I’m putting together a series of short movies illustrating astronomy concepts. The first one illustrates a pulsar in action. I’m at the point where I’m looking for some feedback and some pointers on how to clean up a few aspects. FYI, I’m a relative noob to Blender but I’m a professional astronomer.
I’m posting on putfile.com right now, for lack of a better idea. Suggestions? (Putfile supports images and movies, is free, but won’t allow download…would prefer something with more flexibility)
I’ve written the file out as DivX codec AVI to save space:
http://media.putfile.com/pulsar-movie-DivX-codec (800x600, 750 KB, 10 sec)
The video begins by showing a neutron star (white ball), then superimposes a rotation axis (blue arrow) and a magnetic field axis (green arrow). The magnetic field axis is mis-aligned relative to the rotation axis. Magnetic field lines are then shown (green curves) and the whole system begins to rotate. Emission from the poles of the magnetic field appear as glowing white cones. Then the axes and the magnetic field lines fade out, leaving the standard “lighthouse” model of a pulsar. The camera moves until it looks down onto one of the pulsar’s poles. The video ends with the pulsar producing a few pulses.
What I see that needs work:
The star background (made with the “Stars” button) is ok as long as the camera is stationary. With the camera moving, Blender appears to re-build a new star background for every frame. This produces a mess rather than a smoothly changing background for the pulsar in the foreground. Any suggestions for producing a better background?
The last few pulses are not nearly bright enough. I’d like to have the flashes be more obvious. I’m using spots with halo to produce the beams of emission and I’m lighting the surface of the neutron star in sync with the rotation of the magnetic field to produce the last few pulses. Is there a way to produce both features with the same lighting? I tried turning up the energy of the spots when the camera was aligned, but I got no obvious flash so I used the neutron-star-lighting method.
The overall timing needs adjusting. Sometimes there is not much happening, other times it’s too busy.
Any comments would be greatly appreciated.
Hey astrogeek! I’m making astronomy objects in Blender too :). Our animation studio is working on a Planetarium show that will involve pulsars, quasars, black holes, and even the big bang! I’m sure we’ll be a lot of help to each other, because I’m a professional animator with an undergraduate degree in astrophysics, so I often need science help rather than 3D help.
I liked your animation, but I did think it was a bit too quick. We try to slow things down as much as possible when we’re communicating ideas through 3D animated diagrams. The times when not much is happening feel like the right speed to me, because it let the information settle into my brain before the next thing happened.
The problem with the stars is one I have hit before. It’s not that they’re regenerated for every frame, it’s that when you’re rendering at a low resolution, the stars can get so small that anti-aliasing makes them fade up and down per frame. Sometimes they disappear, sometimes they’re at half brightness, etc. It’s almost a twinkle effect.
I don’t have a solution for this except to render bigger frame sizes or make your starts big. I’ve scripted a realistic starfield for our show based on real life star data, and I had to deal with this problem a lot. Luckily, planetarium shows end up with frames that are 3200 by 3200, so it’s not a big problem
So maybe you can get by with rendering your frames bigger, then reduce the size of your final video in a compositing app, or quicktime pro, or something that handles resampling each frame well.
With the flashes, perhaps it would help you to puff out some particles on each flash, which would make it look like it’s glowing a bit brighter? If you use a halo material on the particles with xalpha turned on, it’ll look VERY bright.
By the way, here is an earlier version of a spiral galaxy I’ve created (sorry about the watermark, I just don’t want it linked around the place). Since this version I’ve made some improvements to the shape of the central bulge and added globular clusters, but you can see the kind of thing I’m doing in Blender at the moment.
Today I’m working on making an elliptical galaxy, which I can do in several ways but they all have problems when we render out a cubic map for the planetarium format. When halos get close to the edge of the render they fade off at the edges. Volumetric particles would help me SOOOOOOO much right now!
I’m excited to meet you and glad to know there are other folks working on astronomy topics and educational material that use Blender.
I’ve intentionally compressed the timing of this version of the movie so that it’s small. I didn’t realize how small it turned out, so I’ll work on the proper timing now. I agree that slower is better for the brain to absorb the information presented.
I thought maybe the stars were blinking due to anti-aliasing. I’ll do some experiements with camera motion and see if I can fiddle the star sizes. Thanks for the suggestion.
I also thought of some things to try with the flashes too. I just don’t see the neutron-star-lighting working very well. Maybe tuning the spot lamp to flare in sync with the rotation by increasing energy and halo intensity as the beam sweeps past the camera.
I really like the spiral galaxy image. I’m also planning on making movies of differential rotation and accretion disks around stars (new and old). Both of these really need volumetric particles, but you seem to have hit on a good recipe for the spiral. Do you have some hints or maybe a blend file I could look through? The blue component of the spiral arms is really good looking.
I’ve found that lots (1000’s) of halo particles with very low alpha (0.02) produces a smooth looking volume of dust and gas. But I’m having the problem that I get artifacts in the particle “cloud” that show the underlying geometry of the emitter. For example, a squat cylinder emitting particles to create an accretion disk produces particles that form a distinct ring at a fixed radius from the cylinder. The render shows a definite non-random shape for views at high angles over the disk. I’ve tried all kinds of random setting, but can’t seem to get rid of the “rings”.
Anyway, great to hear from you. I look forward to posting more WIP for your comments.
I’m sorry but I can’t share the blend files, since this is for a commercial project and we want to profit from it as much as we can so we’ll be able to make more films. So ending up with a library of assets that we can use in our planetarium films or license to other people will be important to us as a studio. But I can give you hints
I actually hand modelled the particle emitting geometry for the spiral arms, and did lots of fractal subdividing on it, and then deleted random vertices to the point that I basically “sculpted” the star-forming regions and other parts by hand. I don’t think there’s an easier way to get rid of the regularity of the emitting mesh than that.
There are truly no shortcuts for this stuff. I’ve decided for the ellliptical galaxy I actually have to write a python script to create vertices in exactly the distribution I want them so I can just put a halo material on the object and there will be a perfectly formed elliptical galaxy.
If you’re doing lots of particle systems, look out for the bug I ran into where they are rendered in the order they are formed, rather than their distance from the camera! This means even if a particle system is obscured by another one, it will be rendered as though it were in front of it! This has apparently been fixed in version 2.42, so in a few days neither of us will be bothered by it anymore
I understand about the blend files. No problem. I suspected that would be the case.
The regularity solution sounds interesting. I’ll try it out soon.
I didn’t know 2.42 was coming out so soon. This will be my first “transition” to a new version, so I’m hoping the interface and functionality isn’t too different. Reading through old posts, it seems several buttons get moved every version. I am looking forward to learning the node system, though.
I’ve been tempted to pick up a bit of python programming myself. Mathematical models of some of this stuff would be handy, but so far I’ve done ok with hand-modeling from images, etc. Do you have a favorite python reference (web or paper)?
I learned a lot from the big Python tutorial on python.org:
It’s really quite easy to understand. But looking at other people’s scripts (for Blender) is a great way to get started!