thispatcher

I have coordinates of an area of a matrix that I want to crop and overlay on top of another matrix. These coordinates will change in real-time.

Let me know if I should clarify -- I expect I'm bungling some jitter and/or video terminology here.

I have coordinates of an area of a matrix that I want to crop and overlay on top of another matrix. These coordinates will change in real-time. 

crop-a-circular-part-of-one-matrix-and-position-it-on-top-of-another

The following might be of some help. Go to 

http://music.arts.uci.edu/dobrian/IAP2006/examples.htm

 and scroll to the examples titled "make a mask image", "mask by multiplication", and "mask by alphablend" near the bottom of the page.

The following might be of some help. Go to http://music.arts.uci.edu/dobrian/IAP2006/examples.htm and scroll to the examples titled "make a mask image", "mask by multiplication", and "mask by alphablend" near the bottom of the page.


just send updated control mesages at framrate to all objects involved.

I realized that what I was looking for was Chromakeying.

I'm poking around with this too. So far, the patch is applying a coloured shape to the area tracked through cv.jit.faces. We're having difficulty converting this to a mask, applying an effect only to the cropped/masked area, then layering it on top of the original, unaltered feed. Any suggestions would be quite helpful. I've looked at discussions about applying a shape over face-tracked areas, but am getting tripped up on the masking.

I'm very new to Max, explanations would be much appreciated!

I'm poking around with this too. So far, the patch is applying a coloured shape to the area tracked through cv.jit.faces. We're having difficulty converting this to a mask, applying an effect only to the cropped/masked area, then layering it on top of the original, unaltered feed. Any suggestions would be quite helpful. I've looked at discussions about applying a shape over face-tracked areas, but am getting tripped up on the masking. 

I'm very new to Max, explanations would be much appreciated!


You might want to look at Jitter Tutorial chapters 10 and 29.

You might want to look at Jitter Tutorial  chapters 10 and 29.


Thanks for the tutorial suggestions. I've gone over both, and am attempting to modify them for a live feed.

Currently, the suckah is recognizing the colour on the modified feed ("Feed A", monitored by a print RGBA message), but when output to the chromakey object through [prepend color] (borrowed from Tutorial 10), there is no output to the p window from the chromakey object, and no output from the alphablend object connected by r alphamask (borrowed from Tutorial 29) that should be combining the unaltered feed ("Feed B") with Feed A. Tutorial 21 wasn't very specific about this kind of situation.

Thanks for the tutorial suggestions. I've gone over both, and am attempting to modify them for a live feed. 

Currently, the suckah is recognizing the colour on the modified feed ("Feed A", monitored by a print RGBA message), but when output to the chromakey object through [prepend color] (borrowed from Tutorial 10), there is no output to the p window from the chromakey object, and no output from the alphablend object connected by r alphamask (borrowed from Tutorial 29) that should be combining the unaltered feed ("Feed B") with Feed A. Tutorial 21 wasn't very specific about this kind of situation.


FOLLOW UP: Ended up approaching this problem through OpenGL, which is a whole new world that is way less processor intensive. No Chromakeying, no suckah. The mask is generated by linking the tracked object to a feed that translates it into a black paintoval on white background in jit.ter, with the effect applied through jit.gl.slab shaders. Still tweaking it (I'm particularly interested in finding more current methods of facial tracking), but will likely post in an OpenGL- or tracking-specific thread.

crop a circular part of one matrix and position it on top of another?