July 14, 2004
'homecoming'
an update on my final for perry's class:
here is a third revision of the script, the one that was actually shot.
ive decided to make the center screen always run in order and non-interactive. ive also decided to not record or use the apartment voice in the script. the two other screens will remain in a shuffle until approached.
the reasons for this are several:
1. i dont know how i want to record the third voice (the sex/person/style of the reading)
2. the scipt is really preachy in those parts and dont really add anything to the entire piece.
3. by making the middle screen static, you separate the two viewers by that middle screen, reflecting the issues in the video itself.
ive found the takes i want to use and have digitized it. next up (tomorrow) is hanging sensors with perry.
March 07, 2004
midterm, the final chapter
well, i think i got it. the bugs seem to all be gone out of this verson, though it took me a few hours longer then i wanted it to. least ill get to go home and have dinner, right?
it is looking for a path to load the movie that you wont have (i cant upload our final movies, cause they are in the range of 2-4 gb). so change the ubumenu on the 'iam' window to a folder full of quicktimes and you should be set.
i have noticed a lot of sluggishness on the large movies and the sliders not cooperating fully, but on smaller movies, they seem fine.
if anyone plays with it and finds bugs before tues afternoon, id love to know. but as far as i am concerned, it is done.
March 06, 2004
midterm redux
perhaps my final version of the patch, though i think kurt is going to pick up from here while i film.
ive been considering the name 'iam' for the project. it seems kinda suitable, though it has generally been agreed that im pretty bad with naming things. but if nothing better comes this way....
March 03, 2004
final proposal
this is the original proposal i turned in for peggy's 499 class last semester that i am going to build as my final for 542 this semester.
new revision of the script coming after spring break, but this should help flesh it out.
February 28, 2004
new max patch, tripp's version
this patch contains a subpatch called mouseloc that reads a mouseclick and pulls up the movie in a seperate window showing the same frame as the screen that was clicked (in the main window).
it does not reflect any of the help/edits/changes/suggestions made by kurt, perry or will over the past week. kurt has been working in the main patch and so ive tried to stay out of it by working in subpatches.
heres a link to our whiteboard session, expertly photographed by kurt.
things left to fix:
add controls for the main window.
orient the frame count text in main window.
set default framestate for gl video planes.
get better control over the popup. (prob best done by using a pwindow and adding controls in that window, all as part of the patch. ill prob tackle this next.)
try to come up with a name for the freaking thing. (and then fight with kurt over the name.)
film content.
for the microsoft research:
(build website.
film demo video.
plan for and show how multiple streams and flagged data can be represented.)
February 23, 2004
more midterm madness
will aided yesterday and gave me some good ideas on stuff. nonetheless, max is going to be the death of me. there seems to be no good way to get the current frame of a movie - regardless of whether the movie is playing or not. wills method uses a framedump to get the current frame. but this requires a bang, which then forces a playback of the movie.
what i want is a static number, so when/whatever you are looking at, the frame number is reported. allowing a report on pause and during scrubbing.
there has to be a way to do this, but im going crazy trying to figure it out.
i though that sending the framedump out of the qt.movie, then back into a pack of framedump back into qtmovie would create a nice feedback loop that would report only on the current frame. but it seems that to send the framedump message to the qt.movie, you need a bang. so im right back where i started. arg.
why cant this program be more oo and open?
February 21, 2004
midterm max patch v1
my drafts of the max patches....biggest issue is getting the current frame of a movie spitting out. ideas?
see also kurts posts with his patch and pics. (though theres nothing else there....)
February 03, 2004
midterm progress
progress is behind by a week because of kurt's absence. we hadn't really gotten a chance to talk about the project. but we met last night and began sketching out interfaces for a multi-author video blog system.
some of our scramblings on napkins as we went back and forth.
there are several challenges right now to what we have - we need a prototype and we need to finalize the interface. location has still been a sticky point between us, with kurt seeing it as essential and me seeing as less important.
this problem arises because location depends on at least 2 dimensions, which really prevents you from seeing a summation of video over time and location with regards to others. kurt envisions a map, where i think it could be displayed more relatively, if at all.
we do have a solution, but im not convinced it needs to be that complicated.
it just seems to me that time (and the summation of the video) are more important than seeing exactly where you where when it was shot. perhaps itll become important down the road, but i just think kurt has been thinking about maps too much.
i expect that we will have a final mockup in the next few days and can begin building the system and creating a demo of the content.
final
i'm going to realize my homecoming project developed last semester in peggy's class. right now, i have the entire piece planned out, though the script needs one more draft before casting and shooting.
id like to rewrite this upcoming week, cast before spring break and shoot either over break or immediatly upon returning to school. tracking pieces should be installed, leaving me to edit, program the basic stamp and display through max the projections.
a bit of work, but doable.
January 23, 2004
midterm ideas
this could easily turn into another paper.
i have been thinking for a long time about the idea of context, defining a character (not by telling or even showing, but by having the character be absent. this was a focus of one of my favorite tv shows - 'twin peaks' and something i played around with in a text game for interactive writing last spring.), blogs and self-expression.
we produce too much text as is, what happens when we look at video? when video becomes as popular? (it is going to happen - technology will allow us to store more and more data and prices will continue to fall on video technology). kurt told me to go back and look at my notes from the mark davis lecture. i have, and unlike his thoughts, i find these ideas to be an extenuation.
so whats the project?
i attach a camera to myself and shoot. shoot where i look - this could be random, it could be setup, whatever, but its strictly pov. and when i say pov, i mean hardcore, the focus of the eye stays in the center of the frame, very directly where you are looking. digitize it and bring it into max. set the playback speed, so the user can vary it from real time to something like 32x (perhaps a better way to think of it is # of frames per second) and add a filter that looks for changes in shots. a threshold, so it becomes a degree of adjustment, how much detail and informarion you want. you can watch the hour uncut or you can watch the 4 most varied shots a one frame every 15 seconds.
it becomes a way to define chracter through context. the filmer remains in control, but the user can absorb as much detail (or as little) as they need or want. this becomes very important in time-based media. we create more things then we could ever hope to watch. once every one of us is doing this daily, we have to have a way to boil down that which is created.
i imagine this as a larger system. if everyone used a system like this, we would 1. be back to data glut. so how do you distill video so it is managable, when everyone in the world produces a copy of their life to be viewed by everyone else. 2. the narrative side of things is entirely context and what you film. given that you created enough of these, you would build up a chracter, not by what they said or wrote online, but what you watched them do, what they looked at and what they controlled.
it becomes more interesting if you had more people using it, much like hte patholog. you could build up entire scenes, days, months from different points of view. in a densly populated area, you could have a complete map of a time/space from every pov.
when i was a kid, one of my idea of heaven/reincarnation was sitting around watching other peoples lifes through their eyes. this seems to be a step.
(kurt wants more metadata, though he has yet to explain why - he wants gps and image recognition. i can see value in these, but i dont believe they are necesary at thsi stage. i believe that given enough of these for a single person, you could begin to understand that person simply from seeing where they looked and what they did. the information becomes the context.)
bueller? bueller?
January 17, 2004
project listing
not a list of just projects for 542, these are most of the im projects that are floating around in my head these days.
1. a sort of homecoming. fleshed out last semester, it involves presence detection, video projection and narrative. the proposal and explanation can be found here.
2. wrist annotator. id love to build a small gestural device that sits on your wrist and takes notes for you. thus far, ive managed to narrow down input to (i think) finger tracking using shorthand. i need to have some serious talks with people who know more than i do at this point though, because its starting to seem a little hairy. (th ebest would be to dump the ascii onto a compact flashcard. then you could view it anywhere - i dont plan on this device having a screen, its for notetaking only.)
3. patholog. will gave a really good explanation of where this is going and provided links to where it has come from and what it is. in addition to the items he listed, im interested in the metadata/mining abilities this system could provide you, your trusted sources and the rest of the world. building in the option for realtime updating, flagging, alerts, temporal/location overlap options and privacy for the system are all things i am interested in building on. (these differ, in some cases slightly, in some majorly, from wills explorations of the same platform.)
4. mobile project / collective responsibilty. in collaboration with imsc, we are working on developing a tracking/projection/multi-player game system. i would like to turn this out for my midterm, but im not sure about that as a realistic timeframe. ill know more after we have a mobile meeting this upcoming week.
5. that paper. im not providing a link because i expect to post my new draft this weekend. this is the same paper i have been working on for several months, with more words now. it still isnt a final draft, but its approaching one. of course, this draft has inspiried more ideas and there are two more sections i would like to explore once i wrap this one (and its two sections) up. the ideas that have come out of this one have been extremely heavy on the metadata side of things - both in creation and utilization. as always, the actual process should be kept as simple as possible for the end user (both the browser and the content creator). currently weighing in at about 7500 words, its a bit to bite off. but i dont want to give up on it now.
theres more, of course, but these are the foundations/fleshed out ones. im still very interested in location-based narrative, sending the player around and interacting with others. while i think the zombie game is mostly done, i think that lessons learned can and should be applied to something new, hopefully on a pda.
im not really all that interested in persistant worlds, but mobile entertainment is still way up on my list. games and accessories are things i am really intrigued by - the type of thing you can use briefly and put back away. i find more and more that i have a few spare minutes to play. not 30 minutes to explore, but just a few while im waiting for something or someone. this is the niche i would like to work on filling with currently available technologies.
