Reading visual data from webcam to trigger sounds in series, in real time
Hello! I am a student in the early stages of working on a patch for an interactive class showcase.
The idea is as follows: viewers will interact with a crossword puzzle, and as they fill it out, a series of sparse looping samples will begin to trigger. With each, say, 10% or so of the puzzle that is filled, more of the samples begin to play, until the puzzle is filled and they're all playing and looping together.
Can anyone point me to some objects that could help me get started with implementing this idea? One of my professors recommended some cv.jit objects, and we were also looking at jit.3m. Anything helps! This project is due in 3 weeks and I am lowkey nervous lol.
your professors leave you alone with hundreds of max objects
to find out how to do this really complex task for a beginner ?
You probably are not allowed to ask for help here on the forum, or ?
or is this more kind of investigative course,
"let's see how they find out or fail" ?
Tell your professors that without text recognition, one can write any bullshit in the crossword puzzle, and for that reson they should come up with somehing that makes more sense.
------
for the part that makes sense - incresing number of samples being played,
I would recommend to first define max number of samples in use,
then grab your 0 - 100 % as simple int, and decide on rules
- do samples that allready started playing keep looping, or do you start
new set of samples each time when % rises or something in between ?
fixed order of added samples, or random pick ? without repetitions ?
Make your logic decisions first, then look for solutions.
for example random without repetitions will tell you to use urn and not random object to pick sample numbers...
At the end you need to tame loudness of few compared to 50 or more samples running.
Hello! Thanks for your response. This project is for a class but the idea was not assigned to me by my professors, it is my own idea and I know it will be challenging to implement but I think it is very possible. I am very much allowed to ask the forums for help.
Also, text recognition does not matter in this project. I honestly don't care if people get the crossword "right" (it will be a very easy puzzle), it just matters that the patcher responds to them filling it out. Could you elaborate on what parts of the idea don't make sense?
I meant getting puzzle right simply by trying
to detect what ?
if field that should contain a letter is blank ?