DART Discoveries

After much frustrating work, discovered how to do synchronized audio/video textures in DART. Also got positional cues to work for a single tracker near static/actor "zones" on the Mac but still haven't had success doing so with 2 trackers at the same time. I emailed the developers at Georgia Tech for help on this. They say they'll look into it.

Modern Markers

Started creating the markers we will use for the modern period.

More on Phicon internals and the "Magic Light"

The high-flux LEDs arrived today for the internal lighting of the phicons. Discovered a nice, lightweight, small battery with a high capacity to use as the internal lighting power source for our Phicons. The CR2477 coin-type battery is 24mm x 7.7mm, 10g, and has 1000mAH capacity equating to some 9 hours of use before going dead based on 120mAH used by the 6 internal LEDs. However, even so, we calcuated that we would burn through these batteries very quickly during troubleshooting and testing and during presentations. At $25 per set to power all the phicons, this quickly looked bad--on the order of a few hundred to several hundred dollars of our budget just to light up our phicons over the course of the project. We then got the idea to place a 3-wire Magnetic Reed switch into the lighting circuit and hooking the LEDs up to the "Normally Closed" contact. Then we would place "rubber magnet" material in the bottom of our table's tray areas. When the reed switches came near the magnetic material, the contacts would then open. Thereby, the phicons would turn themselves off whenever one intuitively placed the phicons in their resting trays. Granted this would only work with certain cube orientations unless a more complex system were installed, this is an infinitely better solution than having to constantly open the cubes up and tearing everything apart to switch them off to save the batteries every couple of minutes when working back and forth. It will also be a cool factor for the user to experience.

A sample battery holder for the CR2477 was also ordered for installation in the phicon prototype.

More Mayan Symbol Markers

Finished drawing the symbols of the Mayan Tzolkin Cycle (days). In addition I drew the Mayan Haab (month) symbols.

Camera Specs Again

A thought occured that we might not want the specs we thought we wanted for our AR viewer camera. We did the math originally to establish specs for a camera optimized to created a more seamless "magic window" effect for average eye-to-viewer and viewer-to-content distances. However, my feeling is that people will tend to want to "close in" on the content to experience the AR in greater detail at times. With our current camera specs, markers would be clipped from view many inches before the viewer got near the content, and in many cases, the content may even be offset from the markers to begin with causing even greater problems with the markers being clipped from view (and hence AR content cutting out). For this reason, a wider FOV camera is probably a much better choice than our "seamless" spec'd camera. However, around 90 degrees FOV or so, the content would be so small and dwarfed within the wide angle view that it would really detract from the effect in my opinion. Therefore, new calculations seem to indicate a nice compromise of around 78 degrees FOV. We will restart our search for such a camera. We might already have a few models logged that we previously discounted.

Mayan Symbol Markers

Worked on the marker set for the mayan time period. I have drawn most of the symbols from the Mayan Tzolkin cycle.

Rethinking Things

After having done our little tests with the gamepad in the Grad Lab, a thought occured to me that the gamepad controller we had previously picked out may not be right for us. In attaching the display to the controller, too much weight would be placed forward of the user's hands making it not very comfortable to hold for long periods and somewhat cumbersome. It would also not be weighted and shaped right to easily lay down. The Microsoft Wingman that we used in our tests would allow for the display to be mounted more in-line with the controller. It had long enough handles to be comfortably held and operated if the display was attached in this way, balanced more directly over the hands. This would also make it the perfect shape to easily set down flat upon a table in a stable position. We should look for a similar type controller, and it is likely that such a simplified controller (compared to our previous pick) will come already configured to be used wirelessly on the PC/Mac.

Also, prior to actually laying down the money for our AR Viewer camera, we did a little more research to confirm that it met our specs as we believed. We found some conflicting info. We will investigate further.

Research, Tabletop, SaveKey

Did more research for the AR viewer camera, TX/RX system, and uhf transmitter. We checked our AR viewer (tv) to see if channel 59 was available as one of the uhf transmitters we were looking at broadcasts to this channel. It appears to be available.

We picked up the acrylic tabletop.

Andrew showed us how to use a savekey script to make the keyframing process of animating in Softimage more animated. I also got from him a reference sheet for the walk cycle of a horse.

Research - AR Viewer Camera

Started researching camera and TX RX options for the AR viewer. While we wer doing this Stephen had the idea to lighten the viewr even more by using a uhf transmitter so that we will not have to attach a video receiver to the AR viewer.

Ghost in the Shell

As per Scott's recommendation I rented "Ghost in the Shell 2: Innocence." It has some stunning 3D graphics.

Communication and Control

1. Investigated positional cues in DART and discovered that the Mac couldn't track two markers separately at the same time. This did work on the PC, however.

2. We also figured out how to send data between platforms (but not DART's supplied cue-system) and what information can be sent.

3. Figured out how to augment video with live textual data (including model 3D position and rotational information).

4. Researched various methods of accessing gamepad controller data from within Director and tested some to little avail. We decided that using a keyboard wedge-type hack like "Game Companion" might be robust enough for our project and easier and faster to implement. We test this by hooking up a gamepad, installing "Game Companion", and sending gamepad controller data from one computer to the other successfully. Related to this, we figured out how to hide the cursor to keep it's movements from distracting from the AR display while using Game Companion since it emulates a mouse with the gamepad in some instances.

5. We also researched more AR cameras and narrowed our choice to 2 models (including the 402c) based on our specs.

Interface Design

Song and I spent most of the day discussing and going over tabletop interface designs and functionality. We came up with start and end schemas that fit into our idea of a puzzle that invites exploration and solution using intuitive cues. Initially, colored squares matching the colors of the Age Phicons will randomly flash on the tabletop surface encouraging the user to place a colored Age Phicon onto the tabletop. A single Age Phicon will change the whole table top surface giving an overview of the age along with an invitation to explore this age/culture's relationship to the Cosmos. This content would orient to any angle around the table based on the orientation of the Phicon so anyone could view the content right-side up from any direction that the table was approached.

In addition to the 4 Age Phicons, 4 tools (Earth, Sky, Sun, Planets/Sea?) will cue the user to use them by projecting jigsaw puzzle-like silhouette graphics around them on the table where they are placed. Intuitively, one would realize that the tool pieces should be "fit" next to the Age Phicons, which have the inverse jigsaw patter projected around them. Each tool would reveal an aspect of the culture's relationship within the tool context. A small content window displaying content in either Flash or video format would open on the tabletop near the tool phicon's position and would track with it so the user could move it around to a different position for viewing.

Two Age Phicon's upon the tabletop at the same time would initiate a final puzzle to be solved if desired. The tabletop would change to the same squares as the tabletop's beginning state except that these 4 colored squares would be next to each other in a particular sequence. Associating the cues from before, this would encourage the user to place the 4 colored Age Phicons on the tabletop at the same time in the given order. However, the puzzle would not unlock unless each Phicon's pattern appearing right-side up correctly matched the solution symbol given as a reward for solving each Age's puzzle.

We also as a tangent researched Mayan Calendar systems and learned about their cycles. We think the Haab calendar system might be easiest to adapt to a puzzle one could quickly understand versus the Tzolkin system which may be too different to be grasped in the short time one will have to experience the project.

Apollo's Body

Added the bone structure to the model for Apollo. I will learn how to animate this in Softimage on Wednesday.

Gamepad Controller Research

Looked around for wireless Mac/PC gamepad controllers or ways to modify standard controllers to be wireless. Found a few controllers that looked like that would work form factor-wise ("Gamester USA FPS" for PS2) but would need a wireless adapter. The form is nice since it is designed for First Person Shooter games, having a more pistol-like grip that would make holding the AR Viewer easier and more comfortable for extended periods.

We also researched and found USB adapters and drivers allowing the use of PS2 and XBox on the PC/Mac, including "Game Companion" for Mac OS X.

Assault on the Airwaves

Dropped off our template for the acrylic tabletop.

While researching 1.2GHz video TX/RX products, it dawned on us that we could utilize our AR Viewer's TV receiver to receive UHF video transmissions. This would help us in 3 ways:

1) Less equipment incorporated means less weight added to the AR Viewer rig.
2) No added receiver means less equipment to power and, therefore, longer battery life / wireless AR content viewing.
3) Less time spent hooking up and wiring the video receiver to the AR Viewer rig.

Began researching available UHF video transmitters. They appear to be within our price range ($100-$150) and transmit on UHF Ch. 59 (434MHz).

Hanging the Projector

Tried the PC DV bridge again to no avail. The PC seems to be the right computer for this project save for this problem of getting our analog video signals into DART. We might want to look at some way to do this again and going PC instead of Mac as we are currently leaning.

Bought supplies to hang the projector in the Grad Lab for testing. Had to modify some hardware to do this but eventually got it hung quite soundly. Spent the rest of the evening doing artistic photography utilizing music visualization graphics projected onto a table and ourselves. Cool.

Tabletop and Lots of Stuff

Went to Blick's to get a re-stickable adhesive to enable switching graphics in our phicon's internal support. We then went to East Bay Depot to see if they had anything useful. We found an Apollo 11 video and purchased it for $1. We went to Cobel Glass and got estimates on the size and price of an octagonal glass top. After consideration of this, we are leaning towards the use of acrylic as our tabletop material. It will be 1/2 - 1/3 the weight and about as strong and similar in price without the risk of shattering.

Stephen worked on figuring out how to do video texturing in the Shockwave 3D world. Since DART's video actor is clunky the attempt was made to use Flash video instead and map the SWF to the Shockwave 3D object (or plane). DART could not do this in realtime. Stephen also found some clues to getting the DVCam to work as a bridge on the PC, thereby reviving the possiblity of our previous intentions.

With Andrew Cho's help, Song rigged her horse model to a skeleton. Next week animation will be tackled.

We established where and how we will hang our projector during our test phase in the Grad Lab. Mounting hardware was purchased to place the projector on the ceiling. The projector we got has vertical and horizontal flip modes as well as mounting points allowing us to hang it upside down and correctly orient the image. It also has distortion correction features.

Gavin confirmed that we could use the department's microphone stands at table bases for our mini-exhibits.

Acrylic Tabletop?

Went to TAP Plastics to get estimates on an acrylic tabletop. We will check with a glass place to compare costs-benefits tomorrow.

Continued polygon reduction of 3D models.

Phicon Internals

Worked on creating a template and constructing a cube made of posterboard to serve as a support for graphics placed within our acrylic cube phicons. It will also work to obscure light transmittance through the AR Toolkit marker borders while allowing it to illuminate the custom marker image from behind since the centers of each face are cut out to correlate to the size of the marker borders.

Stephen re-ordered LEDs.

Boning the Models

I brought in and sized the quadraped and biped rigs for my models of the horse and Apollo respectively. I also worked on Apollo's body. On Monday Andrew will help me glue the bones to the horse model.

Polygon Reduction

In order to reduce the load time in DART, I worked on reducing the number of polygons in my model of the horse. It is now a much cleaner more efficient model.

Research, Shockwave 3D, and Nulls

After noticing the troubling artifacts that appeared on my 3D models when I exported them to the shockwave 3D format, and not being able to get rid of them yesterday, I decided to look for some alternatives to the Shockwave 3D exporter that we have been using in Cinema 4D. I came up with a few different altenatives which Stephen and I looked at more closely today. We tried out one called Shapeshifter which is an extra specifically for Director, unfortunately its results were worse. Then Stephen came up with the idea that maybe Director wasn't having enough time to draw our objects and therefore wer were getting these artifacts. To test this theory he gave the objects more time to load, but this didn't solve the problem. Finally, after looking at the odd triangular white (non-textured) artifacts it occured to him that maybe there were some stray points that Shockwave 3D did not know how to handle since it can only deal with triangles. In questioning me about the structure of my file it occured to him that Shockwave 3D doesn't know how to handle nulls, and when I counted how many I had in my file it matched the number of artifacts we were getting. We then adjusted the file so that it did not contain any nulls and rexported. The model no longer has artifactcs. What a fantastic discovery!

Video Bridge and Shockwave 3D on Windows

I discovered that some of the video and Shockwave 3D visibility problems I'd experienced on both the Mac and PC were due to timeline errors in Director on my part. I also discovered that lit or unlit, objects will appear dark in DART unless a light is specified specifically in DART's "3D World" actor. I also discovered that the PC doesn't like playing DART .dir files correctly since the imported Shockwave 3D objects apparently don't load properly into the cast. They have to be reimported on the PC side for them to work. The Mac had no problem running PC DART .dir files with only changes made to driver file paths.

Determined to give the video bridge one more shot on the PC before deciding wholly against using it as our project's other computer, I went into the Lingo script that opened video connections for DART. DART was successfully able to communicate with the camera but decided to try 1440x1080 (a 2x mode) instead of the requested 720x480. Obviously, it was originally looking for 720x540 (1x)--a 4:3 aspect ratio. Since the DART Xtra only receives the width in the cameraInitString, the question is where is the video aspect ratio set at 4:3 (640x480 or in this case 720x540) without options for the DV aspect ration of 3:2 (720x480)? In the Windows drivers? I changed the Lingo script that constructs the string so that this was the string that was passed:

vc: 0 720 29 dvsd dvsd 1

I also originally tried "rgb" type camera parameters but saw in the Message window the camera was offering "dvsd" modes ("digital video streaming device"?). However, no combination seemed to work. DART simply reports "Couldn't find desired mode" then looks for a 2x mode, doesn't find that, and OpenVideo fails. This whole process worked great on the Mac in DART--without having to change the Lingo at all. DART reported a whole different set of camera modes coming from the same DVCam when it was connected, including a 640x480 30fps rgb mode which worked perfectly with the script as written. It was a beautiful de-interlaced image and highly stable from an AR tracking perspective. As I think about this now, this seems to indicate a difference in the way Mac and PC drivers report the camera's available modes.

We wrote to the DART peeps for some answers to our questions. We are definitely leaning more toward using another Mac now. Gavin said that a 2GHz iMac is available. We'll have to see whether this will work for us.

Also, our LCD TV screen arrived today. We checked it out and it looks brilliant and works great. I think it is perfect for what we will be doing with it.

In addition, my LED order for the internal illumination of our cube phicons was cancelled because they couldn't justify restocking their supply with the number I was ordering. They emailed back and said I'd need to order 20 or more before restocking.

DART, Lights, Models

Today I tested the model of my Chariot with lights and without lights in DART's 3D world. There was no difference between the lit and unlit model. So that question is answered. More questions have now arisen though. The Chariot model is very simple and yet we are seeing strange artifacts (white triangles) on the Shockwave 3D object in DART's 3D world. We aren't sure what is causing this so I spent the better part of the afternoon altering the model in various ways (i.e. removing textures, triangulating every polygon, etc. ) and reexporting it to see if I could get rid of the artifacts. Nothing worked. Perhaps we need a better exporter than the one in Cinema 4D.

Destroying So That We May Create

I began taking apart our poker table. I stripped the felt and lining from it then removed the top, made of fiberboard, so that it may be used as a template for either a glass or plexiglass top.

Video Bridge for Windows

Installed two options for creating a video bridge for cameras on our PC Platform. Neither worked to our liking. The Director Xtra we're using would connect to the DVcam directly on the PC but would only take a 640x480 video stream instead of the 720x480 coming out of the camera. The installed software "translation" drivers would either degrade the signal too much (to 320x240) or crop the image to such as well as retaining interlacing and other non-desirable effects. This caused us to bring up the idea of dropping the PC in lieu of another Mac as our second computer.

I also noted discovered some issues with video and Shockwave 3D objects not working on the PC and Mac as they had before.

Website Connection Works

Gavin fixed the problem we had connecting to our webspace yesterday. I took the opportunity to fix the blog link and update our email addresses.

No Refunds

We went to return the camera we purchased at Mike Quinn's. They don't give refunds. So we're stuck with a camera that we could have gotten for half price and that we may forego in lieu of a camera with a slightly narrower FOV.

Ordered The Window

After doing much examination on all the requirements we have for our "Hubble Window"(i.e. size, weight, price, etc.), we decided to puchase an unbranded 8" car monitor that we found on ebay.

Video Bridge

Met with Gavin at 4pm to fix our computer and check out some equipment. We got a video camera for our DV Bridge idea and an LCD projector. Gavin helped us troubleshoot our cross-platform communications issues with Director and gave us some ideas to look at. Song discussed items with Scott while I wrapped up things with Gavin.

We set up our DV bridge which worked flawlessly on the Mac. As expected we experienced trouble on the PC due to its internal handling of video. Later that evening, I worked some on our cube Phicons, specifically the internal LED suspension array. I also did some research and found a few options for bridging DV from WDM to VFW format in Windows internal drivers.

Clint also stopped by later after his class and helped us with figuring out how we will be suspending our LCD projector.

Project Platforms

Researched other cameras that might work given our requirements. Since it's likely none of these will be directly interfaceable into either the Mac or the PC and will be sent via wireless RF, we researched analog capture options (since DART might not be able to recognize the video signal). Many factors were analyzed (resolution, size, power draw, platform compatibility, and price). The researched led us to the conclusion that what we are planning to do is relatively cutting edge in that there were few options available to us and even fewer within our budget. An alternative plan was devised to use the Grad Department's digital video cameras simply as DV bridges. By using their analog inputs, they could in essence function as analog to digital converters and capture devices for our analog cameras. We also found a good webcam within budget that seems to meet our requirements that would preclude this with the PC at least.

Installed Director on our machines as well as DART. Got some rudimentary results with having the two machines communicate but overall had problems. Discovered some networking settings that might have needed to be changed and ended up messing up the computer to where we couldn't log into it anymore. Wrote Gavin asking for help on this and other issues such as our inability to FTP our project website for the last few weeks and whether he thinks our DV bridge idea will work.

More Shopping

We picked up our poker table at Urban Ore. Then visited Mike Quinn's. No LEDs we could use but they did have a black and white camera which fit our needs. We bought it. Later Song did some research online and found it for half the price elsewhere. We're going to return it. We also contacted Gavin regarding getting LCD projectors for our project. We determined that the projector would most likely need a 44-53 degree FOV to project onto the table from the ceiling with a probable focal length of around 4-5 feet. Also found out our software order came in so we made plans to install and set it up the next day.