Disintegration Effect in Houdini

First couple of steps were similar as had to create a geometry node with a sphere node inside – I’ve already started using new skills from recently learning Houdini’s interface through Lynda.com i.e. using the keyboard for quick transitions going in and out of nodes.

The image above is the beginning processing of me adding the geometry node, using ‘I’ on the keyboard to go inside the geometry and add a sphere. Adding a box node into the scene will be what allows me to dissolve the sphere

After changing the primitive type of both the sphere and the box to polygon mesh, I added a boolean. Before this tutorial I didn’t know what a boolean was however I now understand that I have to use the boolean in order to remove the geometry.

After adding a transform node and placing it between the box and the boolean, I learnt how to add key frames into the scene. The image above shows me adding a keyframe so that the sphere disappears after two seconds (frame 48). I did this by right clicking into the Y axis box and selecting add keyframe.

I added a blast node and connected it to the boolean, which means that when I create an ‘a inside b’ group in boolean and apply it to the blast node, it creates the image you can see in the scene above. This geometry is where the particles will be created from.

The following step was to merge the object together with the OUT_GEO node (image above), this way, the object that dissolves runs as a whole rather than separate entities. When the object was dissolving, it was very smooth and linear – which looks natural for a dissolving object. Therefore, the image below shows what the object looks like when a ‘mountain’ node is added in-between the transform and the boolean.

In the image above you can see that the sphere isn’t smooth round the edges, this is due to the box not being large enough which ultimately interferes with the sphere, to change this I went into the transform node and made the box slightly larger.

The two images above show me changing the way the particles move in the scene. The first image is the original particles moving with gravity force on approx -9  and the second image is when the force is changed to 1. It was necessary to add a ground plane so the particles have somewhere to land originally, but when I changed the gravity, I could remove the ground plane.

I added a POPaxisForce node as this allowed me to change speed in which the particles orbit around the object, this creates a more organic looking set of particles.

As I started getting into editing the particle nodes, I realised that it was easier to screen record my progress as that way I could concentrate on the work instead of screen shotting and explains the progress. In the video, you can see me adding a colour ramp to change the colour of the smoke emitting  from the object. A binding node that allows the other nodes to coincide with the multiple node which fuses the other nodes together.

Time lapse of developing my particles :

Adding principle shaders to both the particles and the sphere allow the colours to be viewed in the rendered view. I then added a principle shader to the ground plane so that my final output would have a more professional look.

Final Disintegration Output:

This particular experiment took much longer than I had anticipated. I believe that it was down to the process of editing the particle nodes as there were many different elements that needed adjusting in order to make it look organic. This being said, I feel as though this was a good step into learning particles that are similar to dust/smoke as this is another experiment that I want to take forward for future and therefore ultimately have developed more skills

Initial Trouble with rendering video sequence after output left Houdini:

I had no trouble rendering out the ‘Disintegration effect’ from the software Houdini, however, when I went to put it into a video sequence and export it, the file kept crashing. Initially I thought it was fact that my laptop wasn’t advanced enough to export such a large file so attempted to render it out on a few other platforms, again each output failed. After explaining to my tutor what was happening, we decided to see if it would render out the video sequence in Blender. Originally, Blender kept crashing too. However, we experimented with changing the settings in Blender which eventually allowed the sequence to render out in a video.

Learning the Houdini Interface

As I have began the process of learning a few simulations (the jelly and water drop) through youtube tutorials, I am becoming more familiar with using the softwares interface as I go along. This being said, I felt that in order to create more intricate simulations, I want to be more confident when navigating around the interface of the software. In order to increase my knowledge on this, I went back onto the ‘Lynda.com’ website that provides information on this aspect (and more).

https://www.lynda.com/Houdini-tutorials/Navigation/571627/629709-4.html?autoplay=true

Below is a time lapsed video that I uploaded to youtube which evidences me learning the interface through Lynda. When I get stuck, I can referring back to this video/back to Lynda.com’s interface section to keep working efficiently on my work.

The headings that make up this sections of the interface contents is:

Navigation: explores the different ways to move around in the viewport

Viewport and display modes: shows the different types of geometry for the objects in the viewport

Panes: How to expand the different sections i.e. making the node section more visible and removing unnecessary tabs

Desktops: How to safe different desktops in the software

Preferences: How you wish to work and what you can adapt in the software depending on your project

Display options and visualisers: explores scene customisation and texture settings

Global animation options: setting up the overall parameters of the scene

Nomenclature: talks about the key words and names in the software i.e. ROP, VOP, VEX

Network view: Area where to build scene and explains how to go in and out of nodes (‘i’ to go inside and ‘u’ to go up)

Node flags: the areas surrounding the node on the right and left

Geometry spreadsheet: learning about points, verbs and primitives

Treeview: being able to see a breakdown of what in the scene

Shelf tools: exploring the tools displayed on the software (i.e. the box, sphere, grid etc)

I feel that when learning through Lynda, my work ethic is more efficient. Therefore, when I attempt a new simulation I will see whether there are any tutorials to follow on Lynda that can improve my work.

Simple Water Flip Fluid Simulation

The youtube tutorial linked above focuses on using Flip Fluids and explains how to cache out the simulation so it runs smoother when working in the software. I have decided to follow this tutorial as it is a simple starting point to learn how particle fluids may work, keeping in mind that I want to develop the skill into more intricate fluid particle simulations.

The first step was to place a sphere object into the viewport and select ‘flip fluid from object’ to apply to the sphere. This created the current image I have below

.

The following step is to create a ‘ground plane’ so that the particles have something to collide with when playing the simulation. Something I have learnt from creating a ground plane is the difference between the grid object and the plane. The grid object doesn’t automatically have any qualities applied to it, meaning that when I press play, my particles will fall through the ground. Whereas when adding a ground plane, the static collision qualities are already applied, meaning I do not need to add any more to the object.

In the image below I have changed the value of the particle separation from 0.1 to 0.01, which means that when the fluid falls, the particles stay closer together which creates a more natural effect. However, this also slows down the viewport simulation a considerable amount due to having more particles to calculate and cache out.

As this is an experiment, the only thing that needed to be added to this was a background plane and adding a light and camera in order to see it in the render viewport.

First Fluid Render:

Personally this was the simplest simulation to develop, however on thing that needs changing (if I come back to it) would be the values of the background plane, as when the fluid hits the ground it passes through the grid in the background, making the simulation less organic.

When going back into this simulation, the few things I needed to change were simple, and I did not require any assistance to do them, as I have become more acquainted with the tools in the software. I changed the background from a grid to a ground plane collision. This means that the attributes that allow the water to collide are already preset, the only thing I had to change was the positioning, so that it could form a background. I also changed the colour palette in order to see the fluid clearer on the simulation. I turned the water slightly bluer, and the ground white. Once I added a distant light, these colours were much more visible.

Second Fluid Experiment:

Frames Rendered: 140

Render Time: 48 hours

Houdini Research

As I have previously already done some research on Houdini due to the project  I did last term being based on the same thing, I can use the old blog post that I made as reference for this project. The link below is direct to the post I did on Houdini research last semester, it includes information on the features of the software, pros and cons, node use and films that used the software. Each of these aspects I learnt were valuable as it gave me a better chance of understanding the software before I went onto it.

https://med3017a-1819-gglevel3projectone.coursework.lincoln.ac.uk/2018/10/15/houdini-research/

As I am now starting to get a hang of the interface, one of the aspects I will further researching into is the node use as I have realised this software is much more heavily based on the technology and geometry side of things.

I accumulated some of my node research from the sideFX.com website, as I know that this information will be accurate due to Houdini being developed from this company.

Types of Nodes: http://www.sidefx.com/docs/houdini/nodes/index.html

Channel (CHOP) – Create filter and manipulate channel data

Compositing (IMG)  – Create, filter and manipulate image data

Geometry – are inside the geo objects and generate geometry

Dynamics (DOP) – set up the conditions and rules for dynamic simulations

Object (OBJ) – represent objects in the scene i.e. character parts, geometry objects/lights/cameras etc.

Render (ROP) – either render scene out or set up render dependancy networks

VEX – containers for VOP networks

VOP – allow definition of a program (like a shader) by connecting the nodes together, software then complies the network into VEX code.

Having a small understanding on how the nodes work in the software and how they are linked together to create the simulation will give me a more knowledgeable approach when in the creative process, this knowledge will allow me to increase my skill set and challenge my current capabilities.

The image above is what a ‘geometry node’ looks like. As I have already begun the process of creating ‘soft jellies’ this node is familiar to me, alongside an object node and render node. However, I have obtained a clearer understanding after seeing what components make up the geometry node. For example, when I was blindly making my way round the interface and struggled to see the object in the viewport, I didn’t realise that I had to highlight/select the component furthest right.

Comparing to other software i.e Maya:

Although Maya is good for modelling and texturing and – in terms of learning the software – Maya would be simpler to learn due to the easier set up and traditional workflow, the thing that makes Houdini good for this project is because i want to learn how to create VFX in ‘the elements’ i.e. water, fire, smoke and things that behave like a liquid or gas. Houdini is known for creating professional outputs that looks realistic. From my previous term I learnt that there were many films that used Houdini to create sequences involving sand (Spiderman) ocean (pirates of the Caribbean).

https://www.educba.com/maya-vs-houdini/

 

First Houdini Experiment (Soft Jellies)

I started looking through a website called ‘Lynda’ (https://www.lynda.com/Houdini-tutorials/Houdini-Essential-Training/571627-2.html?srchtrk=index%3a1%0alinktypeid%3a2%0aq%3ahoudini%0apage%3a1%0as%3arelevance%0asa%3atrue%0aproducttypeid%3a2) as it provides information on beginners using Houdini, however I felt that it wasn’t providing any information that made sense to me as I usually learn through a process where I can see an outcome, so found a youtube tutorial where I could visually learn how to create a simulation before I begin to learn the technicalities behind it.

To begin with,  I want to create a simple simulation in Houdini in order to grasp the scope of the length each experiment may approximately take. As with my previous project I began by looking at rigid body and soft body simulations, I felt that I should follow a similar path as it slowly eased me into the software – hence looking at ‘soft jellies’ as this is similar to my bouncy ball simulation from my previous project.

When researching tutorials to follow I discovered there is little information out there that is straightforward to follow, this being said I found a youtube video that has a demonstration of how to do this kind of simulation.

I found this tutorial difficult to follow as it didn’t have any voice over, therefore with a couple of steps I had to work it out by myself with trial & experiment i.e. pressing certain keys, seeing whether there was any indication in my software that explained what to do etc.

The first step was to create a spherical object which was done after creating a geometry domain (essential for keeping object inside a certain area) and make sure that it is a polygon material. The image below is the stage I’m currently at, and to reach this stage I had to select the sphere tabs and change them to an ‘Organic Mass’ which ‘creates a solid finite element (FEM) object that is pliant and elastic, similar to muscle and fat’ (learnt after highlighting the OM tab) and copying and pasting each object so I eventually have three spheres on screen. To make the objects nodes expand and actually make the organic mass, I had to select the polygon and press enter on each object which also allowed me to put the OM into action and connect them together.

The following step was to add a ‘merge’ node so that I could connect the spheres together – this allowed me to see the three individual spheres through the viewport.

I then added a ‘cache’ to connect to the merging node, as without this my simulation would not move in the motion I desired.

I added a plane  (so jelly balls collide)

Enter the ‘AutoDopNetwork’ node system to change the values of the objects so that they collide more naturally.

added a camera and a sun light (need to change to jelly material)

added material after following the instructions that sidefx provided http://www.sidefx.com/docs/houdini/ref/panes/materialpalette.html

Subdivide and surface smoother (different) 1. affects the whole geometry and is the reason to make things look more natural once rendered out – had to create a separate node in 2. only a temporary affect while editing and doesn’t change the rendered look.

My Jellies Final Output:

Frames Rendered: 200

Render Time: 4 Hours

One of the things I am struggling with is the size of the software being on my laptop as I feel that its fitting onto a small screen and therefore is more condensed, resulting in smaller tabs which makes it more difficult to see accurately what I am doing. I solved the interface issue by changing it in preferences and making it larger, this way, I am able to see what I’m doing with more precision.

 

Learning Agreement

Student Name Giordana Guerriero
Project Title Advanced Physics Simulation Project
Tutor(s) Name(s) Jon Holmes, Graham Cooper, Clive McCarthy, James Field
Group Members

(If Applicable)

N/A

 

Project Type

Please select the optimum assessment strandfor your project, as described in the core module guide. Weighting for each submission component varies from strand to strand.

Type 1

MAJOR ARTEFACT

Type 2

PORTFOLIO

Type 3

EXHIBITION

X

Brief Project Outline

A short outline of your proposed project

Last Semester I intended to learn different skills in the software ‘Houdini’ but as I focussed more on Blender, this semester is where I will take the knowledge learnt from Blender and apply it to Houdini.

To begin with, the aim of this project is to start by creating different physics simulations i.e. fire, smoke, water, in Houdini and produce a show reel of the work in a final output – similar to how I structured my previous project in Blender. This being said, for the first few weeks I will keep my options open as I understand the scope of this project is larger than my last one as this software is more intricate. If this doesn’t go to plan, I could potentially switch my focus back onto using Blender but combine my knowledge of this and Chroma Keying to create VFX in live action scenes.

My project type will be Portfolio, as the nature of the project involves some level of experimentation, I will be frequently updating it in my blog and there is 20% of the focus on research and development. It also gives a clearer aim of my project having a combination of small experiments to make up for the final output of my project submission.

The concept of my idea will be similar to my previous project, as I will be bringing together the small projects in a final show reel of work to present to those in the industry. It will have before and after shots to show what the simulations looked like before the render and after the render. 

My intended audience is those that are already in the industry and looking to hire. This way, I am more motivated to think about industry standard work and also will provide a good portfolio of work to take forward with me when looking for a media-based job.

I will use YouTube to upload my final output so that I can imbed it into the spark page, my YouTube account has many different video uploads from my previous project and is therefore a suitable platform for my work to be seen.

 My technical abilities will be challenged further in each element of this project. Whether I experiment in Houdini or create scenes with Blender, both will be a step further into my skills and the more I learn the more in-depth and refined my work will look. Providing a portfolio that reflects an experienced user with physics simulations.

https://vimeo.com/226906993 I have taken inspiration from this show reel of work as it is from the SideFX Houdini website has been used in 2017 films/adverts. It allows me to see how the work I’m doing can be placed into industry level work i.e. Moana, The Smurfs etc. 

https://www.youtube.com/watch?v=bglBfYdNIvM this YouTube video is where I can take inspiration on if I am to go forward with live action sequences using Blender as it gives me an idea for potential scenes/narratives for short clips to solely demonstrate my skills.

Research Outline

A short outline of the research you will personally undertake as part of the project, including relevant texts. Try to add specific details where possible

Primary Research: Experimenting primarily in Houdini to test whether the scope of the project is possible for this semester. Following this, I will conclude on whether to continue in Houdini, or to revert back to Blender and create simulations as part of small live action sequences.

Secondary Research: The main source of secondary research will be from YouTube tutorials as I believe that these are a fast and efficient way of gaining clear information on my project topic.

I can refer back to my old blog that I did that has previous research on Houdini. One of the websites I found last term was https://filtergrade.com/houdini-overview-vfx-3d-modeling-software/ which is a Houdini overview and will be appropriate for this project as I can prepare for the difference between using Blender and the complexity of Houdini.

https://www.lynda.com provides a lot of different video tutorials for beginners, with navigating the interface to understanding elements like lighting/dynamics/particles etc. This will likely be my ‘go to’ in terms of learning Houdini.

If I was to take forward compositing a physics simulation/object made in blender into live footage, https://www.youtube.com/watch?v=CCBZifZQEf0 this YouTube video has resourceful tips to take into consideration when experimenting with combining blender simulations with live footage

To evidence that I will be developing my project alone, and after discovering that the use of blog posts gave me better time management, I will be uploading blog posts that explain each experiment, and demonstrate my work through discussion and comparison with the use of still images, time lapses and towards the end potential videos that can be used for other people’s tutorial purpose – to show that I have learnt new skills and can relay the information to others.