Monthly Archives: April 2013

Cambot pt4

Untitled

Continued from here

 

I’ve made progress on Cambot. I drilled up some stock aluminum bars to fit between the gearing and the rail mounts so now it is all secured together. I also swapped out one of the axels for a carriage bolt so I’d have a way to tighten it down.

So far I’ve worked out the pat/tilt mechanisms and it seems to be working pretty well. I’m driving the two DC gear motors via a Dynamic perceptions AT2 controller.

I’m still a bit concerned about the overall weight and size. I wanted to avoid unnecessary strain on the motors, so I took the camera off the mount while testing it.

Here it is fully assembled;
Untitled

Sadly immediately after this test there was a ‘pop’ noise and since then the 12v DC adapter I’ve been running this from stopped working. This was a new power supply, purchased with my AT2 and used only 4-5 times no longer than an hour each time, and and it was hooked up to a surge protector, so I’m at a loss to explain what happened to it, manufacturing error perhaps?. I have no way of testing to be sure it didn’t fry the controller or motors, I can only assume(hope) that is as far as the problem extends. I’ve got a replacement on the way …

I keep reminding myself that I started this project to learn what goes into building something like this more than I did because I couldn’t afford a packaged system.

Continued here;

http://bradisdrab.wordpress.com/2013/10/08/cambot-pt5/


Gynoid Robot Doll #5

This week I’m sending off one of my ball joint dolls, the Robot gynoid, the fifth of the series.

IMG_5422

Its always sad to see them go, but I’m glad she found a good home, and I will be casting a new round of dolls very soon.

IMG_5420

This was available through the Zerofriends store, where more will be posted early next month. See more photos in the full gallery; here;

You can see my work in progress ball joint dolls in the flickr Gallery gallery here;


IClone Matte painting set-extension

I did this as a bit of a quick test to see how possible it’d be to create a digital set-extension element.
wasteland_matte_extension

I spent 4/5 hours the first night, and another hour last night using iClone and a few models from the 3Dwarehouse. There was no real time spent texturing yet, Just the stock diffuse map colors that the models had and a couple just a single light lots of ambient occlusion and basic IBL, dropping models in position, and setting up the camera.

I even took the camera for a quick zoom into it to see how close I could get before it looked… not so great. Yeah its pretty mid-90s video game, but better textures and lighting will enhance that a lot. Here are a few stills too, try to spot the Easter egg!

wasteland_matte

wasteland_matte0

wasteland_matte2


Shooting Miniatures; GoPro

I’ve got a follow-up to pick up where I left off with this post; Advanced Camera; shooting miniatures

I did some thinking and talking and testing, and it seems you can only get so close with a scale model before the scale becomes apparent. The camera is essentially the viewer and the lens’ size dictates things, like camera angle, and how close you can get to the ground (table surface) etc. For example in this test I was using a 35mm camera, even with a macro lens and exaggerated angle there is only so close I could get.

However, this has got me thinking; what about smaller Cameras? The new GoPro Hero3 black can record very high quality video, and it’s size is just insanely small. Makes sense that you’d want to scale the viewer along with the scene. Could it be the best way to shoot scale models? We shall see.

But this opens a lot of questions; What about the wide lens distortion? How bad/noticeable is the auto adjust for exposure/color? what is the focal distance, and how close can i get to the model before the image is unusable?

Fortunately, I have awesome friends and one lent me his to test with 🙂

What do you think? I’m impressed. Obviously I pushed it out of it’s confort zone. I wanted to see where it’d fail, so there is a bit of too-close, to-wiggily, etc. but overall it performed better than I expected and I may need to pick up one of my own soon…


Cambot part 3

Continued from; Cambot pt2 (BRAIN storming)

I’ve been thinking a lot and working on my Cambot. I try to keep the end goal in sight as I am teaching myself just enough to make this work. In the end I’m hoping to be able to set it up for both live action and StopMo, on a stage and out on location. So it needs to be somewhat modular and self-contained in the end, so not (solely)controlled by a computer.

Untitled

I’m not much of an engineer, more of a hardware-hacker and I’m basically teaching myself this as I go along, which is partly why I’m working it out in parts. I’ve also not got access to a proper machine shop, so I’ll either be at the mercy of my friends, looking into buying a mini mill, or attempting it with hand tools.

But first some inspiration from Vimeo user Displacment1, who has some really amazingDragon Stop-mo MoCo rigs;

I’m really liking these designs, modular and adjustable. As much as I like the simplicity of my Geared tripod head I think I’m going forward with this sort of a nodal pan/tilt head design.

Mohan has made some really slick 15mm-based pan/tilt/slide rigs. He appears to be using carbon fiber 15mm rail rigs, so it’s both light-weight and strong! which he has custom built using stepper motors and custom gearing;
Another similar 15mm pan/tilt head was found on the dynamic perception’s forum here. He says he was able to assemble the whole thing for $450, which seems low to me, especially after seeing other offerings.

The miniengine seems to be a really cool way to run time lapse, or shoot-move-shoot based moco. It seems to be well-supported and designed, and actively developed. The next version seems even better, definitely something to keep an eye on…

BFG motion control seems like a great system. They provide a hardware encoding box and GUI interface for designing complex shots. However their price of $1000, which is reasonable for something of this kind, is outside my budget for this project. I’m really hoping to put the whole thing together for 1k, and there is a lot of hardware to put together…

Cambot Plan

I’m getting close, my research and brainstorming seem to have brought me to a clear path. Well I say one path, but its more like a three phase plan.

      -Realtime mode
      -Playback mode
    -Program mode

Realtime mode – Hardware rig

The first step is assembling the rig, the real hardware of it. That will show me the physical limitations of the rig; size, weight, as well as the number and positions of axis.

This is where I am now, essentially a manual stopmotion rig;

IMG_1196

cambot’s geared pan/tilt head

Bogen geared pan/tilt head, mounted to Jib(rigged to raise and lower with an automotive jack), on machinist’s rotary table and linear slide. Current arrangement is set up for manual moments. Turning the wheels incrementally of the geared head and cross slide allow for an animate-able camera.

cambot’s cross slide x/y axis

Pros/Cons; It works… mostly. It is limited to either stop motion animation for camera moves, which takes a lot of patience and focus for long, smooth camera moves. To move increments as well as animating the puppet or prop actually in the scene, or for an actor I can release the jack and move the jib smoothly, for hand-operated movements but its a bit wobbly, and requires an operator.

On to the next phase!

Playback/Live mode

The next phase involves motorizing it all. I’ve got the motions plotted out, but it needs to drive itself, So I’m beginning to add gears and motors now. The goal is to rig it so that it can be remotely operated. (with speed controllers). I’m looking into offset gear motors, DC motors, and gearboxes. I’m planning to work out/test 2 axis at a time in realtime, and possibly set it to record/playback once that is working.

I’m starting off by setting up my hardware to run off an Dynamic Perceptions AT-2 controller, starting with ez-swap dc motor or a a gear motor like these or a DC motor like these.Which will get me the chance to set up all the mechanical parts of the rig properly.

Untitled

This real-time controlled live action remote rig gives me the ability to control and test two axis of DC motors.
Here is my rig as it currently stands, 2 geared axis hooked up to a joystick

From here, an idea for self-contained playback in the mean time, this servo recorder, seems like it could record up to 3min of motion for 4-axes, and with speed controllers it could control DC motors, then play back recorded motions.

This really seems like a good place to get, since it seems like I’d be able to control pan/tilt/focus and movement, either; slide/track/jib/rotate, however it’ll be $3-400, for it and I’m still not 100% sure it’ll work reliably/repeatably… DC motors aren’t really known for their detail in movements, and while my gearing may help I just don’t know, really people tend to use other types of motor for this sort of design, i.e. steppers or servos.

Program mode MoCoBus Rig; stepper motors, nanoMoCo, upgrade device…

Once the mechanical parts are working for live playback, I’ll be looking into swapping the DC motors for steppers driven by a computer, or potentially some portable brain. Ideally I’d have a magic box which allowed me to move the rig however I wanted, then have it repeat the move and export a 3d camera move which could be imported into effects software.

This post from Dynamic Perceptions got me very excited. There are plans to take the AT2 and use it to communicate with record-able repeatable systems. Naturally it wouldn’t work with DC motors.

If I rig my mover with stepper motors, like these Phidgets or NEMA-17,
or  another stepper, I’d expect I need some sort of driver card for each them. I’d need something like the easydriver. I don’t really know a lot about steppers, yet. But I’ll learn.

NEMA-17 stepper

Currently I’ve got my sights set on Dynamic Perception’s MoCoBus, driving the stepper motors via their nanoMoCo cards. The whole system chain (MoCoBus) will feed into a computer, (potentially a netbook) for movement recording/edit/playback, StopMo, etc. (Software Workflow)The upgrade device will also allow me to drive the steppers and record the AT2 joystick’s movements into the MoCoBus chain.

Then when the upgrade becomes available I can upgrade my whole system to use MoCoBus. Then I would need to swap the DC or gear motors out for steppers to use the nanomoco on,

Itd be great if I could rig all my motors with nanomocos as drivers, control 2-axes of them analog for now. Once its all working smoothly, I’d plan to delve into chaining all the axes via mocobus and get more complicated multi axis moves through Graffik on the computer. And hopefully by that point I’d have access to the adapter and this updated workflow.

Continued here;  Cambot pt4