UFO- after feedback

This was again written over the weeks that I worked on it, trying many different things starting over many times. I am happy with the result and think it fills the brief pretty well.

I love dynamics it is fun and exciting to be able to produce simulations which move dynamically. It is the field of special effects which has been developing ever since movies started. “A century ago, Georges Melies was the first film maker to discover that sheer spectacle was not enough to keep audiences coming back. Hollywood has always sought to provide ever-more spectacular images in order to keep audiences coming back for more. But the advent of digital effects has started to challenge the old rules. In 1993, cinemagoers queued round the block to see the computer-generated wonders of Jurassic Park. Just a few years later such images were almost commonplace.” (Rickitt, n.d.)

Now we are learning the lastest techniques with the hope of building on what we have learnt and trying new ways to produce what a directors’ imaginings could be.  “Particles are versatile and can be rendered in a multitude of ways. Relatively inexpensive to simulate, they can look like sparks or a splash of water. Particles can also be used to drive geometry or be driven by other simulations. Their adaptability makes them a powerful and resourceful tool.” (Palamar, 2010)

Inspiration in images

First, examples that I found inspiring from the client and other images with things he said he liked

moving references that I found: plasma tendrils- movie of sun plasma arms


plasma ball

In these images I am inspired by the moving tendrils and the beams of a varying transparency, the blue colours and the small dots floating in it, also the plasma balls have arms which move about a lot  I would like to have tendrilly arms to go down inside the wider beam and pick up the man. I like a cone of cloudy light and tendrils in the images I think the tractor beam works by first sending down a wide general beam to get the specifications of the position of the target like a radar and when it finds a target it sends down more intense static electricity and plasma ball like tendrils that find the target and draw it up into the UFO.

I looked at plasma balls and the way tendrils move in that, my tendrils were too static and I tried giving them some speed and direction but I can’t get them to move like the arms in a plasma ball so I am trying a cylinder to emit from surface set around the area and goaling them to the tendrils hoping they will move back and forwards more. Not good so I decided to make the tendril trail particles emit then goal to the tendril trails with smoothness to make them oscillate between the 2. No good again.

Made 2 cones of differing width and goaled the new particles to them both but they swung between the top and bottom rather than the sides as I had wanted. Deleted them and made a donut stretched up it fit around the area and goaled that, getting worse. Gave that up and rethought it all from the start. I can’t make the existing tendrils move like an arm it falls apart as dots when I try so maybe I have to have some solid geometry to do the moving and have particles on the surface? But that doesn’t seem to be dynamics enough.

I reread the brief and it is inserted here: “In this shot a UFO (cg element) will swoop into frame from behind the building in the background. The UFO will hover above the grassed area in the quadrangle. It will then open its doors and use its Tractor Beam (cg element) to lift the Mannequin (cg element) up off the ground and into its cargo bay. The doors will then close and the UFO will fly away.” So according to this there does not have to be arms that lift it up and it can be a beam with tendrils shooting around in it. I am going to just make them move and not actually lift the man.

No matter what I cannot make them move after they are emitted! Finally! I gave them a mass, a ramp in the mass attribute, and used a uniform and a turbulence field and I keyed the shape of the newton field which keeps them in, a cone, to expand and contract a few times so it is much more vital now.

I don’t care anymore that the tendrils break up into particles because that was what was restricting me and the feedback asked for  more fine detail in the tendrils and look more like they have a life of their own, to keep moving the entire time and have some wispiness too. I think these details are all present now. Also the colour of the tendrils, it was said, should be more like the reference so I am comparing the 2 and adjusting the ramp in the RGB PP, it is actually lighter than I thought.

The feedback suggested that the small points looked better in the cone so I have gone back to that rather than the cloud and am exploring with points and multipoints. I thought about it and decided I could add more particles emitting from the tendrils like this


– so I have added them and now it looks like this and I think it fits the brief and feedback nicely I decided to add a bit of ‘fluff’ around the edge of the cone on the grass. I made a torus emit particles and made a volume axis field around that in a torus shape too and timed it to the cone. I eventually decided this was too much and not needed. When I referenced the new animation I have to move and lengthen my fx things to match which meant readjusting settings.

Heat Exhaust

I made a cube and sized it and placed it to fit the exhausts then copied that 4 times, 1 for each vent. I adjusted the angles and placement so that the particles emerged at a good place and angle v  selected them all v  emit from surface volume emitter- 1 particle exists for all of them

  • start at a rate of 100 and increase to 300 to get more
  • directional speed 10
  • direction attributes  X=0, Y=1, Z=0
  • away from centre 4, away from axis had no effect unless I changed the shape to a cylinder but then it emitted a hollow tube so I just changed it back and tis worked

v  on the particle shape node

  • live for 6 seconds
  • blobby surface radius 3, threshold .662 so they merge in render view
  • RGB PP- added and given a white to black ramp so they start white and end black
  • Radius PP- added and given a black to white ramp so they start small
    • edit array mapper- min .3, max 3 so they start a bit bigger than 0 and end even bigger
    • Opacity PP- added and given a black to white ramp so they start transparent
      • use the same black to white ramp and edit array mapper- min 1, max 0 so they swap the small and large values, start visible end invisible

On to assigning a shader I remembered that the client said in some feedback that he wanted the tentacles to pick up the man and when I was looking at mine it was very weak with no real connection to the man so… I decided to experiment with emitting from curves as the tendrils to see if I could get more control over them to go down and pick up the man and then go back up.

I drew 3 curves and emitted from them with no speed (speed had them just flying everywhere with no resemblance to the curve).  I changed the particles to spheres to see them clearly and I was amused that the spheres change shape with the curve , now there are points. I put in this creation expression so that they emit all over

tendrilParticleShape3.goalV = parentV; tendrilParticleShape3.goalU = parentU;

I then put in this creation expression to make the particles stay where they are born

tendrilParticleShape3.goalV = tendrilParticleShape3.parentV; tendrilParticleShape3.goalU = tendrilParticleShape3.parentU;

However they stayed in the air where they were when they were born so I group them under the curve and now they stay on the curve. I put a small speed to give some movement and variation of their position Now I want these particles to emit particles which will live for a short time and thus create a wispy trail. But first I thought I had better see if I can goal the man to these tendrils some way  so as not to waste time if it didn’t work.

I made some poly spheres and grouped them to the end of the curves. I was amused that the spheres also change shape with the curve but they will be invisible and it doesn’t matter. I will goal the man’s head and hands to these spheres And I need to animate the curves to move organically too, found that all I could do was rotate it to change it as I had already animated the scale and translate y so I made a skeleton and bound the curve to it so I can animate it.

Had to break connections I had made trying to animate the curve in the first place and made a skeleton for each curve and bound them. That was a mistake it made me spend time with rigging and animating and it moved mechanically. I found I could select the cvs and move them and key them by pressing S so I could animate the curves without a rig just by pushing points around. Then I had a thought what if I somehow made parts of the curve goal to things and make it move organically, So I am trying that.

I was watching a despicable me  mini movie with a ufo and tractor beams in it  and got inspired to dynamically move my tentacle curves. I created 2 poly cylinders and made them long and thin with 3 faces and 60 sudivisions length wise, made them soft bodies, duplicate and the hide the original, and make it a goal of .2 so as not to be too hard.

I made a sphere into a hard body and applied a force to move it through the tentacles, duplicated it a few times and made a newton field to move the balls through the tentacles not successful but when I attached a tentacle to the field it moved- light bulb moment! I deleted the newton field and attached a mild disturbance field to the tentacles and they move organically and dynamically!!!!!

Then I made them start up in the UFO and reach down to the guy then up again- moving all the time unlike my previous attempts where I found the particles wouldn’t stick together if I moved them too much and lost all recognisable shape.

When I tried to make them emit all over with the expressions every time I tried the bloody things slid up and down from the bottom end of the base tentacle shape. So I decided to try to increase the polys and realised I would have to start again so then I thought I would make these particles invisible and emit from them as they move.

I did start again I made the poly tubes wider and they move and change shape dynamically all the time and emit from the particles too so it has a fine trail. I made it irridescent and partly transparent so it looks like tendrils

Now to make the guy be taken up by the tentacles. The class notes say- Pin constraints constrain two Active Rigid Bodies to each other, or, an Active Rigid Body to a Passive Rigid Body. => The pinning point/pivot point is adjustable and it can be key framed. -perfect, seems just what I want!!! I connected the body parts with pins and hinges and chose spheres and cubes as stand ins, gave them mass and other values and it moves like a man’s dead body in response to fields.

I started again on the tendrils in a new scene and made the triangle tubes again with about 40 divisions lengthwise , made a soft body copy goaled to their original, applied springs with stiffness 11 to let it stretch the right length to the mannikin’s hands and painted goal weights on at the top so the top sticks there and off at the bottom so it falls with gravity or uniform field.

I goaled the bottom particles to the hands I made a newton  field emit from these particles to attract the hand and fiddled with the rigid body settings mass, initial velocity, damping  etc and the goaling of the tentacle soft body to its original and it’s stiffness, strength of the newton field etc  til it moved with the tentacle.

I played again with a rig and an ik spline on a soft body curve to try to get the end of the tentacle to stay with the man without much success.

Then it occurred to me that a collision with high friction may work so I made the soft body of the cylinder collide with the hand and split with friction at 1, and it stuck to the hand but did weird things with large solid areas. So I made it not split or emit on collision, the hands rise dynamically in response to the field and the ends of the tentacles sit on them and do not pull up and off like before because because I deleted the animation on the field affecting the tentacles which had it going down then up, now it just goes down all the time and it is the goaling to the original which pulls them up after reaching down. Because they go up slower than the man rises they stay on him.

When I put the tentacles and cubes back in the real scene  and attached the cubes to his hands they rose at a different rate so lots of days were spent trying to make it all work together again.

Lots of time was wasted with the computer freezing with interpenetration of hard bodies.  Finally I made the joint balls not rigid bodies anymore and hid them hoping that would not be too visible and fiddled with fields to make him go up fast enough. I need to get things rendered because the compositing of the renders is due a week earlier so I am trying to get things working well enough to render.

Working on the man again Maya went stupid once more and changed the position of 1 foot and anchor pint for its rigid body in every scene so I had to start again importing a new mannikin and making all the rigid constraints again. I kept the joints in but squashed them a bit and made them close to a part of the geo that I parented them too, skipping these in the constraints. The man goes up dynamically nicely with no collisions.

Oh I also forgot to mention I made each piece have collisions off except the head hoping that the tentacles can stick to that. The cone. The feedback was go back to the smaller points of the cone, at the start there was a cone with another inside If that is what the client  liked, it was in the original brief,  I made that a soft body with springs and attached a turbulence field to move it because the brief was a non static beam.

I added a plane below the man with a surface shader so it blocks the particles that go below ground. I made the plane and particles collide so none would go under ground. But then it showed up in the render so I didn’t use it just fixing that issue in Comp I am really unhappy that I can’t make the tendrils or tentacles move well dynamically while going down and picking up the man.

Then I noticed the way the inner tendrils in the end of the plait of hair of the characters in Avatar move, they are rooted and the ends wave around and also they are attracted to things and entwine around them. I thought maybe they were using paint effects plants so I have been trying out that.

I have made 2 long tubes come from 1 curve and manipulated, the creation, growth, behaviour, forces and flow animation to make it grow down from the UFO in a writhing manner and then curl up into the UFO with the man. I  made it emit from the geo for a short life with expressions in goal U and V and parent U and V to emit all over the surface.

Then I decided it was too boring and wanted to add some small tendrils like in the references so I added another emitter with a slow rate and limited the number to 6 or so and made them emit and those in turn emit again to have small tendrils coming down from the ship in the cone.

I also randomised the shape of the beam by using soft select on the vertices of the original cone geo and moving various levels in and out and making the animation oscillate because the original brief said he didn’t want a plain cone of light.

experimenting with paint effects plants and the results

I composited it all together with the rest, UFO passes, sky, UFO shadow and backgrounds, added some light glow and brought branches to the front in Nuke.


Here is the movie-


Palamar, T., 2010. Maya Studio Projects Dynamics. Indianapolis: Wiley.

Rickitt, R., n.d. Special Effects the history and technique. s.l.:Billboard Books.


UFO- more

I wrote these notes as I did it. I tried many different things and techniques and I hope it is not too hard to follow.

I have the models now so I am trying to reference them and I have changed the tractor beam. There is a man in the bottom of the beam and he gets taken up into the ufo.

This slideshow requires JavaScript.

Here I am playing with rendering with mental ray, some of the particles in this one are spheres


I decided to abandon the thin inner beam and just stick with the wider cone and tendrils.

Below I have put my tractor beam in a file with referenced models etc. I have changed the size of the beam to match the opening in the UFO but I am having trouble making the outer beam particles be transparent since I made Mentalray the renderer. I have made all except 2 streams of particles visible or dynamic.

For the main part of the beam I tried a polypipe into a conish shape with emit from surface, visibility off

I put these expressions on to make particles emit randomly on the UVs

pipeParticlesShape.goalV = pipeParticlesShape.parentV;

pipeParticlesShape.goalU = pipeParticlesShape.parentU;

pipeParticlesShape.goalU = pipeParticlesShape.goalU +0.01;

I had a bit of trouble working out how to use referencing in maya but I think I have it now.

At first I sized my fx all up to the size of the models and so increased the point size on the particles as points and multipoints, increased gravity and the size of the cone geo and the fields.

After I received the scene with geometry and animations etc I scaled it all down to fit my fx and made the particles live longer to reach the guy. Because it looks better in the smaller original scale.

I shrank the particle tube in to fit the guy’s size and particles to goal to him by scaling in the x and z scale of the newton field which constrains the tendrils and they curl down unevenly around just the man now and I like it.

I animated the start and finish of emitting rates to time with the opening of the doors and then animated the cone to start flat and extend to the ground, then reverse at the end.

I tried reversing gravity to make the tendrils go up with the man like it is pulling him up and goaling them to him then a sphere that I animated to ascend before him.


I want the wider beam to be more like a beam of light as in the references but below I have gone back with the maya renderer so that I could get some transparency on the wider beam so I am using a particle sampler node and reverse node to reverse the info of opacity to transparency on the phong as per the notes.


After trying to get tendril particles to follow the man up with little success I made a cylinder with Y scale at .001 and radius 3 and subdivisions axis 6 so that here would be 6 emitting points and animated that to descend and ascend while rotating slowly. I animated the emission rate to be 1000 as it descended and it created an even spiral pattern. I tried a newton field and gravity to bring them up, uniform field worked best turning it on and adjusting the magnitude to match the ascension. I attached the vortex and they flew out so I also attached the newton of the original tendrils. The vortex was too strong so I tried a low turbulence to mess it up a little. I like it. Now the tendrils descend in an ordered spiral then go wiggly and then ascend with the man into the spaceship. I shortened their life to 5 because they were shooting out the top of the ship.

I have tried a new way of making the tendrils and I have playblasted just them not the whole beam below. Now they go down to the man in a nice spiral as per one of the examples in the brief then they go squiggly and pull back up with the man, (so excited that they go up with him now)

Maybe the client will like the more irregular tendrils. Having worked out how to bring the spiral back into the ship I have applied that method- uniform field- to the original tendril particles and shown it below.

Now for the cone particles of the larger beam, I started them again because something broke in Maya and they wouldn’t work at all and I thought I would try something else. This time I made the cone flat and up in the UFO then made gravity pull the particles down then up to drag them back up and placed a drag field under to stop them going way down. Then I tried removing all the fields and making the flat cone go down then up and leave its particles as it went fading over time so as to finally disappear in the upward direction. I then tried attaching a uniform field to drag them up at the end to make it a bit more convincing. I added a drag field at the top to stop particles flying up out the top after they are brought back up.


I am having trouble rendering so I think I must stop fiddling with the particles and spend the last 3 days trying to render. Liar, anyway I made a turbulence field for the cone particles into a cylinder and trapped the particles inside and had it on low.

I made the particles to be software rendered to have translucence so the cone is cloud and the spiral is tubes, decreasing density makes it more transparent and big cloud cell size makes it all like a glowing cloud of the shape I chose coming down from the opening. I am pretty happy with it. I used particle sampler nodes and reverse nodes to get the opacityPP of the prticle working correctly in the transparency of the particle sampler because they are opposites. The roundness of the spiral particles was ramped down to make it more tendrilly. I batch rendered that and it is very regular and a tad boring? It does however have the same spirals as one of the things the client really liked.


This done I wanted to render the tendrils one too. Having trouble with  maya playing up and not rendering in the render view so it was taking too long. I batch rendered some of it with 8 passes and it works so when I get Maya to correctly show particles and changes I am making I will be able to render. Restarting Maya again!

The tendril particles separate as they go up so I decided to try a drag field and trapping the particles as it scales down in the Y scale and translates up into the ufo. Not real good. I changed it to a new gravity field just for the tendrils and fiddled with the strength til it went up the same speed as the man.

I decided to do just a beauty pass at this stage for speed but I made a separate layer each for the ufo, man and beam, if I had more time I may have done the 2 parts of the beam separately so I could just replace 1 part if needed. I am doing the 2 versions of the beam because they are different.

I had a few goes at rendering because of accidentally not confirming the correct camera and incorrectly setting up the render passes for the layers.

I finally rendered, with MentalRay (which worked on batch renders– discovered with small test batch renders– but not when I hit render in render view) – separate beauty passes for the ufo, man, wide beam and tendrils. I realised too late that I had accidentally chosen the wrong format in render settings so I used reformat and transform nodes in Nuke to make up for that when I composited it and finally I rendered it in Nuke and have a real movie of it at last to hand in for the week 9 progress assessment.

This slideshow requires JavaScript.

It  took so long to do just the one scene that I didn’t end up finishing the rendering and comping of the regular spiral one hence I only handed in the irregular tendrils one, that was all that was asked for anyway, just one and I like that one better.

UFO start

tadaaah see I did do something.

Here are some references that I like which are similar to the ones given with the brief. I want a cone of cloudy light and tendrils like in the images

My first attempt-

I made emiiters and particles to emit particles and changed the particle render types. I used fields such as turbulence, newton and vortex. I  assigned particle cloud shaders  to the particles and changed the colour, transparency and incandescence and the colours of them too.

I think it’s really funny how it looks like it is probing the ground or you if you are under it looking up.


and again-

I have been trying to get a cone within a cone too and so far this is what I have

I did a batch render of 150 frames as jpgs and opened it in Nuke and rendered it there to make a movie to see in Quicktime, uploaded it to youtube and here it is as it stands at the moment. The outer cone may be a bit too faint because it is hardly visible.

dynamics homework week 6- walking on rock, sand and mud


These are the steps I did to make 1 surface react like rock, sand then mud as a guy walked over them:

  • select ground, soft/rigid bodies tab, create soft body options, duplicate make copy soft, create.
  • with the timeline at the beginning select the particles, then soft/rigid bodies tab, create springs.
  • I copied the guy with duplicate special and duplicate input graph ticked and deleted all but the feet and made these collide with the ground then hid them at the end.
  • created clusters to move vertices on the feet up that were going under the ground and causing points of the ground to drag up.
  • set conserve on the particles shape node to 0 to stop the momentum or the continuing movement after the feet have passed on.
  • oversamples, which is in solvers tab, set to 4 for more accurate simulation
  • created selection sets for rock, sand and mud particles.Changed the goal per pp in the component editor for each surface- rock to 1, sand to .2, mud to 0, which makes the deformation occur or not and to varying degrees.  Put the goal weight on the particle shape node to 1 so that it multiplies by these new settings.
  • (Originally I tried to achieve different surface attributes by seperating the 3 parts of the ground geo but rereading Chris’s notes reminded me about the above method).
  • set keys to animate the conserve on the particles to change for the different surfaces as he walks over them- 0 for the rock and sand and .8 for the mud so that it will slowly move back after being disturbed
  • and here are renders of the footprints on the surface types which I have coloured with a ramp



settings on the particle shape nodeImage

different particles goal pp settings in the component editor for each surface type

  • here it is

dynamics homework week 5- waterfall

We were taught the secrets of the waterfalls in Avatar by one who knows using goal weights over deformed geo.



Some references: (I can’t make the movies show on here just the links)





Here is my attempt.

I looked at the references and noticed the way the water doesn’t just fall straight down but sprays out and goes around bumps.

I made the water surface ’emit from surface’ and wrote expressions at creation and ‘runtime before dynamics’ to make the particles emit from random places all over and stick where they emit from and then die straight away if they spawn after U=.05, and to live for 4 seconds if they spawn before, run to the end at random speed and direction and fall off. I changed the time that they fell off from u=.95 to .999 so that it was closer to the edge because some were falling off from underneath a bit too far in and sprayed back, this seemed to help. 

I also made them fade out by a ramp in their opacity attribute and I put a colour ramp in their colour attribute.

This screen grab shows the expressions in the attributes –


the next 2 screen grabs show the runtime and creation expressions in the expression editor-waterfall3


I used a gravity field to pull them down after the edge and turbulence for un-eveness. I experimented with how large the dimensions of the turbulence field should be and thought it looked a bit better in a bit from the edge so that it didn’t spray so wildly out to the sides at the level that it needed to be at to stop horizontal gaps in between the particles but sometimes the edges still look like that so I am stretching it out again.

gravity field:waterfall15

turbulence field:waterfall14

Variations in the turbulence field:

I also made a bumpy wall behind for it to bounce off and made the particles collide with it.

bumpy wall:


particle attributes:


The emitter now emits 100,000 per second, I had it at lower and higher levels but it didn’t either look as good as I wanted or was very slow.

this image shows when it had multi points and 1,000,000 patricles per second and a narrower turbulence field:waterfall16

I tried points, multi points, streaks and finally multi streaks with a short tail and a multi radius of 1.5 so they don’t stick together in clumps but are not too wild in the stream. I gave them a life span of 1 with a random of .5 and the opacity makes them fade before the disappear like the falls in Avatar.

This had multi streaks and 100,000 per second and a longer lifespan:waterfall18

After this I decided to fiddle with the length of the fall and make it vary a bit at the bottom as in the Avatar falls. I tried another turbulence and air and drag fields. I had to keep restarting Maya and also ramp them way up to have any effect. I deleted them and tried a vortex field and made it a sphere which I lengthened up the length of the field and not much wider than the fall. It stirs them up a bit the bottom of the falls.;_ http://youtu.be/YhSxwCMurm0


close up of waterfall- http://youtu.be/eH-EZtZmy54