spring quarter
5/19/05
TRUE SUCCESS
Last night was not only our usual testing the installation session, but the CSUEB annual student art show opening. We opened our room doors to all & got a good taste of what Just Playing With It is like - vs. User Testing.
Michael & Kristin's 4(5?)yr. old son Oscar fell completely in love with the installation. Didn't want to leave, ran around in front of it, leaping and dancing, swinging his arms around, tried to name all the fish. His joy and involvement was radiant in intensity. His uninhibited play is the payoff for this second year. Luminance plays. We've done it.
5/12/05
POLISHING FISHES
Second version of fish, slightly altered re color and shape from original fish, really make a difference re feeling of a Koi Pond in Fishes scene. Need to increase size variables more as well inside scene.
This is an advantage of Flash again, to be able to swap out images and reuse the same tween cycles. The contrasts in proportions and fin shape between the 2 fish types gives an illusion that they aren't swimming in exactly the same manner.
This particular combo of techniques mimics post-production of mid- 1980s computer animation: the individual poses are animated in cycles of basic movement ( our fish have 3 swim speed cycles ) and then their trajectories & interaction with each other are created with another program. Then, the second program was often another machine altogether. Here it's the canned movement code of Flash + ActionScript, which not only controls trajectories, but interactive responsiveness. Cool.
5/7/05
LAST USER TEST
Good mix of returnees and newbies. Users really had much better time. Space to move around helps their feeling of freedom. Addition of sound creates needed atmosphere. The different visual manifestations of their bodies are intriguing. We need better images for FoggyWindows. Folks go "AAHHHH!" when the fish scene first appears, but the user/activity connection is so vague that users quickly become frustrated. Need to work w/ Scott on what changes can be made. Most obvious: users need to be in control of changing the scenes themselves. A third of them listed Roadway as their favorite -sigh - although sluggish performance was definite. We did run into serious lagtime on scene switching from Fishes back to Floating Eye.
5/04/05
QUICK - VIDEO CONTENT
Now that a water "theme" has emerged re digital content, much to and fro re including more video for FoggyWindows, either stock footage loops bought off the web or shooting it live. Absolutely need them as a loop, which is trickier to pull off. Need such for Fishes scene to work as well, in a composite with some pebbles graphic to give illusion of depth.
4/28/05
FAREWELL ROADWAY
Damndamndamn. The Roadway scene just won’t fit with the other sequences and seems like it’s from some other project – however well graphically it relates to the Eye look.
Although I've spent more time with it than any other scene, I do agree that it feels more like a traditional game interaction than the lively, straightforward activity of other scenes. The difficulties mentioned in my blog entry below re data points spread applies, even though my solution to have each changed object revert automatically to its first version after a few moments does work to cover accidental changing.
But surely wish we had time to pursue the idea I had about creating the nodes' screen manifestation as a MASK WIPE between layers of moving visuals. This could serve as the same storytelling function envisioned for Roadway (light into darkside and back) but in the direct manner that we've created for most of the other scenes.
4/22/05
DATA POINTS SPREAD
We've been creating some scenes - FoggyWindows, Roadway, and Fishes - by first enabling a mouse click-based interaction and then translating that into the data point based installation version. There are dramatically different behavioral characteristics for these scenes that go beyond just the contrast of single to multiple activating points.
We learned a while back that the system could effectively handle only 40 data points well. (We also saw how changing the size and shape of the visible tracking nodes on the screen gave very different impressions to the user as to what was going on).
And we learned early on that there's no way - using our code system - to distribute the nodes evenly around the perimeter of a figure's shadow like this:
Instead, the nodes cluster in various areas of a user's shadow:
Where they cluster is dynamic and depends on several factors - one of them being whether the user is wearing light-colored clothes or not! But the conclusion is that the nodes' location along the user's body cannot be predicted.
For FoggyWindows or the Increasing Blocks Wipe, this doesn't matter. The onscreen alterations brought by movement are visually gross and erratic. No precision necessary. But the scenes which depend on fairly exact proximity of nodes to digital elements, like Roadway and Fishes, are gonna be REALLY tricky to make obviously responsive to the user. Groan.
This could mean a more intensive attention activity for the user: since he/she can see the nodes on screen, he/she can move hands and arms to "push" nodes to activation areas … but don't know how enjoyable that will be AND whether that pushing puts user more into traditional game mode.
4/20/05
¡¡#@#%X!! SCREEN MATERIAL, PART 3
Screen from Rose Brand appeared one day late and made WRONG!! Aaargh! They only grommetted (sp?) the top hem. GGrrrr. Obviously a mistake: they had sewn webbing, which is basis for grommet installation, along sides and then just forgotten to do it. Calls to co. gained almost no satisfaction. They offer a redo (no time for us to do that) or a dinky $30 rebate. Rotten business. Shipping alone cost us $40.
Hope that weight of material will help in tautness or taping sides can make up the slack.
4/18/05
FLASH TRAJECTORY ANOMALIES
The canned Flash movements strike again. On the roadway motion paths, the objects' spacing decreases AUTOMATICALLY as the objects reach the 50% mark on their route! From their original positions:
To this:
AAAAAAHH! Much discussion with Steve McEntee as to a solution. As is too often true, only solution is to insert scaled key frames of each object in relation to others on that side of the road, i.e. on the same motion path. More workwork. Simple tweens will not do. Sigh.
SWAP OUTS FOR INTERACTIVITY & DARK ROADWAY
To include one good thing re Flash: straightforward to substitute the result of 'touching" object: swap symbol for another. This also means creating the dark version of the roadway will go 5X faster. We can use EXACTLY the same layout, motion paths, timing, etc. of Roadway 1 and swap out current symbols for dark side ones. Of course, there'll be numerous adjustments due to varying sizes of new objects, but at this point I'd much rather match objects' proportions rather than start over.
4/15/05
BASIC FLASH ANIMATION, NOT ACTIONSCRIPT
Final art on eye and its default cycling animation work fine, but ActionScripted cycling background swirls I created jam the overall project .fla - too dense for operability for program to handle data points, interactivity and animation.
Therefore, my animated.swfs must be created w/ Flash's "canned" programming. Shit. Much less powerful and not as flexible. So I'll need to redo swirl background animation with more mundane tweening, motion paths, etc. and re-design roadway scene. Fish = no problem, was going to use cycling symbols anyway.
PESKY ROADWAY
Serious concerns for roadway scene: other than speed, any movement changes cannot be global; latency in image delivery; sheer size of resulting file - with now necessary many multiple layers will slow overall speed; questions of image density - richness of roadway scene pivots on lush amount of objects & variety of animated events.
Plus, illusion of walking down the road means a slow pace - but not a stuttering pace. Achieving smooth slowness is not a characteristic of files this large.
4/11/05
¡¡#@#%X!! SCREEN MATERIAL, PART 2
The temporary, too-small screen material I got has revealed the appearance of a dramatic hot spot, clearly visible on user side of the installation. Otis' had very minor version of this.
Continued struggles with wretched screen fabric has taken me through 6 more audio visual companies and ebay (one entire afternoon for just online possibilities - all too expensive) into the world of theatrical supplies! This proved much more to our needs, since these companies sell PARTS and customize rather than just offer whole, pre-constructed items. And they are accustomed to working on a large scale - 8' fabric width is nada to them. Plus, they all exist online, so I can actually take a goddam LOOK at what is available.
Worked out the order w/ very helpful Teresa of Rose Brand. Went w/ 3 grommeted sides and bottom trimmed. Have learned more re scrim construction than ever wanted. Delivery not until next week, damn.
4/07/05
LAYOUT RESTRICTIONS
User testing revealed how low interactive areas must be for users to easily make changes. This brings the horizon down on roadway design. Yet setting up aerial elements would get us into variable movement rates (what's moving on land vs. in the air) really don't register as believable illusions. Very draggy.
Having looped video cloud footage would help, but all good res clips I've found are very $$$. Not possible to create effective amateur loops 'cause of cloud mass dynamics. Last Apple project I directed was a time/money sinkhole due to a pro editor wrestling w/ same.
No real time for only one person to explore the possible graphic looks to convey properly engaging density that is also lean in file size. Really need a Flash graphics expert and all the folks I know don't create visually rich Flash animations - the Flash STYLE is a simple one. More complex looks seem to be achieved by imported video. If I had the time and $, I'd create looping backdrops for this in After Affects.
Even more of a puzzle is how to place changeable objects on the road trajectory in such a way that users just standing/moving left & right won't trigger the objects.
We want them to KNOW that their actions cause things to happen. As we learned at the user testing session, people assume after the initial dots sequence that whatever goes on is their doing. Confusion set in when there was no perceived direct correlation.
4/02/05
PSEUDO 3D, with SORTAPHYSICS
Could happily spend the rest of the quarter playing with the variables of Flash movement programming as shown by Keith Peters at http://www.bit-101.com. His supersimple tutorials have been (finally) a window into the process of programming for animation.
Downside = trial and error basis for finding correct variables, meaning no time to fully explore more sophisticated nuances myself for project. Extensive net search has not found code I could use for either cycling eye background or roadway. But Peters' examples are adequate building blocks, however clunky my resulting code may be.
3/30/05
¡¡#@#%X!! SCREEN MATERIAL
Have spent 5 days wrestling with 12 different possible sources for the damn screen material. Our intended 10' X 8' size, necessity to be seamless + rear projection requirement make this seemingly an impossible item! Plus the discussions of translucency over the phone = age old art-over-the-phone conundrum. Muslin too thin, etc.
I have always been rotten with this sewing stuff and find myself caught in the depths of ignorance. As usual, copping to this w/ sales folk helps, but doesn't get us the damn material.
Ironically, Otis' lending of his screen delayed our pursuing our own (past the initial few $k price tags), so now we are in a super rush.The final screen material could have time consuming adjustment impact on visual content. Phooey.
3/25/05
REAR PROJECTION & GREYSCALE
The switch to rear projection hasn't impacted tonal thresholds re greyscale recognition parameters, but only solution to greyscale restrictions is to fine tune each image extensively. Since the most dramatic images are usually high in contrast, this is a bitch.
Painstakingly altering hue and saturation in "danger" zones of each image as a substitute for dark levels does enable the images to work in the installation. The last minute capability of ActionScript to alter the intensity helps as well as a last level tweak.
Our original idea of shooting a live image of a user's face as a basis for manipulation would probably have been nixed by this constraint, or necessitated a completely controlled environment re ambient light.
3/21/05
METHODOLOGY MUSINGS
In no particular order:
#All team members must truly AGREE on the main point, the premise of the project, no matter how long it takes or delays start of the production.
#Decide on the main arena of interactive experience over some period of exploration, don't decide entirely based on current areas of expertise (unless some looming deadline is a factor). Interest & desire are key motivating factors. Possibilities: portable game, living sculpture, community-building events, immersive environment, online puzzles.
#Once you've started building a project, as a creator or part of a creator team, you must be able to imagine complete ignorance of what you're supposed to do if you're a user. You must clear your mind of knowledge in order to see the project as a new user will. Otherwise, assumptions will poison the interface & activity.
#ASAP, gather all the essential characteristics of the intended interactive experience, but remain open to what the testing users actually reveal about what they're undergoing & understand (or think they understand) about what they're doing.
#Single user or multiple user = deeply profound, but can be unclear until fairly far along in development.
#The creator-controlled >< user-controlled spectrum has a tipping point that varies depending on the creator: if the content is entirely creator-controlled (with only the illusion of user control), there is a greater likelihood of effective drama, story arc and satisfactory feeling of closure. But this is a well-trod avenue. Seeding an fertile environment that is evocative and unpredictable CAN be emotionally engaging - the involvement of Luminance's testers was palpable - but the lack of progression also frustrated them. Trick is to create activity that is markedly rewarding, so there's a feeling of closure.
#PEOPLE NATURALLY CONNECT THE DOTS. Often, even when there aren't really connections. Task is to establish specific enough triggers in imagery/activity that folks fill in the blanks themselves. Creators "merely" need to not care what that filler is IF the user-controlled experience is the aim
3/18/05
PRODUCTION CHARTS
Although a complete pain in the ass to create, especially the original template, my production charts do give us a first pass at the scope and possible content of the project. Served to make the winter quarter thesis committee presentation clearer too. Combined all elements into one doc: storyboards, interactive movements, illusion, underlying programming aims, assets to be created and who's to do what.
Problem = who's gonna update/maintain this sucker? Would have to be me, since we have no production assistant or asst. director.
|
