eikonoklastes

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:

From the media release: https://aws.amazon.com/blogs/media/aws-thinkbox-products-now-available-free-of-charge/

Yeah, should be an excellent learning resource. Can't wait to sink my teeth into it myself.

From the media release: Promo video: Link to downloads page: https://animallogic.com/alab/

The algorithm, yes. Personally, I haven't a clue what it's doing. Maths and I parted ways a long time ago after she found out I was spending way more time with Pulling Numbers Out of My Ass.

So I finally dived into that algorithm and we do seem to have lift-off houdini_N3ojd2H3fb.mp4 Common Tangents.hiplc

A quick hack to get the outside tangents would be with Convex Hull. Then I just delete the small edges. Have to think about the inside tangents.

They are accompanying videos to the Houdini Foundations PDF document. Perhaps they should have made that a bit more clear. I agree that the density of information is quite high, but I still think he's doing a good job of going over the basics.

Well, you don't animate, so no reason for you to know it. I, however, animate all the time, so...yeah - really should have known that, because it can be pretty helpful for camera animation.

For me it was - Object Display as Subdivision - Ribbon mode on the Pose tool

Or to learn a couple of new things, as I did...

So I've watched it now, and yes - definitely a really nice kickoff point for new users. It only touches the absolute basics, but all very critical stuff. This is the playlist to send to people looking for a gentle entryway into Houdini (we'll kick them in the teeth when they're in and the door is closed behind them...)

SideFX just dropped a playlist of foundational Houdini stuff. I haven't watched it yet, but should be a great primer for new users.

I think (I might be wrong) that C4D just rolls the depth extrude and inset into one step, whereas Houdini separates them. So to do what the OP asked would need 2 PolyExtrude nodes (one for the depth, one for the inset/bevel) in Houdini, but will achieve the same result.

The designers at work threw me an unexpected challenge today. They had me put a bunch of random country names along a spline, with very specific alignment and spacing. The names are being read from a CSV file, and the excellent Chain SOP made the distribution pretty easy. Then they come along and say, "Hmm...how will it look with a line going across the top of the names", like this And it's challenges like this that make me appreciate Houdini all over again. I took a likely unnecessarily convoluted route, but it worked like a charm, and didn't require any VEX or expressions. With the method I used, it works nicely for lower-case text too, ignoring the middle characters. I'd need extra logic to handle sentence case, but our design won't be using it, so I'm not bothering. Scene attached, but I would love to see alternative ways to tackle this. Connect Font Tops.hiplc

Thanks. I used Karma XPU, and was getting 2-3 minutes per frame at 1080p. I have a Ryzen 7 3800X and an RTX 2070 Super, so a pretty lower mid-range spec compared to what is available today. I rendered with 128 samples, which is quite low for a production quality render, so there was definitely noise, but I used the Optix denoiser, that worked pretty damn well, even across frames. It definitely softened out some detail, but I also did another render without any denoising to compare, and honestly, for this project, the loss of detail was totally fine.

So I finally got around to rendering this thing out. Was a fun project to work on.

The ol' SOPs gotcha. I really should have picked that up from your screenshot (it says Object in the corner). Glad it's sorted.