Is messy duck taped spaghetti scenes the norm in the VFX industry?

[acey195]

unless the end product IS machinery, for others to work with.... I guess this is maybe not as true for FX,

but if your product requires a lot of iterations, in games for instance, it is certainly worthwhile to optimize stuff and make it readable.

[magneto]

But would having an uber messy scene come back to you negatively if someone else needed to reuse it or wanted to work on it?

 

I saw some scenes with a single geometry node where everything is there and there was a lot of copy pasted sections, not a linear network but one that has connections all over like Spider Man's next masterpiece.

 

I guess artists trust other artists to understand what's up when they scene their scene and not complain about it to others.

[anim]

while messy scenes are very common due to lack of time, tight deadlines and unclear idea about proper solution, you can always try to keep things organized, I bet no one leaves messy scene on purpose

 

- the creative process itself is messy and it's not important to keep your network clean when you are trying to find a best solution for your problem, but after you find it, it's worth to clean it up even if you are not going to wrap it as an asset

 

- if you can find some time to optimize and clean up your scene from time to time it can be a huge timesaver later, when you need to do some changes or reuse or pick up after a few months etc. 

 

- picking up scene after other artists can be always painful, as even tidy scene can seem messy since everyone works differently, however actually messy scene is much more worse

 

- name your stuff, leave notes, etc. It's the same as in programming, leaving the code messy and undocummented will bite you later, or your co-workers

 

- trying to keep your network procedural is a choice, not a must. So destructive changes are perfectly fine if that seems to be like a best think to do at the moment, as others said result is more important than the process, however be smart about it

[mestela]

http://scriptsofanotherdimension.tumblr.com/

 

 

[Kardonn]

So in the end it seems like studios don't care even if your network is the biggest mess this world has ever seen with mind boggling complexity, even if you hand create/edit/delete things as long as you make great looking things, right?

 

Whereas if you make something so simple and so elegant that it's a masterpiece in design but just looks ok, or average, they won't be impressed by the cleverness of your setup.

 

Honestly on a super complex FX shot with dozens of things going on, I don't care how clean you work...it's going to take me hours to pick apart the scene no matter what if I had to take over the shot from you.

 

I definitely DO think it's best to work clean and smart of course, even just for your own personal benefit. But I also very strongly think production managers need to stop with this ridiculous idea that ANYONE in the department can pick up ANY shot for notes and revisions. It wastes hours of everyone's time for no benefit.

 

And to me that applies to any department, not just FX, though they'd be up at the top of my list with how custom almost every shot ends up being. Often, shots at our studio will bounce around to 2-3 people (per dept no less) over their lifespan...it's an absolutely massive time killer, and frankly an even worse morale killer. You lose the pleasure of seeing the task through to the end, and I gain the distinct displeasure of losing hours figuring out how your scene works and what I need to do to address these notes.

Edited October 28, 2015 by Kardonn

[magneto]

What about having your fx rigs as generic OTLs? Anyone does this?

 

I ask because it seems like managers expect a completely procedural, reusable tool after you finish your shot. Isn't this unreasonable? Even for SESI it takes at least one person per release cycle to develop a mostly usable major tool that can be used by anyone.

 

So how can you do the same when you don't even have a fraction of that time, while the requirements keep changing daily?

 

So I have a hard time imagining how one  can develop a system that can deliver any sandstorms, or structural destruction, etc at the click of a button with minimal setup.

[acey195]

Well if you think of your problem as tool development, instead of FX directly, I do not see why it would not be possible.

It may not fit within the schedule of your particular project,

but if your team has any long term members that can work on these kind of things alongside the project,

it should be done at some point (or in between projects), at which you will make up the time "lost" by creating this tool, over a period of time.

 

That is at least, how I would prefer to go about it, for it is often about long-term gain,

but even in the short term (2 weeks) in some cases it is already possible to already win back the extra time spent on a certain tool, multiple times over.

[Guest tar]

 

 

I ask because it seems like managers expect a completely procedural, reusable tool after you finish your shot. Isn't this unreasonable? 

 

It can be that the manager is a bit rubbish too.  My personal favourite is working with an animator who was critiqued as lazy, and 'just strummed guitar at their desk instead of doing what we pay them to do!' Hmmm that so-called loafer has been consistently nominated for many oscars at a totally different company...

[Guest mantragora]

... at a totally different company...

Probably some music company

[michael]

What about having your fx rigs as generic OTLs? Anyone does this?

 

I ask because it seems like managers expect a completely procedural, reusable tool after you finish your shot. Isn't this unreasonable? Even for SESI it takes at least one person per release cycle to develop a mostly usable major tool that can be used by anyone.

 

So how can you do the same when you don't even have a fraction of that time, while the requirements keep changing daily?

 

So I have a hard time imagining how one  can develop a system that can deliver any sandstorms, or structural destruction, etc at the click of a button with minimal setup.

 

on Happy Feet 2 I built an otl that would take in the animation of the little Krill characters from Maya and procedurally create the little hairs on all of the arms and then render them

 

it took a couple of months (it shouldn't have but well...that was Happy Feet 2...)

but once everything was rolling I had one of my little monkeys write a python script that would allow us to run as many characters in as many shots as we wanted right from a command line - never had to open Houdini at all for these characters ever again.

on The Wild we had, near the end, a system that would let us render any shot, with any version of any character/camera from the command line - because everything was a tracked asset - sets, characters, animation, cameras, lighting etc - you set it off and in the morning you had 100's of rendered elements - send them to comp and go to the pub

 

so these kinds of tools are possible to make - and actually work...

but often they depend on regular and consistent input....which in FX you often don't get.