Particle blow effect?

[caskal]

Hey magicians,

Just saw this fantastic spot by Lewis Orton and I was wondering how to achieve this effect where the particles are calm and explode up (sec 0:9 and 0:22), I tried with a ripple sop to drive the particles but the effect is very poor (and particles go up and down instead of up only, tried vector to float here with no luck)

Any tip to get me on direction will be really helpful!

Thanks!

[caskal]

Talked with the author, he gave me this article (I translated with google), it has some nice info

http://mp.weixin.qq.com/s/aL72NA03SGn4jCs27f_lgA

[Atom]

Here is the Google translate.

Quote

Part II to analyze some of the technical points in the Lenovo Yoga 6 Pro promotional video participated in the production in October. This article will not step by step to resolve the entire production process, but all the points should be easily overlooked should be taken into account, a friend should be able to more easily understand.

Each screen went through several rounds of technical and visual tests, and finally found five versions that could meet the expected screen effects.

The final particle solution for each shot uses Redshift Instace to select 6 models from 6 different sand models based on the particle @ id pair, which greatly improves the rendering speed and maintains good picture quality. All pscale and @orient are randomly controlled and visually each sand grain appears to be sufficiently morphologically different.

Mobile sand:

The first 2s of the scene and the above scene in the production is actually not the same scene. The first terrain was created by two other designers in the C4D, mainly using Octane Scatter to sprinkle the particles. This next set of close-ups and all subsequent particle shots were made directly in Houdini and rendered using Redshift.

The close-up of the first paragraph is actually a FLIP that renders the water particles as grit. This is a good example of how dynamic design does not need to be used in the original purpose of the tool.

Points:

1. Solve the input form for the square sheet box, only a pop vop to velocity input, pop vop internal only a 2d curl noise effect in the horizontal direction, output to the force;

2. In order to prevent the particles from disappearing, you need flipobject. To fully close the Close Boundaries, turn off the Solver's Kill Outside Volume Limits and close the Reseed Particles.

3. PIC / FLIP in the data exchange and advection process will naturally produce errors resulting in particle alienation or compression (resulting in "drawing" phenomenon), so to open the particle Separation;

4. FLIP default particle label, need to hook Add ID Attribute;

After adding an ID, you can use Time Blend and Time Offset to slow down the SOP. The color is controlled by the position of the x-axis in vex using gradients. The fluid is constantly distorted by Curl, and in order to maintain constant visual distortions, two time offsets need to be created with a steady time difference of 10 frames, for example, and then by adding rest to the former and passing the former through Blendshapes Offset to the latter, so you can rely on the frame difference to arbitrarily control the ideal fluid color distortion. Once all the above steps have been completed, use the Point VOP in the SOP to add the terrain relief based on @ P.y plus vertical noise.

(Tips: Solve the particles after the first use of clean will not need to remove all the properties of the cache, the follow-up read speed will be greatly improved, but also reduce the storage required hard disk space)

Jan sand:

This lens has been solved three times, the initial setting and the previous one is about the same.

Points:

1. Into the first FLIP DOP curl by a Curl Noise to a nice shape after a cache, add rest, connected to Point Replicate ten times the number Upres;

2. Enter the second Grains Solver DOP to switch to particle solver so that all particles are next to each other under the influence of gravity and freeze the frame state with Time Shift freeze;

3. The third DOP is still Grains Solver, but in the SOP it is necessary to prepare a line that moves right each frame as a second Input into the DOP.

Line segments use Point VOPs to control the orientation and length of @N and @v. Use VEX in DOPs to move the @v of points on the line segments to nearby points at points close to the line segment, and control the acquisition strength of @v based on the distance . This line segment swept across the area near the point will be the initial speed and was quickly pulled up to form the sand effect. Finally, use Time Blend and Time Offset in the SOP for speed changes.

expansion:
The main part of this footage is still FLIP, but a force field is designed to achieve the desired effect.

Points:

1. The basic settings and the first one is exactly the same lens, the only difference is that the initial length of the box is three times the width of the top view of the solution to take the middle of the fluid area, both left and right extend to the outside of the screen is to force field buffer use;

2. DOP settings and force field as shown below
Visual description, the force field will be divided into left and right images on both sides of the region, the left of the particles pushed to the left, the right particles push to the right, but the entire force field only in the central horizontal line of the screen area, upper and lower regions without force. The final result is shown as the result;

3. Based on the solution above, we use Point VOP to add a wave diffusion effect, and only do the Y-axis position replacement on the particle. The basic principle is that the particle's abs (@P) Surplus, map the remainder to Ramp, and add the resulting value back to the Y axis value of the location (to be more precise, the input source is abs (@P * {1, 0, 1}) and finally a little Curl To make the waves slightly imperfect).

Shock wave:
Although the final effect of this lens is relatively neat, but the production process is more complex, different links have gone through many rounds of iterations, replacing the different ways to achieve the final determination of the final results. The entire shock wave is divided into two steps, the core diffusion that occurs in the first and second shots, and the large spread behind the screen in the third shot pen. The following only say the final version of the production process.
key step:

1. Prepare the ground.

Create a nice gradient map in other software as the basic hue of the entire floor. Adding a Scatter to the Grid To densely populate a layer of dots, add a pscale with random variations, import it into a DOP and use POP Grains to solve it once, so that all ground particles are separated and slightly stacked, and use the Attribute Transfer Obtain the base color from the gradient map made from the very beginning to complete the ground surface sand.

2. Prepare core diffusion speed source.

We need to get an initial speed source, the shape of a circular outflow, but need to be able to freely control the speed and direction, as shown below:
(The blue line indicates the direction and intensity of @v on the circle point)
(Speed ??source preparation node)

First introduce a point into the POP net, generate a new point every specified number of frames, and within the DOP make @pscale grow with @age and then copy the circle to the above point using Copy to Points Continue to generate a new diffusion ring effect.

Followed by the first round of two-node speed change, making the ring first uniform growth and then quickly accelerate and then slow down. All operations below are based on a variable speed diffuser.

Create a Point VOP, create two vectors internally, one for Normalized (@P) (ie, on a horizontal plane facing away from the center), one for {0, 1, 0}, use a ramp to mix the two with noise, and Give random values ??for all points on the ring, adding visual irregularities to the diffusion ring.

3. Prepare core diffusion particle source
Using the above-generated velocity sources, VEX is mapped to @Cd and @chance (custom properties) of the Grid and the attenuation amplitude is controlled based on parameters such as the radius on the circle point, the distance to the ring core, and the result is taken as The particle input source introduces the first Input of the DOP, introducing the above diffusion ring as the velocity source into the second Input of the DOP.

4 core proliferation solution
In DOP, the POP Source is set to Scatter onto Surfaces, the probability of generation is determined by the above @chance, and constant particles are continuously created using the constant density method and a group is assigned to the newly generated particle.

For newly generated particles, use POP Wrangle to get the initial velocity from the nearest point based on the distance to the nearest point on the Input2's velocity ring and to attenuate the particle velocity in each step, add POP Wind to add random air perturbations to all particles influences. This step solution is pure POP, did not use Grains and FLIP.

5 pairs of core diffusion particle solution for two-node speed and cache the results;

6. Repeat the above steps to make the outer ring part, after two-node speed change and then add the time difference (Outer ring shock wave in the proliferation of core proliferation) and cache;

@Crests the @Cd from the ground grid using @rest to solve the particle;

8. According to the lens number, the merger needs to be part of the solution, the corresponding time offset, the merger of ground sand.
In addition to the many steps and the middle of many R & D, this lens also has multiple rounds to control the rhythm of the screen. The number of shift and offset frames needs to be repeatedly adjusted back and forth. To perfect the animation of the imported alembic camera and pen , Can not appear even a frame of error. Finally, the sequence of frames rendered directly in Houdini requires motion blur with particles but without camera motion blur, back to C4D with the original camera using the previous rendering results as a texture import and Camera Projection on the computer screen, Once again, the well-crafted material and lighting screens and pens are rendered with Octane as a whole with motion blur on the camera, which ensures perfect matching of all picture effects, far beyond the effects of rendering and resynthesis, respectively. So this set of shots actually uses two different renderers to render at the same time

At this point all particles in the project using Houdini produced focus analysis of all the key difficulties, hoping to be helpful and inspired friends.

 

 

[Jesper Rahlff]

beautiful spot and a very insightful paper! Thanks guys

[anto187]

really admire for this fx, and really great to if there any hip file.

[Bleip]

On 19/12/2017 at 3:04 PM, Atom said:

3. Based on the solution above, we use Point VOP to add a wave diffusion effect, and only do the Y-axis position replacement on the particle. The basic principle is that the particle's abs (@P) Surplus, map the remainder to Ramp, and add the resulting value back to the Y axis value of the location (to be more precise, the input source is abs (@P * {1, 0, 1}) and finally a little Curl To make the waves slightly imperfect).

Sorry to resurrect such an old thread. I'm working on something similar trying to get a decent wave effect post-sim and found this. Can anyone work out what they mean by the above?