Is this realistic?

[reelinspirations]

Hi,

 

I'm attempting to assess whether or not what I'm trying to do is reasonable and realistic.

 

I've been dabbling in 3D for a number of years and a bit of experience in Cinema 4D, After Effects and Lightwave.

I just started learning Houdini because of wanting to produce some VFX for a documentary I'm helping a friend with.

 

What we're wanting to do is track some aerial drone footage of a hillside and have a portion of the hill fall away to reveal the films title.

We are wanting the realism that Korhon pulled off for Gimpville (see image below)

 

I'm not needing water, or the entire hill to be falling away, I'm only needing a portion of it to slide away and reveal text.

 

Given that I have been learning Houdini over the past month, is something like this even remotely possible to pull of single-handedly to that level of realism?

I've watched some great tutorials by Sidefx, Digital Tutors on controlling your fractures, etc.

 

I need to know because I'll work myself into oblivion trying to do it because I don't know what I don't know.

But the more I get into Houdini and read posts about fracturing, clustering, etc. the more I feel inadequate to even attempt this.

 

Thoughts?

[michael]

I think it's realistic for one person - depending on the time you have

the important thing to do would be to break the shot down into it's basic steps and then ask the right questions for each stage

there are lots of tutorials that either do exactly this or are close enough to be easily modified

so here is a VERY basic list of steps:

1 - get the track into Houdini - Houdini doesn't have a tracker but can use the information provided by most on the market

2 - identify the assets you'll need to build - the cliff (based on the track), the title, you can probably use projections to texture

3 - rough out the animation - hand animate the chunks of wall falling to reveal the title - just to get a feel for what you want to end goal to look like

4 - start fracturing the cliff and test some basic RBD sims (maybe look at Grains also)

etc

 

start a new topic in the Work In Progress section and post your progress with any specific questions you have - you'll get lots of help along the way

[reelinspirations]

Thanks for the reply Michael.

 

How much time should I expect this to take?

As I do more research, I keep finding things I didn't know. It's a DEEP rabbit hole :-)

 

Right now, I've been working on learning to fracture things, glue networks, controlling fractures and the RDB system using Bullet.

 

I have Syntheyes and can get a good track, so that's taken care of. I'm sure there is something somewhere about getting a Syntheyes track into Houdini.

Building the cliff based on the track is tougher to figure out. I'm guessing I need to model something that resembles the exterior of the hill, plus the lower portion of the hill for the pieces to react with.

Then somehow marry it with the track so it all 'jives' together.

 

Yeah, I was wondering about Grains. The hill is very sandy looking, so having some grains in there would be good.

 

The WIP section is a great idea. I'll start that soon.

[Guest tar]

SynthEyes exports a Houdini Script

 

You can build a cliff mesh in SynthEyes, page 312 'Creating Meshes from Tracker Positions ' in the SynthEyesUM; in NukeX with 'PointCloudGenerator' node, and probably Houdini with SDF /VDBs convert to mesh. MeshLab can also be used to process the mesh, and probably build one from points, and finally PhotoScan probably could work. 

 

You can rework the mesh in Houdini with the Sculpt/Edit node and Zbrush for more better control.

 

Also checkout Instant Meshes to generate a nice quad bashed mesh if Zbrush is not available.