Dust Update

Here I'm using an AttributeTransfer to transfer colour from the water to the particles. A delete node was then used to delete based on colour. This way, I could kill the particles that were entering the water without having to resim. However, it wasn't perfect and a resim might still be needed..

Rumbling Dust update

Here, I've tweaked the dust on the sides after Mr Michael suggested that the dust should start building up from the top of the valley instead of all appearing at once.

Since I have a file reading in only the geometry that is emitting particles, I animated a delete node so that only the geometry near the top would emit particles at the start of the shot.

Crack Dust Update

Update for the dust from the main crack. References I used include the earthquake scenes from 2012 provided by Mr Douglas. I tried rendering the smoke out completely white and adding the colours in compositing as Mr Ron suggested, however it did not look very nice with a single colour. Hence I used a ramp for the colour based on the lifetime of each particle.

Below is a test comp with the single colour.

Managed to get some groups out from the geometry by deleting based on colour. This gave me the polygons that were created during the fracture in DOP, I could then emit particles from those surfaces, making it look like dust and debris emitted during cracking. However, this was not enough, as I realised that some of the cracks had been made with a Voronoi fracture in SOP and the inside group was no longer existent.


Currently rendering the final main dust cloud, and trying to render the dragon's aura again.

The high res rocks had to be ropped out again due to the rocks not having a rotation value after filing in from DOP. Mr Ron then showed me how to get back the rotation values, by doing a DOP import, importing the rocks as points, then ropping out those points. The rocks were then copied onto those points to enable them to rotate.

Crack Dust

Here is an attempt at tackling the challenge of creating the dust produced when a fracture occurs. I could have used the inside group to emit particles if the fracture had been generated in SOP level, however, in this particular instance, it was done in DOP. I could have also used the impact data if I had set the network up to record impact data while doing the fracturing in DOP, however I did not due to the lack of experience and foresight. So the effect would have to be faked as there was no more time to resim.

Here I'm using the surface that Mr Ron created to sliced the terrain in half to emit particles upward, colliding with the terrain. As can be seen in the above flipbook, a lot of particles are being trapped below and those that do get to the surface are flying too quickly.

Here a more elegant system is in use. Initially, a curve was made out of the cutting surface with a carve node, however as it gave problems later down the network, I traced a curve out instead. This curve was then plugged into a creep, together with a sphere. The translate was animated to match the opening of the ground. An AttributeTransfer node was then utilised to transfer blue onto theterrain. From here, particles wold be emitted from the blue ares of the terrain. In order to optimise the simulation, non-blue areas of the terrain was deleted procedurally.

Result of rendering with Houdini 12, the sprites managed to be rendered, however they came out as a brown mess.

NDP Citymaker!

A digital asset that I created for the NDP project. The purpose of this asset was to create a "city" designed by the NDP team and the original request was the internal surface. To ensure that no 2 extrusions would penetrate each other, two group nodes were used to select every alternate face both vertically and horizontally. This produced the desired result, but it was too regular. In order to randomise the placement of the "buildings", a random function was used to create a selection of polygons for extrusion. The team later requested for the outer surace to have "buildings" as well, so a reverse node was used to flip the normals before extruding. The 2 surfaces were then merged together.

In the render below, I added a curved top to the hemisphere as the team did not want it to be a flat surface. The curved top was done by extruding the top upwards and using a mountain SOP on those points.

Dust Test render and playblast

To create this storm of dust, I used an AttributeTransfer to transfer red colour from the spheres to the terrain and emitted particles from wherever there was red. To speed things up, I also used a delete node to delete all unnecessary geometry; geometry that was not red.
In order to achieve the effect of the dust building up, Mr Ron showed how to use a render loop. A render loop basically adds data to the previous frame. It achieves this by reading in the previous frame, passing it through the network, and writing it out, where it is read back in and passed through the network again.
This was required because the spheres were rolling about, so the emission area would be constantly changing. However, with the render loop we could get the accumulation effect.

For the particles, sprites were used, with a ramp on the
Some problems with the rendering, Mantra keeps crashing! And I have no idea why..

Compilation of progress so far

Smoke sprite test

More updates!

Falling rocks breakdown: The rocks were read into the scene as a bgeo file, passed through and connectivity and partition which would make each rock a group of its own. The connectivity and partition method works by grouping all faces that are joined together. A foreach node was then used to replace all the rocks with spheres. These spheres were then used as proxies to simulate the rocks rolling down the valley. After that, they go through another foreach node to transfer the transform values back to the high res rocks.

Tried out some smoke simulations to simulate dust being raised by the falling rocks. A box was placed by the side (off camera) to emit smoke into the valley. It did not work out very well, and I turned to using POP and sprites instead (thanks Jiabao!)

Latest update for fracturing

Latest version of the fracturing, tweaked the position of the box so both sides would fracture equally. Previously, only the right side was fracturing properly.

Previously, there was a problem with the fracturing as the terrain was too thin, resulting in large gaps of nothingness after it fractured. To solve this issue, I extruded the base of the terrain so it had more height. This gave the collided geometry some depth instead of it collapsing like a crust.

On the right track at last!

Here the fracturing looks more like the look we're after. A box is being used to collide with and fracture the terrain in DOP.

From the camera view:

Fracturing valley

Here, I'm testing out the fracturing in DOP before carrying out the actual simulation: The intended effect is to have the pieces crumble as they fall.

After successfully doing the fracturing test in DOP, I've started some test fracturing on the actual geometry. However, the falling geometry (red) is not falling realistically, and is falling together as a chunk instead. Hence, I'm using a sphere to break up the chunks as they fall.

Here, I'm experimenting with animating the sphere to get a more realistic result.

Smoke test comp & Fracturing R&D

atmospherics

Test flipbooks & Renders

In order to prevent the smoke from rising too quickly, I reduced the uniform force, however, the side to side motion was still a little fast. To solve this a drag was used to slow the smoke down. The sine function of the vorticle's transform node was also further tweaked to get this movement.

Finally I needed to render the smoke without the dragon, however the depth of the smoke could not bet seen if the dragon was just composited on top or below the smoke. A matte of the dragon would have to be rendered with the smoke to achieve the desired effect. The matte shader did not work out in this case, giving an undesirable alpha. In order to generate an image with the dragon completely black, I applied a color node with a constant black on to the dragon. I could then shuffle the red channel to the alpha in Nuke.

Yet another test render

Here, I added an upwards uniform force to the smoke so it would rise slightly. However, it seems to be rising too quickly, causing it to look like it's boiling.

Aura test 3

At this point, I'm trying to create movement in the smoke, instead of having the smoke stay put after emitting from the dragon. By increasing the resolution of the smoke container, a lot of the finer detail appeared, but at the expense of calculation and render time. The values of the sine function in the transform of the vorticle field were tweaked to amplify the movement of the smoke.

I also discovered that the guide smoke displayed in the DOPnet was inaccurate, a better representation would be the DOP import created with the smoke object.

Aura test 2

Aura Test

I'll be doing the aura for the dragon, so I did some tests using voxel fluids in Houdini to get the look of the aura.

Using a curve to control the path of the smoke:

To use the curve to influence the path of the smoke, I swept the curve with a circle, then made it a fog volume and scattered points in it. Normals created for the points on the curve using a Point SOP. The points were shifted using a Sort SOP and fed into the second input. This set of data was then used to make the normals follow the direction of the curve. I then created an attribute "v" (velocity), using the normals as the value. This attribute was then transferred to the points scattered on the fog volume. This set up was then fed into the DOP network via a SOPGeo. A field force and gastofield was used to create a field force from the geometry information. The field force then influences the smoke.


Testing with an animated object:

For this, a gas dissipate was used to make the smoke fade out over time. Turbulence was created by using a vorticle field. However, this was not satisfactory and I went into the node and added a transform, using a sine function to translate the field back and forth over time, creating more turbulence.