I appreciate the level of detail the author goes to in describing the creation process. Very informative. I did this myself once, but using a patch I made in Ma

I appreciate the level of detail the author goes to in describing the creation process. Very informative. I did this myself once, but using a patch I made in Max/MSP for drawing and listening to line segments on a pitch vs time plane. For this particular use case, I generated the input score with a Python script rather than drawing them manually. I found that detuning the sustained tones made the biggest difference in matching the original sound, which the article author mentions. Here's a video of my patch: https://www.youtube.com/watch?v=xl4C4zsy9LY https://www.youtube.com/watch?v=xl4C4zsy9LY