Hey there guys!
I have just entered the world of vvvv and it just rocks!
I study audio-visual arts in Greece and i want to develop a project with my fellowstudent.
On our project we will have a black room, a camera in the back of it and a projector.
We want to take the figure of anyone that enters the room and according to the height of that person a musical instrument will be projected on the opposite wall.
I have just done the tutorials and i think is simple to get tha point of how vvvv works, but i think we probably need some help.
So, can anyone tell us any ideas how to develop it in vvvv?
Thanks in advance.
(nikolasgr@hotmail.com)
hello
try thinking from the perspective of the sensor. camera in your case. what does it mean the ‘height of that person’? how would YOU perceive the height of any person entering the room? and by the way… is the height of the person really the trigger that moves your art?
freeframe plugins in vvvv are pretty accurate in means of color and frame to frame changes.
just think in the way what your camera is perceiving and how would you like to convert it into the information you need.
for example, you could sense any object of any color differnet from surroundings using the ColorTracker node (and that could be any object brighter than black) and get the Height output of the tresholded sample. just conect the VideoIn to ColorTracker, turn on ShowROI, play with it and you will see. some math can be applied depending on the position of the camera.
note that multiple persons i.e. multiple bright objects from camera perspective can not be recognised as separate objects with one ColorTracker node.
but, back to what i am saying here, creating art is always about taking some information from around you and interpreting it into the art form. you have to know or feel what you want to do with information your sensors are giving to you. using technical devices would need some kind of project plan.
i am curious to see how your idea is going to develop.
br
thanks a lot for the help.
i have made a project plan, and we know that the project we are going to do is something more like a game, and not an art project.
I am not that good in english so i can not explain, and express everything the way i want to. But i am possitive that with a little help from the forums and with our time spent in vvvv, something will be brought up.
I am now testing my camera with camshifttracker and i was trying to do what you have suggested but i do not understand what ROI is and how will i exactly get the height and the width (Adjust ROI position and size (or the camera position) , so that your object is covered by the ROI.).
thanks in advance
Hi there!
i was trying and at last the camera shift tracking worked and here is one image.
The problem is now that i don’t know how to tell the program when it has color on the video output to give me in another screen a circle for example.
Can anybody help?!
thanks thanks thanks in advance!
shifttracking.JPG (31.6 kB)
hello again!
your english is very good and please do not worry about that. the most of the people here are from parts of the world where english is foreign language, so you can see. :-)
it is great you found your way arround CamShiftTracker plugin (it took some hours in my setups too). but what about ColorTracker? those act quite differently and you can combine their results (trasholded video streams) to focus on exact targets.
if you post your patch here and describe your test setup, we could work it out somehow.
i am not experienced vvvv user, just to mention, but i have several solutions for my ideas. what i usualy do is:
- prepare the background, either black or white, horizontal on the table or floor
- adjust lights to have as uniform lighting as possible
- focus my camera on the background
- do whitebalance to adjust levels
- prepare sheets of background colored paper, large enough so my camera will not see my hands when i move the paper (use gloves, or cover hands if setup needs that)
- put some different colored objects on those paper sheets so to be able to move them without actually touching them
- work with vvvv from here and fine tune the results of tracking
this helps me to create all math and relations from camera perspective.
after that i try to work out the dinamics of real setup. in real setups dinamic range is kinda narrowed because of lot of reasons, but once you set it up, you can use the same math from your simulation.
i know some other guys here have lot more experince. maybe they will join in.
there is also one interesting module you can try to analize: blobtracker
i can’t recall where i took it from.
and… thinking about your project, i would say you will have to use at least two cameras. not only to successfully track your targets, but to add some more interactivity to your project.
how about that?
o_o!
Thanks a lot for your help!
Your advise are useful and we will try to get around with them when we will have the apropriate room ready. For now we are just testing in a room with a lot of objevts around and withoutuniform lighting.
We have tested the camshifttracking and the colortracking nodes and i think we can get quite good results!
I want know to have a renderer in another screen and in addition to the height of the person that enters the room, a specific video with music instruments.
I can not understand much of all these things and thanks a lot for a user that already helped me a lot!My little problem is that i am working on a mac and i do not have all time a pc in my hands, except of not going that well with programming.
Anybody who could help my msn address is the following nikolasgr@hotmail.com.
Thanks in advance!
(blobtracker?!?)
