Author Archives: mingyuay@andrew.cmu.edu

Project2-Mingyuan Yu

In the movie Hobbits, there is a scene about the hero face a dragon. In my point of view, the dragon’s voice wasn’t sounds nicely because it did not have much reverb on its voice, which does not make sense because the scene happen in a city in a huge close environment and it should have lots of reverb. Also, I want to add some spatial effect on its voice, so the sound and video can be more interactive. In this project, I extract the voice track from movie Hobbits, split it into two parts – main voice and others. Then I use leap motion controller to control the sound image position for the dragon voice. Finally, I add a reverb to the dragon’s voice.

 

 

 

 

Project1_Mingyuan Yu

You can image this project act as a cell phone noise cancellation. It has two inputs which simulate two microphone inputs in a cell phone, one Mic pick up voice + environmental noise and the other Mic pick up pure environment noise. This app combines both amplitude cancelation as well as phase cancelation. It also provided a gate therefore the pure environment noise signal won’t be heard by the user. The following video demonstrated shows this app can separate noise and wanting signal pretty good in both musical environments, speech environment, street noise environment and pure white noise, as long as some condition fited. However, if those conditions don’t meet (ex: not playing simultaneously, phase no match), the app won’t perform very good noise cancellation.

 

Assignment 4_noiseCanceling_Ming

In this assignment, I create a noise canceling tool. I first import a background noise environment sound track, then import a voice-noise-mixing sound tract. By playing with pfft~ I can use either amplitude or phase attribute to cancel the noise from the voice-noise-mixing sound tract and leave me only the voice. You can hear the difference between the two methods.

In the youtube demo video, I first playing 2 original sound tract (noise background and voice-noise-mixing sound), then using amp noise cancelation, follwing with phase cancelation.

PS: the youtube video shift my sound about 1 second for some reason.

 

Convolve_HW3_Ming

Sound source: SUZANNE VEGA ‘Tom’s Diner’

IR source: First two are ballon IR—— one from Doherty Hall and the other one from Bathroom.

Rest IRs are made by White noise in an anecoiec chamber at Columbia College Chicago, and I use Kemar human head microphone to record them. These sound track are record horizontally, from 0°-360° with 30° interval (0°, 30°, 60°, 90°…etc.) and two different distance (<1 meter and 6 meter away from Kemar).

178ada1aedc3fdca-1

627170107a6265c4-2