Dr. Hamel and I met up for Week 2 and discovered the [suckah] object was not working properly just on my computer. It seems that others are having issues with [suckah] as well (if operating on Windows 8 or 10).
So, we worked through an entirely new approach to track color within 2 hours. Here we go...
Back in November last year, Laptop Orchestra was tasked with a colour-tracking piece in groups. If you're interested in seeing our performances, you can check them out at: https://ubcsubclass201819.wordpress.com/2018/11/11/colour-tracking-performances/
and my group MALeR's performance:
Below is the template patch Dr. Bob sent out for us to work with. Let's break it down...
Color Recognition
[jit.grab @colormode] grabs our color data in argb
[jit.matrix 4 char 320 240 @name sourceVideoMatrix] creates a 320 x 240 dimension matrix named sourceVideo Matrix that stores characters (data value type)
this all goes into our [jit.pwindow] that displays the Jitter data and images. It's also where we select a colour to track.
the information from that mouse click (including rgb values) and the tolerance are taken into a subpatch [makeMinandMax] to create a pair of min and max colours for colour tracking
Some new [objects] and terms I played with and learned today:
[jit.grab] - digitizes video into a Jitter matrix (the @colormode attribute outputs the matrix's color encoding in argb!)
[jit.matrix] - a matrix for data storage and retrieval and other operations! This could be useful for the #future for manipulating the tracked colors.
Colour Tracking
the subpatch [setParameters] takes in and unpacks the min and max colors to their separate r, g, and b components.
These are taken by [jit.pix] which determines if every pixel is within the colour range of our selected colour's min and max values. Only pixels that satisfy this are let through.
Will incorporate this in the #future but the [jit.fastblur] blurs or sharpens our tracking colour's image.
And then using the jit.findbounds we used a previous post, we can find the location of (the x and y values of a rectangle containing) our tracked pixels.
Some new [objects] and terms I played with and learned today:
[jit.pix] - determines what to do with each Jitter pixel
[jit.fastblur] - blurs or sharpens our image based on an optimized algorithm in Jitter
So there we have it. No [suckah]. Colour == tracked. True.
Now to track multiple colours...
Comments