Computer Vision Patch Post

PP3 Patch

 

Screen Shots included within zip!!!!


Selecting Scenes

https://vimeo.com/141210159

I put together an animatic for my first idea for the scene selection.

The icons are purely arbitrary; we would probably want something more representational of the concept behind the individual scenes.

I am totally flexible on the idea, and it is dependent on being able to get reliable X Y positional data on the performer from the overhead camera…


Class notes on 9/30 regarding PP3

PP3 GOAL: Create an interactive ‘dance’ piece (as a group), each should create a moment of reaction from the ‘dancer’

Group goals/questions:

  • Focus on art!
  • What will we do?
  • How will we do it?
  • What is my job?
  • Establish perimeters and workflow?
  • Understand the big picture?
  • How is the sensing being done?
  • What is the system design?

Resources:

  • 4 projectors
  • 3 projectors on floor
  • 2 HDMI Cameras
  • Top down camera with infra red
  • Kinect
  • Light/sound system
  • Isadora
  • Max/MSP
  • Myo (bracelets)
  • MoLa
  • Various sensors (?)

Group 1: Sarah, Josh, Connor, “Computer Vision Patch”

  • Turn on MoLa, video input, write simple computer vision (CV) patch that gives XY coordinates
  • Program lighting system, provide outputs and know what channels they’re on

Channels (1- 10)

  1. x
  2. y
  3. Velocity
  4. Height
  5. Width

All x2 for the 2 cameras (top down and front), 2nd camera would be same  identifiers but on channels 6-10.


Group 2: John, Lexi, Anna, “Projector System”

  • Listen to channels 1-10 of 2 cameras, take data, create patch and a place for them (6 quadrants were discussed, each in charge of one)

Options:

  • Create scenes and implement triggers
  • Each create our own user actor in the same scene

 

(This is by no means an exhaustive or exclusive list, just my general notes from today!)


Pictures from the Class MoCap Patch from Monday 09/28

 

IMG_5235

On this screen

  • Video In Watcher – Camera input
  • Difference – areas that change (eg. moving around on the screen) are brighter
  • Eyes ++ – finds “blobs” (eg. an area where the actor/user is on the screen)

***

IMG_5220

On this screen

  • Start Live Capture Settings (don’t forget to click the show preview, or you will not see anything)

 

***

IMG_5221

On this screen

  • Eyes vs. Eyes ++
  • no “Blob” detector, but it will give you tracking data (eg. Center, Height, Width, Velocity)

***

 

IMG_5222 IMG_5223

On this screen

  • Eyes vs. Eyes ++
  • no “Blob” detector, but it will give you tracking data (eg. Center, Height, Width, Velocity)

***

 

 

IMG_5224

 

IMG_5225 IMG_5226 IMG_5227 IMG_5228 IMG_5229 IMG_5230 IMG_5231 IMG_5232 IMG_5233 IMG_5234 IMG_5235 Anna PP2


Repetition & Repetition

I made several pressure projects because I realized that in order to learn to build a system independently, I needed to start over multiple times. Each beginning was formed with support. Once I had created an initial course of shapes altered through actors to land in a projected stage, and I could progress through adding shapes and additional actors, I found I could look backwards to my initial connections in order to firmly establish which trigger in/outs were correct. I need to know I could manifest them again. So I started over. Replication was teaching me, but repetition was going to teach me better. And so, I’ve arrived at a rather mediocre little looping film of shapes with timed motion and transparencies as its the fourth miniature project of building similar projections. No doubt this restarting choice, with multiple short films was certainly influenced by a few other it’s-real-life factors: a flu, a power outage during my last project, and a subsequent broken computer in the last hours of plans to submit. (A note of gratitude to the OSU media lab in which I am now writing this post.)

The familiar challenge of building a looping project that sustains attention is not lost on this dance and filmmaker. Repetition tends to breed the encouragement to “look” elsewhere, e.g. that dancer has been doing the same forward and back run for a minute, he’s likely to continue, now I am free to examine other elements of the experience. When there are no other elements, attention may fade–or build with the sensation of subtle change or detail. I found with the initial shapes moving through the space, I engaged with an aesthetic draw; even if it repeats and is “boring,” if it’s “attractive,” it’s engaging. Or is it: if it’s engaging it’s attractive? In this last patch, there are three main shapes. How to make three shapes be just the most fascinating thing; this frankly, is not something I have yet arrived at, but with this project I am beginning to find the possibilities that I can actually do. Right now the bluntness of my abilities is blunting not my imagination but my ability to manifest ideas into an actual projection. (I’ve got some grand ideas for a new piano sonata, but I’ve got to learn to play the piano first, etc., etc.) Discovering the Dots actor and extending its own timing while changing its rotation felt like I was able to bring in an element that changed enough to sustain a slight furthering of a viewer’s attention. I had originally played with the Kaleidoscope Actor (misspelled in Isadora), imagining a wealth of change, but discovered it was not applying as generative change as I had expected. I applied Dots instead, recalling a colleague’s mentioning of their current favored actor for the project. I altered the color in addition to the rotation, played with its timing, and set it going in hopes of having a minute experience of moving shapes to keep you watching.

 

Anna PP2


Pressure Project 2 Documentation

For my pressure project I created 2 scenes which looped back and forth to one another. I wanted to play with as many new types of actors as I could while maintaining something comprehensive and interesting.

Click here to see the video!

Scene 1

PP2 Scene 1

Scene 2

PP2 Scene 1


Isadora Live Camera Settings

Here’s the screenshot of the live camera settings for you all!

ACCAD 5194.01 - Live Cam Input Settings


Computer Vision Example from Class

CV EXAMPLE.izz

Here is the CV example from class today.

This file has been compressed. So make sure to unzip it before trying to open it in Isadora.

 

-Alex

 


Square Dance (literally-ish): PP2

For Pressure Project #2, I created 4 different scene’s in Isadora, using various actors but the most helpful was the Ease in/out 2D to coordinate and consolidate the movements and sizes of the boxes. Enter scene triggers, values and delays were used to time the scenes and various other actors were used (dots, reflection, envelope generator++ etc.).

PP2 sg3 PP2 sg2

It was delightful in its simplicity and choreographed movement and it was automatic through the use of scene triggers to loop and repeat. Complexities came into play by varying the movement of the shapes and sizes after setting a constant with the first frame.

PP2 sg1I *believe* we watched it for about a minute, time lapse achievement: unlocked. Some of the feedback from my classmates included the video being delightful (even looked like a bowtie at time) and that the shapes all had particular pleasing places to go as they overlapped. One suggestion/criticism was to include more boxes, which I agree with if time would’ve permitted. 

PP2 sg4

All in all: https://www.youtube.com/watch?v=Uuzmub5PXNQ

 


Duck… Duck… Duck. Duck…… Duck Duck……. Duck… Goose!!!

Let’s see…

Delightful: Check

Visually Pleasing: Oh yea!

Self generating patch of shapes, lines, and color: Check, check, and check (with a goose for good measure)

 

Level 1: System is fully automatic and only requires being ‘turned on’

Fully automated??? It’s practically self aware!

Level 2: System produces multiple visual ‘looks’ or ‘feels’

You got your duck look; you got your goose look.

Level 3: Any underlying pattern in the systems movement and visual state is complex enough that a human takes more than a few seconds to ‘understand the pattern.’

Everything is randomized, even the starting and stopping of the random sequences.

Bonus Level: System produces unexpected results over time.

No one expects a goose…

Bonus Level: Maintains a watchers attention for more than 20-30 seconds. (Much harder then one might guess. How do we do this in the theatre and dance? [Can your system tell a story?])

A complex visual tale of binary, conflicting pairs. Up and down, or left and right; never together.

 

This is a compressed copy with no audio.

Proj1VideoCompressed

I will elaborate on all this after I write the paper that is due tomorrow morning.