Alarming Trends I see with Houdini marketing & development

   1199   15   2
User Avatar
Member
4 posts
Joined: Dec. 2019
Offline
Recently I can't shake the feeling that the team's focus, is all over the place.. Making development stagnate & things increasingly confusing to relatively new users like myself.
Especially with the development and marketing/ social media teams feeling completely out of sync.

I get the strong impression that increasing amounts of time is spent on what I would label as obsolete projects.
From questionable at best, learning resources to, the development of Mantra & Karma. Does Houdini really need its own integrated alternative to established industry standard engines like RenderMan or Redshift?

I would love to see more go toward making the creation of feature film quality effects easier & faster.
Things like making good-looking vegetation, especially trees, or proper rain -- are still a major pain & somewhat difficult to get into from a new user's perspective.
Forcing me to almost across the board spend major amounts of my time finding the right workflows, creating my own tools, or turning to 3rd party tools altogether.
I'd be more than happy to pay more & see improvement on this.

The learning resources and social media in general published on the channels feel especially all over the place. Many of which I would describe as down right horrible content.
My main issues are:

  • The workflows detailed by most tutorials often seem suboptimal at best. In the professional realm, from what I can tell, using USD with RenderMan or any other renderer seems to be the undisputed "new" meta. With resources like Kit Bash, Substance Source and Quixel seeing majorly increased use.
  • Most tutorials published or otherwise advertised, actually create a very unprofessional looking end result. This makes the entire tutorial not only unappealing, it also doesn't make the detailed workflow feel viable whatsoever.

This is making Houdini's learning curve, much more steep than it needs to be.
In general, I would love to see quality over quantity when it comes to learning resources of any sort.
With the focus being detailing viable professional workflows creating state of the art visual results.
It doesn't even have to be a professionally recorded tutorial. Even a page like the technical documentation, but properly outlining workflows & approaches, would be a tremendous help already.

Thanks for reading, I hope this gives you some insight into the feelings and thoughts of a relatively new user in your community.
Edited by Yannik - July 21, 2021 08:09:36
Midphase
Member
752 posts
Joined: Jan. 2018
Offline
Someone please pass the popcorn!
>>Kays
For my Houdini tutorials and more visit:
https://www.youtube.com/c/RightBrainedTutorials [www.youtube.com]
User Avatar
Member
946 posts
Joined: March 2009
Offline
Yannik
but also is more often than not blatantly bad content.

Can you post an example URL of one that you find bad?
Martin Winkler
money man at Alarmstart Germany
User Avatar
Member
224 posts
Joined: April 2017
Offline
I would love to see more go toward making the creation of feature film quality effects easier and faster. Things like making good-looking vegetation, especially trees, or proper rain is still a major pain.

The Labs team has been working on tools for just this purpose-- take a look at the various branch and leaf tools they've constructed. They might not be perfect for the exact use case you have, but that's why these tools are all open to be modified by anyone.

proper rain is still a major pain.

"Rain" is a bit hard to quantify from a VFX standpoint. Depending on where you're looking, this could manifest as a simple particle simulation, a POP Fluid simulation for crown splashes, volumetric effects for sheeting, or some kind of custom gradient descent solver for rivulets and clinging particles. There's a lot of potential effects there that wouldn't be solved as a single unit, so it's hard to just build a prefab "rain tool".

For example, why mantra and karma are still being developed is a marble to me.

Mantra is not still being developed. Karma is, but it's still a bit young for production work, so yeah, if you want to use USD for now you're better off with engines that are more fully fleshed out. I don't think Mantra or Karma are going to be everyone's first choice for rendering any old scene, but there is definitely a strong case to be made for an engine that's as tightly coupled to the rest of the Houdini framework as Mantra currently is, and Karma eventually will be.

I would not even try to market or cater toward non-professional users, as they aren't going to use/ learn Houdini either way.

There are plenty of CG hobbyists out there who are using Houdini, alongside people without a traditional VFX background who are interested in using Houdini for purposes other than pure VFX or game assets (such as scientists using Houdini for simulation or data visualization).

I do agree that there's a bit of a signal-to-noise problem on the tutorials end. Part of the problem may also be that SideFX is a relatively small company (and this is a good thing!) and develops new technologies very rapidly, and so it's naturally harder to always have up-to-date educational content available with the cost of admission. There are some series that would probably better suit your needs: the Masterclasses from SideFX in particular are some of the best deep dives you're going to get into Houdini's various systems. And if you're willing to pay, there's always Entagma, the Applied Houdini series, and many others that are widely recognized as being very well-constructed.
MOPs (Motion Operators for Houdini): http://www.motionoperators.com [www.motionoperators.com]
User Avatar
Member
228 posts
Joined: April 2018
Offline
I hear what you're saying, but I'll offer some counterpoint. I actually really like that they're going in a more, hmm, "prosumer" direction recently. I haven't watched every video they've put out, but the ones I've seen are informative and fun. I like the new presence on ArtStation and I LOVE how responsive the SideFX guys are on social media. Makes the Houdini world feel welcoming and supportive and a good place for artists.

I'm definitely on the same page as you regarding Karma. It would be great, again for the prosumer audience, to have a "free" world class renderer in Houdini to match Maya Indie and Arnold. But... Karma needs some work. It creates great images and has a lot to recommend it, but there are some nagging issues that I hope to see addressed in Houdini 19. If they can't add that final polish in the next major release or two, then I do wonder if it's worth the dev time.

Perhaps redirecting focus towards some of the things you talk about, making common creative tasks easier (vegetation, rain), would be a better path forward. I do find the "sim" part of Houdini still pretty hard to use and hard to get excited about. I know how powerful it is (first in class), but for intermediate level tasks it's much faster just to fire up nCloth in Maya...
Watch Perceptarium on Amazon! https://www.amazon.com/dp/B0846GB3WG/ [www.amazon.com]

http://youtube.com/brianhanke [youtube.com]
http://twitter.com/brianhanke [twitter.com]
https://www.instagram.com/brianrhanke [www.instagram.com]
http://facebook.com/brianrhanke [facebook.com]
User Avatar
Member
734 posts
Joined: Oct. 2008
Offline
It's never been easier to learn Houdini. Not so long ago there were practically no video tutorials. Everybody has their own pet peeves and wish list. I agree that a proper veg system would be useful for a lot of people. But I also want cops updated, and sims to get faster, Karma on gpu, and machine learning nodes.
User Avatar
Member
81 posts
Joined: May 2017
Offline
So you are saying that the development is all over the place, and immediately suggest to make it even more over the place.
I see you are looking for "magic 1 button solution", which is a realm of AE and all that kind of thing or so it seems.

Too many thoughts jumbled together anyway, for example marketing teams do not develop rendering engines. And I personally think that buying more hardware is an interesting alternative to buying a 3rd party render engine (sometimes same money for sameish result).

Long story short, feature film quality effects involve heavy computation (HEAVY computation), lots of testing of variations, yada yada yada. After that more often than not you bake it down into hundreds of gigs of data, then assemble it together, render in passes, then composite it all together, list goes on.
To be fair, H is actively working towards simpler solutions - SOP based dymanics is a good example, as well as some 1-click explosions presets. But what you are asking is kinda like getting ripped without doing the exercise. What can possibly go wrong.
https://twitter.com/oossoonngg [twitter.com]
User Avatar
Member
19 posts
Joined: Aug. 2020
Offline
osong
But what you are asking is kinda like getting ripped without doing the exercise. What can possibly go wrong.
ha ha!

User Avatar
Member
4 posts
Joined: Dec. 2019
Offline
toadstorm
"Rain" is a bit hard to quantify from a VFX standpoint. Depending on where you're looking, this could manifest as a simple particle simulation, a POP Fluid simulation for crown splashes, volumetric effects for sheeting, or some kind of custom gradient descent solver for rivulets and clinging particles. There's a lot of potential effects there that wouldn't be solved as a single unit, so it's hard to just build a prefab "rain tool".

Maybe rain is not the best example. It came to my mind, since I have seen much demand for interactions with rain effects recently. Be it splashes or rain hitting objects, forming droplets and trails running down those objects. While this is definitely possible, figuring out workflow, realistic feeling settings etc. Just feels so much more demanding than I personally feel like it should be. Providing a more easy to use and setup preset or at least some proper resources detailing such workflows for fairly standard effects, I do believe would be extremely beneficial for especially new users.

Mantra is not still being developed. Karma is, but it's still a bit young for production work, so yeah, if you want to use USD for now you're better off with engines that are more fully fleshed out. I don't think Mantra or Karma are going to be everyone's first choice for rendering any old scene, but there is definitely a strong case to be made for an engine that's as tightly coupled to the rest of the Houdini framework as Mantra currently is, and Karma eventually will be.

I did not know Mantra was completely cancelled -- my bad.
I mainly made this point, since I see too little benefit in Houdini competing with well established production ready engines such as RenderMan or Redshift. I doubt professional and even hobbyist users which turn to Houdini, will be looking to use or learn Karma over well established industry standards.


There are plenty of CG hobbyists out there who are using Houdini, alongside people without a traditional VFX background who are interested in using Houdini for purposes other than pure VFX or game assets (such as scientists using Houdini for simulation or data visualization).

I believe, Houdini's selling point ultimately is quality. If users are happy with a medium quality effect, I do believe that they will usually always choose the cheaper, easier to get into alternative.
If anything, I might consider investing into making it easier to achieve high quality results. Supplying better resources and easier to work with tools - effectively flattening the learning curve to achieve state of the art results.
User Avatar
Member
224 posts
Joined: April 2017
Offline
I mainly made this point, since I see too little benefit in Houdini competing with well established production ready engines such as RenderMan or Redshift. I doubt professional and even hobbyist users which turn to Houdini, will be looking to use or learn Karma over well established industry standards.

Smaller studios are definitely not using Mantra much, but the last big studio I worked at used Mantra for all of the heavy effects-driven stuff we did for feature films. Nobody is using Karma at scale yet, but there will still be room for it once it's a bit farther along... the tight integration with the rest of Houdini (CVEX, VOPs, etc) is really important for complex shaders, volumes, heavy scenes and the like.

I believe, Houdini's selling point ultimately is quality. If users are happy with a medium quality effect, I do believe that they will usually always choose the cheaper, easier to get into alternative.

"Quality" is a bit relative. I think really the selling point is "flexibility" in that you can get whatever you want out of the program with few restrictions (other than your own ability). Originally Houdini was intended only as a vehicle for effects work for feature films, but that's expanded dramatically since into games and motion graphics. The relative "quality" of a motion graphics animation versus a feature film environment is a bit tricky to define... the goals (and usually timelines) are very different.

If anything, I might consider investing into making it easier to achieve high quality results. Supplying better resources and easier to work with tools - effectively flattening the learning curve to achieve state of the art results.

This is more or less what I've been trying to do with MOPs; building high-level tools to help flatten the learning curve and solve common animation problems faster. There's other guys out there like qLib, ODtools, and SideFX Labs who are doing the same. The core Houdini developers, in my understanding, have been a bit hesitant to build higher-level "wrapper" tools in the past because flexibility and scalability of the core SOP tools is so important to Houdini as a whole. It's a really tough balance to have tools that allow for maximum flexibility while also making them quick to set up and simple to understand (such as preset rain effects)... the more you try to set up for a user, the more things can go wrong when the user wants to break the rules, and now you get stuck supporting these gigantic lumbering prefabs instead of staying closer to the bare metal.
Edited by toadstorm - July 20, 2021 17:36:23
MOPs (Motion Operators for Houdini): http://www.motionoperators.com [www.motionoperators.com]
User Avatar
Member
4 posts
Joined: Dec. 2019
Offline
osong
I see you are looking for "magic 1 button solution", which is a realm of AE and all that kind of thing, or so it seems.
No one is looking for a one click solution to make a movie..

I'm looking for maybe a solution to get a basic working setup going within like 20clicks instead of often more like 500 required & learning how to write code, so I can start tweaking and developing my scene quicker.

I'm mainly asking for proper resources to tell me where to plant my initial clicks, be it 20 or 500. As currently, even though the technical documentation is rather good. There is little to nothing usable on workflows and approaches. And the stuff there is, often details suboptimal or non industry standard workflows.
User Avatar
Member
4 posts
Joined: Dec. 2019
Offline
BrianHanke
I'm definitely on the same page as you regarding Karma. It would be great, again for the prosumer audience, to have a "free" world class renderer in Houdini to match Maya Indie and Arnold. But... Karma needs some work. It creates great images and has a lot to recommend it, but there are some nagging issues that I hope to see addressed in Houdini 19. If they can't add that final polish in the next major release or two, then I do wonder if it's worth the dev time.

Perhaps redirecting focus towards some of the things you talk about, making common creative tasks easier (vegetation, rain), would be a better path forward. I do find the "sim" part of Houdini still pretty hard to use and hard to get excited about. I know how powerful it is (first in class), but for intermediate level tasks it's much faster just to fire up nCloth in Maya...

While I'm on board with having a free world-class renderer being great to have, I don't think it's realistic or beneficial enough to pursue. Even if they could, against the odds, pull off competing with RenderMan 24 (which looks to be light years ahead).
I don't think it's a good investment of time and resources.
Realistically, I don't see many users really "needing" an integrated free solution.

Especially with so many things being far from perfect, I firmly believe time could and should be better spent.

To me, the missing of up to date resources detailing the workflow & approach of most things makes many things feel hard to get into and excited about.
It doesn't have to be a professionally recorded tutorial. To me, something just like the technical documentation but for workflows & approaches - would be a tremendous help already.
User Avatar
Member
81 posts
Joined: May 2017
Offline
I made my remarks based on your initial post.
Also, why you guys are thinking that working on Karma is eating up resources from developing tools for easier vegetation solutions?
Imagine coming into Kai's office (https://www.sidefx.com/profile/KaiStavginski/) and saying "Yo yo yo, wrap up this Karma thing, you are working on trees now!".



Also, I'd love to have a good render that is built in with my indie (or FX, does not really matter) license. I very much need it, do not take it from me. The fact that you might happen to own Rman or Arnold license does not magically result in no necessity for a render engine. What this idea is even.

Finally, although I personally share your sentiment that things should be more approachable, the thought "feels so much more demanding than I personally feel like it should be" happens to all artists across the globe all the time.
Even such a trivial task as holding a pencil and trying to draw a person sitting on a chair would result in a necessity for practice of anatomy, gesture drawing, and perspective (with 2 or 3 vanishing points).
https://twitter.com/oossoonngg [twitter.com]
User Avatar
Member
111 posts
Joined: July 2005
Offline
Houdini has always been about providing the fundamental building blocks that allow the expert user to build their own custom solutions. I liken this to the programming concept of "simple vs easy" (google it). Houdini is simple not easy, on purpose. Other softwares have gone the "prosumer" route where you have point and click tools for particular sorts of assets, which can be nice and easy and look pretty in a short amount of time, but they can be ultimately limiting in the way you can use them. Don't judge tutorials that are trying to teach a particular concept by the rendered output unless the point of the tutorial is actually trying to achieve that goal. Notice that many of the Entagma tutorials don't include the steps that they took to achieve their final rendering, even though they nearly always include a beautiful rendering in the intro.

For the nerds in the audience, the speaker in this talk makes a similar argument in the field of programming - https://www.infoq.com/presentations/Simple-Made-Easy/ [www.infoq.com]
Edited by drew - July 21, 2021 19:37:33
User Avatar
Member
102 posts
Joined: Feb. 2016
Offline
Now Adobe has moved in on Redshift, Karma is an important part of what SideFX is developing. We have seen what happens to software when Big Corporates get involved. They don't develop the just buy development. Redshift is at future risk of stagnation and price gauging.
User Avatar
Member
19 posts
Joined: Aug. 2020
Offline
Yeah, but apart from all these, what has SideFX ever done for us?
  • Quick Links