I worked all day yesterday on the animation patch and I think it’s pretty complete. The patch now has two modes, an active mode where frames are displayed as the coordinates come in, and an idle mode where random animations of Larry scratching his head, rolling his eyes, … are played. The patch switches in and out of modes depending on what coordinates it gets as input. Read the rest of this entry »
Salut!
I corrected some flaws in my patches. There was a mistake in the calculation of the center coordinates (the ones that were not going higher than 20). I’ve tested it… it’s working fine (within a 50cm x 50cm testing area).
The only thing that is not working properly yet is the calculation of the speed. I couldn’t find how to save the position of the previous frame (in order to substract it to the current one). It’s probably stupid easy… I’m just not used to the graphical coding environment thinking. The speed can go from 0 to 233 pixel/something (I’ve haven’t figured out which time measure to take.
I’m posting all my patches as text files… have fun!
I worked on a preliminary version of the animation patch today. It takes a position as input and goes down a binary search tree, trying to match it with one of the target positions. This outputs an index, which is then used to load the correct frame in the animation window. The patch files can be downloaded here. Read the rest of this entry »
I captured the idle animation frames today with the help of Patrick Meimari, my Orange Productions animation buddy. These sequences should play whenever no-one is using the installation, kind of like in video-games when the user doesn’t do anything for a while. We made a total of five sequences:
Each of these last a few seconds, so it should not interfere with the action or the animation. Read the rest of this entry »
I captured the animation frames today. The final matrix consists of 7 rows and 9 columns, with the corners omitted, totaling 59 individual frames.
I decided to go against our original idea of a tree structure starting at the center and instead, to go for a more “free-for-all” structure where any position can be reached from any of its neighbours. I think this will result in more natural character movement. Read the rest of this entry »
Hello from Sao Paulo!
I don’t have much time to explain this but I’ll do what I can! You should both have received my patch that I sent yesterday by email. Sorry, I cannot post it on the blog because I don’t have it with me. So here’s how it works.
The first inlet of 3dtracker is for the top camera, the second for the colour to look for in that input, the third one for the side camera and the fourth for the colour to look for in the side camera. You should be able to figure it out from the comments I wrote, same for the outputs.
The most important thing you should know about the patch in order to use it is the coordinates relative to the cameras and how to place the cameras.
Projection is on plane x,y at z=0.
Side camera is on plane y,z (projection on its left)
Top camera is on plane x,z (projection on its left)
It’s important that the projection, side and top cameras are placed in these directions for the coordinates. The origin of the coordinate system (0,0,0) is at bottom-left of the projection (when facing the projection). It would be easier if I could draw a diagram, but I don’t have much tool to do it. If you don’t get what I’m saying, send me an email to my gmail account, I’ll look at them tomorrow afternoon.
Notes: the speed is implemented but not thoroughly tested. As for the direction of the ball, I thought we could use the speed for that (negative vs. positive speed) but from what I remember it was not working. I’ll fix this when I come in next Monday.
See you!
L’idée c’est que je prendrais un feed de Lucie pour determiner le niveau
d’action du participant, c’est-a-dire, la frequence a laquelle la personne
lance le ballon au mur et la vitesse a laquelle il/elle le lance. Plus que le niveau d’activite est eleve, plus l’intensite de l’atmosphere sonore s’intensifie.
peu d’activite = son calme, pas fort, stable, plus ambiant, relax.
beaucoup d’activite = son chaotic, intense, fort, debile.
il s’agit de savoir si c’est possible de determiner la niveau d’activite avec le data recueuillie par le tracking du ballon…
J’vais essayer de linker des loops pour les 2 exemples…
Here is what I’m thinking for the installation of the cameras.
Cameras installation and placement
Since the angle of view and the height of the grid restrict the trackable space, combining the cameras this way will allow to enlarge the trackable space. It would not be much of a 3D tracking system anymore since we’ll loose an axis at different places but we should still have a good idea of where the ball is. I think that placing the side camera so that its angle of view is parallel to the wall will allow to more easily detect the distance left between the ball and the wall.
Let me know what you think about it.
Here’s the first version of the tracking patch. I mostly tested it with red objects, but it’s working good with most colours. The tracked colour can be chosen by clicking in the first pwindow.
note: I was planning to add background substraction to the patch, but after hours of messing around with matrices, operations, etc. I gave up. Unless it’s a real issue when testing in the lab, there won’t be background substraction in the patch. It’s better that I concentrate on getting the coordinates to output properly.
I just completed the animation tests. I made them pretty quickly, and the characters are not completed yet, so keep an open mind while watching.
I originally liked the intr_ceptor model better (the weird robot thing), but after animating with it, I am realizing that it will be a pain. The tree structure that we have come up with works from the centre of the frame to the extremities, but since intr_ceptor rotates around its centre, we would need a different scheme. And it would be impossible to simulate a catch right in the middle of the frame since the shields cannot reach that position.
And animating Larry was a blast. It went really smoothly and he is much more expressive so I think it will be a better choice. Read the rest of this entry »
