work less of a chore. Basically, a tool that removes more technical barriers between an idea and an output. Replacing a tedious, technical process isn't necessarily a bad thing in the pursuit of content creation. There is definitely purity and value in pure mechanical and technical prowess, but in a capitalist world, easier methods will always take general precedence, especially if they can reduce costs. For me, a "significant human intervention" needs only be the creative idea or direction that a human can come up with. Being able to translate that effortlessly onto a certain medium should not detract from that. I'm in the beta, yes, and have played with it a bit. It's promising, but extremely limited in what it can currently do. Each generation is random - any further tweaking will require traditional methods. It also struggles with people, especially non-Caucasians. It's also heavily censored. I couldn't make an "exploding text" effect, because the word "explode" is banned by them, so good luck trying to make a "cool guys don't look at explosions" poster. I think Adobe's approach of adapting Firefly tech into its core apps (like Generative Fill in Photoshop, and the vector recoulouring in Illustrator) is the correct approach. By making it a tool that can do a certain something, and be complemented by the existing toolset, it works a lot better for actual work, rather than being a self-contained quasi-random image generator that's only useful for generating concepts, and falls apart when you get any specific notes/feedback. I don't believe there are any generation limits. It's online only, but I imagine that the final product should be able to work offline, at least optionally. If it works offline (like it should) I wouldn't want to pay anything above the existing CC sub. If it's online, I can imagine it's fair for them to charge an extra amount, I have no idea how that might be structured though. Having said that, I would be miffed if it was online only. There's zero reason for them to do that, except be greedy (and this is Adobe, so I do actually expect that).

I haven't tried it yet, but it seems to have some decent USD support already. This can potentially be huge for doing some advanced post-work. I really need to get my hands on the trial and start poking around.

Another big win in my books, straight out the gate, is it runs on Linux. I know Linux fanboys like to crow about how Linux performs so much better, but I am not a Linux fanboy - I don't really like the OS. But I have tried it out with Houdini, and it absolutely flies compared to Windows. It feels like I upgraded my system hardware, when all I have done is reboot into a different OS. Adobe, and Ae specifically, was the one thing that was keeping me from moving my work to Linux entirely, but with Autograph I can potentially finally do that.

Contrarian opinion here, but I think the pricing is ok. It's not ideal, but it's not terrible either. $60 for a monthly sub is reasonable when even a low day-rate is several times that. I am at a stage with Adobe (and have been for a while), where I will happily pay double the price for a better experience. The thing that irks me most is the time-limited trial. That's not ideal at all for unvetted software, and for busy schedules. I don't have the time to cram a brand-new app's learning into 30 days. They need to remove the 30-day limit and offer an unlimited trial. I don't care if they put a massive watermark over the UI and clamp the output to 640x480, but you need to be able to thoroughly test something like this before jumping in. Houdini is the gold standard for trials. Apprentice is fully-functional and entirely free, including all updates. Just copy them.

That's precisely why people want an alternative to After Effects, because it does pretty much only those things, but with a much worse user experience. The future does seem very bright, with amazing AI tools on the horizon, but in today's world, animators and compositors still need to grind out work keyframe by keyframe, layer by layer, node by node, and have been waiting for a tool that isn't as poor as After Effects currently is. Adobe have slept on their motion graphics monopoly with Ae for around 20 years now, mirroring Cinema 4D's laughable development pace, and very content with 3rd parties filling in the very large gaps for features that should exist in the app itself. Houdini and Blender becoming more accessible over the last couple of years seems to have finally woken the C4D devs up, but no competition exists for Ae. I'm hoping that Autograph is the one that can at least wake Adobe up to make Ae suck less, so that something as basic as exponential easing may finally exist out-of-the-box, in an app that has been a leader for over 30 years. Ae only just added (literally in the latest release) functionality to re-use a track matte without having to duplicate it. It's embarrassing, and they really need a bit of shaking up.

Templates perhaps, but the whole point of wanting a more robust, extendible application is to not have to rely on a slew of 3rd party plugins/scripts, at least for basic work, which has been the bane of Ae forever. An analogy can be drawn with C4D - hugely popular app with a ton of plugins and templates to extend it, but then you have Houdini, and you don't need any of that. Then you factor in the infinite cost of Ae, along with its myriad bugs (in the latest update, H.264 video is a gamble on whether it will actually work in Ae or not, and just show a green screen instead), and things start getting pretty irritating. Regarding the other apps, for compositing and effects, Nuke and Resolve (Fusion) are miles better than After Effects. It's not even a contest. For motion graphics Nuke doesn't enter the conversation. Fusion, while it does have a good foundation for mograph, needs a ton of work to make the UI and UX streamlined enough to actually be able to do any sort of meaningful work without wanting to tear your eyes out. A basic example of this is that you cannot select two nodes of the same type and change the same parameters on them simultaneously. I cannot comment on the other two as I've never used them.

I'm cautiously optimistic. I previously put my faith in Cavalry as the potential replacement, but after many years of an inordinate amount of bugs and almost negligible foundational updates (it's still not able to handle even a few high-res images, and its support for video is very poor), it's clear that that's not the one. Autograph looks significantly more robust and capable, straight off the bat, but we'll just have to wait and see. Pricing will play a big role, as will stability.

Sorry to derail this thread a bit, but I don't see Autograph being comparable to Cavalry. With all due respect to the Cavalry devs, it's a product seemingly aimed at hobbyists looking to do something isolated and cool with its procedural toolset, to post on social media. It is fundamentally not built to handle any sort of complexity in a project. It has a miserable caching system (if you could even call it that), and the whole app will choke with even a handful of high res images on the timeline (by "handful", I literally mean less than 5). I don't believe the current version even supports video footage. What makes things even worse is that the app is unbelievably riddled with bugs from top to bottom. They claim a 1.0+ plus release, but honestly, I have used alpha software with fewer bugs and inconsistencies than Cavalry. I have lost an inordinate amount of time to bugs and reporting bugs in that app, and I ran out of patience with it a while back. The devs seem keen on pushing more big title features than knuckling down on the consistency and robustness of the app, which is a shame. I desperately wanted Cavalry to be a viable alternative to Ae, but unless they revamp their QC process, Adobe has absolutely nothing to worry about from Cavalry. What irritated me more is in their latest release, they have locked the Shortcut Manager behind a paywall - you need the Pro version to customise shortcuts. When I asked the question on their Discord about this decision, a dev stated that they feel that anyone who wishes to use Cavalry enough to want to customise it, will see the value of their pro license over the free version. This would normally be a good approach, but if they think that I am paying to beta test their incredibly buggy app, they are dead wrong. Back on point - Autograph is foundationally built to handle video and high res images, including OpenEXR. I cannot imagine that their app, even at 1.0 will be anywhere nearly as broken as Cavalry, so I'm cautiously optimistic about it.

I've had my eye on Autograph for a while now. It looks and sounds almost too good to be true. We'll have to see how it actually performs, and I'm expecting the price to not be cheap.

It's been that way pretty much since the launch of Solaris, iirc. You probably just didn't have the desired template as part of the same stage. Here it is in H19:

Yes, it's definitely the direction SideFX are taking Houdini. You can do everything inside Solaris via SOP Create. The one reason I'd avoid that workflow, is when dealing with heavy datasets. Since the SOP Create LOP is a wrapper for a /obj > USD pipeline, there should be a delay every time you jump in and out of that LOP if you are converting a massive scene into USD on-the-fly. I assume there is a delay, anyway - I don't know what tricks SideFX might be pulling to mitigate that, and I haven't really tried it myself yet.

Ghosting works in Solaris. You just need everything to be part of the same stage: The main sell of USD is not its interoperability. Alembic already does that very well in production. USD offers much more than interoperability, in the way of non-destructive scene management and inter-departmental scene and asset layering, apart from other nice features built into its core. It is a godsend in production, and no other system even attempts to do what it does. It is not a technology aimed at individuals, so I guess if you have to ask, you very likely don't need it (I certainly don't). This is not true at all. You can definitely animate positions on any light, and here is an example of lights in Solaris instanced onto points: Leaving aside the Solaris-exclusive interactive highlight and shadow placement (that SideFX could easily bring to /obj if they wanted), Solaris offers a trivial way to non-destructively set up and lights, and reuse them across shots. You compare it to takes, and there is some overlap in this use case, but you would need to generate a fair few takes to achieve what Solaris can do with a little branching. The other, less mentioned, advantage is the visual layout of your scene structure, something that takes does not offer at all. Knowing what lights affect what shot - at a glance - is a massive productivity boost. I'd say to each their own, though. If takes works better for you, you should use takes. I vastly prefer setting up lights in Solaris, and that will be my preferred method going forward. It's good to have choice.

I believe this is appropriate here: