mychang@andrew.cmu.edu – 18-090, Fall 2019

mychang — Final Project

mychang@andrew.cmu.edu — Mon, 09 Dec 2019 06:36:10 +0000

I created my final project using the Leap Motion Controller and machine learning. The sensor will take in the data from my hand movements, read them as different gestures and convert them into digital numbers that the patcher can understand. Each different gesture will output into a different piano chord. The ending position of the hand will also rotate sections of the cubes on the screen.

I downloaded an external helper patch for the leap motion controller from Masayuki Akamatsu.

Link to my google drive folder.

mychang — project 1 — talk to me!

mychang@andrew.cmu.edu — Wed, 06 Nov 2019 07:09:49 +0000

My project displays a face and when you talk to it, its eyes, ears, and nose will move in response to your voice. After contemplating for a second, its mouth will move (accomplished by delaying the signal). This face is also rather shy and feels uncomfortable in environments with a large amount of noise, and the size of its blush depends on the amount of high frequency sound present (accomplished by Fast Fourier Transform).

There are a few sources from which I took materials from:

https://www.youtube.com/watch?v=PuHvNxyOULQ I took some code from the project that is being demonstrated in this video, which is published by Sem Shimla.
https://www.turbosquid.com/FullPreview/Index.cfm/ID/614232 The ear model is taken from a website called TurboSquid. The model is published by a user under the name turbosetro.
I also took some snippets of code from the patchers that Jesse made in class for fft.

Update: link to Google Drive

mychang — assignment 4: duck overflow

mychang@andrew.cmu.edu — Wed, 16 Oct 2019 05:56:06 +0000

This max patch deforms a duck model when a loud sound is inputed into the microphone. I started with the duck fft patch that we made in class, and added modifications to the position, scale, and color fields of the duck model.

mychang — assignment 3: Michelle is sitting in a room

mychang@andrew.cmu.edu — Wed, 02 Oct 2019 04:54:24 +0000

I used Lucier’s “I am sitting in a room” audio because everyone is familiar with this sound track already.

The first IR response was recorded in the A level of Baker hall. The sound of the balloon pop was pretty reverberated, creating a strong echo in Lucier’s voice.

The second IR response was recorded by me tapping in my refrigerator. It has a wiggly sound to it, instead of an echo like the first IR response.

For the third IR response, I recorded myself saying “hi!” in order to see if putting a recording of my voice as the impulse response can transform Lucier’s voice into mine. The result was actually somewhat close to my speaking voice.

The fourth IR response is recorded when I bite into my breakfast toast. Although my room does not have much echo when I recorded other impulse responses, because the crunch sound is composed of a repetition of similar sounds, when I convolve with this impulse response, Lucier sounded as if he was in a room with a lot of reverberations.

mychang — Assignment 2: d/dt

mychang@andrew.cmu.edu — Wed, 18 Sep 2019 01:14:24 +0000

This patch makes use of two patches we have seen in class. One is the patch where we combined two video signals by overlaying one on top of another; the other one is the patch where a delayed video played next to the real video. By adding a RGB analyzer, I made a patch that has a time shifted frame overlayed on top of the real-time video, where the particular time frame was chosen from the past with the RGB values of the current frame. By moving around and messing with my desk light, the RGB values changes and random frames flicker on the screen. If I try to stay still, one frame will be chosen and shows as a stilled image.

In the video, I first stayed still in a few positions, producing the same average RGB values over a period of time and hence one frame was held still each time. Then I started moving around a bit more and putting my hand on and off the webcam, producing a continuous flashing of frames from the past. I call this assignment d/dt because a video is only produced when there are changes in the RGB values; otherwise, it is more like a plain image.

mychang — Assignment 1: I am me 1+12 times

mychang@andrew.cmu.edu — Wed, 04 Sep 2019 04:26:46 +0000

This project explores the idea of seeing how my identity can be manipulated and interpreted differently by people I know; going further, how small deviations can resonate into the crowd and become significant, though they were originally negligible.

I started by drawing a simple self portrait, sent the picture to one of my friends, and asked her to draw in response. After receiving her drawing, I sent that picture to another friend of mine and asked for another drawing. To make this process more controllable, I gave them the following instructions:

Tell me your interpretation of the drawing that I sent you.
Make another drawing of this interpretation; your drawing does not have to look similar to the one that I sent you.
Finish the drawing in one stroke.

I repeated this process 12 times, and my picture was already fairly distorted by then. I included their interpretations as well as their drawings accordingly in the video. A notable point is that many of them chose to draw in similar fashions, which may reflect our tendency to follow the precedents.