Story ideas for Lunatics! have arisen from a deep collaboration between the series creators, Rosalyn Hunter and Terry Hancock, but Hunter is the one most responsible for the writing.
She is known for her witty and strong characterization, character voice, and unusual perspectives. A botanist, a molecular-biologist, and a science educator, she has been a lifelong advocate for space development and space agriculture as well as an avid fan of science-fiction, anime, and manga.She cites many creative influences, including writers as varied as Larry Niven, Anne McCaffrey, Tanith Lee, and Robert L. Forward. She was also strongly influenced by fantasy writers like J.R.R. Tolkien, Lewis Carroll, Hans Christian Anderson, and George MacDonald. And she also enjoys the cyberpunk genre, Greek mythology, and classic literature.
She has taught science at different levels, including high-school chemistry and biology, college biology for both majors and non-majors, anatomy and physiology, and botany.
In the early 2000s, she wrote an online blog called “The Space Homestead Corner”, in which she looked into some of the everyday problems that might face people living on the space frontier. Some of those ideas have inspired stories for Lunatics!
This is a steam locomotive, modeled on the one on display outside the Baikonur train station (included in the exterior set). Modeled by Chris Kuhn. Textures and materials created by Terry Hancock.
Development of the character “Hiromi Lerner”, from the silhouette on the character sketch, through concept and model-sheet art by Daniel Fu to the character model by Bela Szabo, and the rigged and toon-rendered character models used in the pilot, created by Keneisha Perry.
In our pilot episode, we have a montage in which Georgiana and Hiromi go through some of the traditions of a Russian space launch from Baikonur. One of these is signing the door of the room they stayed in at the “Cosmonaut Hotel”
This is an in-depth discussion of the animated writing technique using Blender’s “Dynamic Paint” simulation to reveal a pre-drawn signature texture (how I did the signature tests published earlier this week).
Stages in the development of the character “Georgiana Lerner”, from a silhouette that was included with the character sketch to a posed and toon-rendered version of the character used in the pilot.
An improvement over the writing animation from yesterday, with a moving pen added. The point of the pen tracks the same animated curve used to move the brush object for the dynamic paint simulation. This creates an animated texture which is used as a key to reveal the drawing texture.
The top of the pen tracks a different, smoother curve added on top, to create the impression of the pen being turned as it moves. I’m hoping this will more accurately simulate writing mostly with wrist motions.
I will have to see how well this can be integrated with the IK system for the character’s hand and arm.
Test animation (no sound). Blender’s “Dynamic Paint” simulation is used to reveal an already-prepared texture of the signature. The brush may be a little too big, as it’s revealing some strokes before they would actually be seen (particularly with the star). Still working on it! I suspect I can improve things a bit by animating the brush object some.
The drawing path will also provide a method for animating the marker pen and hand, I hope.
I plan to explain this more fully in my next Production Log.
The opening begins with a shot discovering Georgiana on the train. That scene ends with the camera pulling away from Hiromi and watching the train disappear down the tracks. But this is first we’ve started to work on that shot, and here’s an early render (no Freestyle yet).
It has now been years since we published a storyreel animatic for “No Children in Space” after our “pre-production” Kickstarter. That was originally conceived as an hour-long pilot episode.We’ve now broken that into three smaller, more manageable pieces (variously labeled “Part 1”, “Part 2”, and “Part 3” — or if you like, “Prolog”, “From the Earth…”, and “…To the Moon”), as well doing lots of trimming. And of course, a great deal of what we set out to create has been modeled, and some parts of it exist in various states, ranging from “blocking” (which just means figuring out where the characters, cameras, and lights will be) to “fully rendered” (which means, the shading, lighting, ink lining, and other details are all in finished form).
Keneisha Perry and I have discussed the animation after having produced a fair amount of it, and we think there’s a good chance that the animation for the “Prolog”/”Part 1” can be finished in December or January. Keneisha has nearly completed the “Press Conference” sequence, and is already started on animating for the “Touring Baikonur” sequence. She got a start on the “Train” sequence at the beginning some time ago, although that needs a lot of work. I’ve hardly started on mech animation, although it’s generally quick to do (thanks to very good modeling and rigging work by Chris Kuhn).
That still leaves quite a bit of work: detailing sets and other models, rendering (no small task in itself!), and a complete revision of the audio production (we’ve used Audacity so far, but will be migrating the whole chain to Ardour as well as adding additional ambience, hard effects, group ADR (possibly), and foley sound to scenes that need it.
With all the changes, it seemed like a good time to create a new “previz” animatic. There is much work remaining, but we are way past the storyreel.
We are about two-thirds of the way through with that — in the edit pictured above, you can see about 8-1/2 minutes of story plus another couple of minutes of credits. My guess at this point is that the final edit will be around 12 minutes of story with 3 minutes of credits.
Most of what remains for the animatic will be a lot easier after I finish some more of the mech re-rigging, so I’m going to complete that first. That will help with the “Launch” and “Flight Into Space” sequences (there’s also a “Suiting Up” sequence missing, but the issue there is mostly that I want to make changes to it, so I’m rethinking some of the shots).
Yesterday and today, I’ve been learning how to work with “Dynamic Paint” in Blender (this is a feature in the Physics panel).
I’ve never done anything with it before now, but it seems to be the ideal solution for a couple of shots where the character is writing something, and we need to see the words appearing on the page.I was going to post something about this today, but I’m not there yet, so maybe next week!
Neat picture, though. Still wrong. But neat. 😀
Things are going well, though. We’re now doing various work on the “Touring Baikonur” montage. We have about 2/3 of a previz animatic for the first part now.
I am weighing whether to put the completed previz animatic up for Premium-and-above patrons. It’s still very much WIP but also kind of “spoilers”, since it shows the composition for the whole episode. I’d be interested in reading comments on that if you have an opinion.
Blender has a number of different ways that data can be linked from one Blender file into another. Blender files can contain multiple scenes, and objects can be linked between the scenes so that, for example, an animated object can appear in both scenes, while each scene might have other, unique objects, such as cameras and lighting for different angles.
There’s also a nifty option which allows you to set a “background” scene. This is very handy for putting the action on a set, while keeping the set out of the way for animation (because you can’t accidentally select or modify the set and it’s easy to turn it off and on).
Also, I don’t know how widely this is used, but scene backgrounds can be daisy-chained — the background can itself have a background, and so on (all of them will be included in the final render).
You can achieve some pretty nifty effects with this.
Yesterday, I shared a screenshot of editing the blocking version of this shot, in which Georgiana starts skipping on one set and then skips right through the transition onto the next set, without breaking stride.
To do this, Georgiana is actually animated all in one scene, while the two different sets are linked as backgrounds.
But we have to actually generate two shots, with some overlap for the transition that we will do in the editor (Kdenlive). So, each camera also has its own scene, just as we did for the “multicam” setup in the Press Conference.
In order to keep these two shots perfectly synchronized, though, there is only one camera rig, shared between the two scenes with “object linking”.
The characters are “group linked” (or “dupligroups” as some Blender documentation calls it), with their armatures “proxied” for animation (this is just the normal way characters are animated in Blender).
Group linking has a unique advantage over other methods, in that the group can be positioned, rotated, and scaled in the scene (there is a control “Empty” that provides the coordinate transformation).
Putting all this together does start to get complicated, though!
As you can see, these two shots are created from eight Blender files, starting at the lowest level with the backgrounds and sets for the two locations in Baikonur. There’s also a “relocated” set — which is to say, the set was group linked like the characters so that I can move and rotate it to match position with the other set (because we need them to be lined up just so).
We keep the character animation in one file, while the camera movement and lighting is in another file, and sets and characters are in still other files. The primary reason for this is “separation of concerns”. Since it’s quite hard to merge Blender files together for version control, it’s not good to have two people working on the same file at the same time. This way, Keneisha Perry’s character animation is isolated so she can do pretty much what she wants with that file without having to coordinate with Terry Hancock who is modifying the cameras and lighting in the “camera” file. Similarly, Keneisha is most likely to alter the character files, while Terry is most likely to modify sets, and these are also in separate files.We only really have to coordinate when we’re signing off on the shot to render it.
This is one of the more complex examples, but the same principles are (or will be) in a lot of our shot setups.
Had a little inspiration on how to make the montage of Hiromi and Georgiana touring the town of Baikonur much tighter with some unifying camera direction work. So I’m re-blocking this sequence, positioning things a little differently than they were in our storyreel. This is a lot of fun to do!
(The red markers in the shot are character locations — literally just blocks for blocking).
Recently, we decided to move more of our incidental work-in-progress postings to postings here on Patreon. A lot of these had previously been going up on our Facebook page, with only more complete and finished work being published here on Patreon. We’ve decided the opposite is a much smarter use of our social media time, especially since the newsletters had to be discontinued.
We have certainly not been idle. Most of 2017 and some of 2016 was dedicated primarily to infrastructure work, which have been documented in a few posts on our hub site ( http://lunatics.tv ). We’ll put some details about that into separate posts.
During that time, Keneisha Perry streamlined her character animation workflow, by converting the character face rigs shape-keys to bones, and over the next few months, completed most of the animation for the two major character animation scenes in Part 1, the “Press Conference” and “Opening/Train” sequences. This leaves a number of much shorter animation sequences to complete the character work.
Also during the last year, Rosalyn Hunter completed a working draft of the third “Lunatics!” script, for “Cyborg”. Hancock and Hunter have also been working on story development for the fourth and fifth stories, “Rocks” and “Death’s Door” (which are linked, and mostly to do with introducing Sarah Allison, the colony’s resident geologist). Which of course, puts the writing well ahead of where we are in animation production.
In October, Terry Hancock opted to get back to production work for awhile, and most of the posts you see going up lately are from that work. All of the sets for “No Children in Space – Part 1” have already been modeled. Most of what Terry is doing is to apply materials, textures, lighting, rigging, and other minor detailing — collectively “finishing” those sets.Keneisha Perry and Terry Hancock think it will probably be possible to complete principle animation on “No Children in Space – Part 1” (a.k.a. the “Prolog” a.k.a. “Pilot Mini-Episode”), by sometime in January, possibly early in the month if things go well (and lately they have been).
We’re getting the hang of the workflows we’ve established for animation: the multicamera scene-based workflow is both functional and faster than shot-based workflow, as we had hoped. Papagayo (now Papagayo NG) is working for lip-sync. And we are able to cut together complex character scenes.
Terry Hancock has been learning to convert Chris Kuhn’s contraints-based control panels into armature objects, which makes proxying them into animation files easier for mech animation, as well as doing some additional work on effects (mainly flame effects).
He’s also finished the “Aquarium Room”, “Mission Control Room”, and “Soyuz Interior” sets. The “Traincar Interior”, the two monument sets (“Glory to the Conquerors of Space” and “Stele to Science and Space”), and the “Train Station Exterior” sets all need some minor finishing work, but are quite close to finished. Most of the other Baikonur sets are still pretty raw — just models without much dressing or materials work, so they still need extensive finishing work (this includes most of Travis Souza’s set models: “Train Station Interior”, “Cosmonaut Hotel Interior”, “Cosmonaut Trees Exterior”, and “Bazaar Exterior”).
The “Launch Pad” set still needs to be relinked with the corrected version of the gantry model and the on-the-pad version of the rocket has to be updated to match the flight version. All the mech models will also need the same sort of rigging conversion that has been done on the Soyuz Interior set (“Launch Pad”, “Soyuz Launch Vehicle”, and “Soyuz Orbiter Exterior”).
And finally, there’s a few props missing: the ornate box for Georgiana (as well as the others on the shelf), the spacesuit crate and empty spacesuit, a brush for Georgiana’s hair, and so on.
Over the next couple of weeks, we’re going to be blocking out the rest of the animation for the episode. This should result in a new pre-viz/blocking animatic for all of “Part 1”, with an updated running time (probably still about 10-12 minutes).
After that step and after nailing down our infrastructure problems, we will probably be in a position to set a firm release date. The biggest unknown is whether the IT work and the Render Cluster build are going to work out, and how much time finishing them is going to take. It’s also very clear that some more funding is going to have to be found to make finishing Parts 2 & 3 possible on any sort of reasonable schedule (meaning “less than the four years Part 1 has taken”). We have a few ideas about that — most of which involve Part 1 being finished. So that remains our top priority.
I’ve gotten requests to publish a list of free-licensed resources used for Lunatics, and I have accumulated a very long list of them in my browser bookmarks collection. I don’t really have time to edit this down into a more digestible format, but it wasn’t too hard to export the bookmarks and import them into my news blog here.
Repost of a useful links list from the Production Log.
Soyuz-SF (Descent Module) Interior Set For “No Children in Space” (will appear in Part 1/Prolog and Part 2). This is an updated and almost entirely rebuilt version of the model that appeared in our first teaser trailer. Chris Kuhn modeled almost all of this, although it is based on the earlier model created by the team in 2012.
[NOTE (2024): Our ResourceSpace service has been disrupted since this was written. This model, prepared for download, is likely in the “exports” directory of the Production tree.]
The new interior set for the Soyuz scenes in “No Children in Space” has been completed and uploaded into our Subversion repo and ResourceSpace DAMS (the link is to the ResourceSpace page for the Blender file).
By the way, some people have been spooked by the “license agreement” that ResourceSpace puts up. This is actually our upload terms, and it basically says you can’t add content to our DAMS without it being compliant to our Creative Commons Attribution-ShareAlike license.
Most of the content in our DAMS is under CC By-SA 3.0, including this file.
View of the periscope from the pilot’s point-of-view, with (probably!) final materials and lighting. The periscope can be dark (the periscope is retracted during launch and shows nothing then), show an Earth view like this (this is Timothy Giet’s Earth limb painting again), or one of a few animated sequence created for the rendezvous in Part 2.
Work-in-progress preview of the effects for the retro-rocket used to be used in the landing sequence. Of course, our lander uses LOX/LH2 propellant rather than the hypergolic mixture used by the Apollo LM, which means the flames will be a different color — not sure I’ve got it right, but I think they will have a more visible glow like this.
This video includes the effect, a wireframe GL render, and an animation of the node-tree as the textures animate through the shot (the textures are shown in addition to the flame effect itself).
This was done almost entirely by animating “cloud” procedural textures on modeled flame envelopes. It’s a pretty simple effect, but I think the result is more realistic than what we’ve been able to do with simulation approaches.
The Periscope is a little more complicated than the other screens, because I’ve provided a means to select from several different videos (the selector nodes are in the framed box), and because the result is composited with the reticle. There are two rigged parameters for this material: one “selector”, which determines which texture is seen, and one brightness control that balances this against the darkened screen.
The result for one set of parameters is shown in the lower right (this is actually just one frame out of a video, that shows the Earth slowly rotating around the frame as the spacecraft rolls). Other possible videos include the approach to the space station and the final docking approach.
An overview of the materials rigging on this set is available on the Production Log.
Earlier this year, Keneisha Perry came up with a way to rig the character’s faces without using shape keys, and this is a lot easier for her to animate. Most characters have already been converted by now, and the existing character action scenes have been converted over to this method.
This approach is also/still supported by Papagayo, so it doesn’t make a very large impact on our workflow, other than simplifying the refinement to the animation.
I think this is just about the last thing: cabin lights. I almost forgot about these, but of course, they needed to be in the control rig too, so they are now.
The two red lights in the back provide dim lighting which doesn’t interfere with night vision, for better visibility through the windows and periscopes. I don’t think current Soyuz has these, but it seems like a good idea, and it looks cool. The lights are assumed to be LED-based, which is another change.
This angle will probably not ever be in the show, it’s an extreme wide angle shot looking straight up from the middle (pilot/commander) seat.
Next comes setting up an animation file. Hopefully, all the rigging will continue to work when the set is proxied. We shall see.
Scheme for classifying assets based on how they are used.
Migrating data from one system to a radically different system typically requires a lot of thought and reorganization, and this post is about the plan I’m formulating, which I hope will be workable. We still won’t want to save everything, but on the other hand, with TACTIC, we will be able to practically manage a lot more assets.
Some thoughts from our Production Log, regarding how to handle several different types of data to be migrated from existing sources into TACTIC, and how this will affect our requirements.
Still not quite done here, but the materials are looking better, and of course we need to make sure Hiromi’s perspective works, too. I remapped the labels with some proof-reading edits, as well.
One of the features of the new set is that the hatch opens, and I’ve added some detailing to the back side of it. Clearly, our intrepid colonists must be able to get into the spacecraft!
This is the new control rig for the Soyuz set. The main mechanical aspect is the acceleration couches, which are mounted on shock absorbers. This panel allows each one to be controlled separately, or all to be moved at once. It’s adapted from Chris Kuhn’s original constraints-based rig, but uses an armature for the control panel, which not only makes it not render by default, but also much easier to proxy into an animation file (because it’s all one object).
The other controls on the panel are for the indicator lights and video displays on the Soyuz console.
