Lab 9: Advanced Function and Audio/Image Data - Wednesday, 2016-04-13

This lab will combine the advanced functions discussed in Lecture 9 with the audio and image handling from Lecture 10.

Contents

Vocal Reducer

Everyone loves karaoke, but it can be hard to find the latest karaoke version of your favorite song. Fortunately, there is a neat trick to reduce the vocals on most stereo audio files that works well for creating karaoke songs. Let's start by reading in the audioclip.mp4 file.

[y, Fs] = audioread('audioclip.mp4');

Examining Audio Data

First, inspect the audio data before we move on to the audio processing. How many dimensions does the audio data have? What is the sampling frequency? Is this stereo or mono audio data? Lastly, use the dimensions of the audio data and the sampling frequency to determine the time duration of the audio clip.

[samples, channels] = size(y);
fprintf('The audio clip has %d samples and %d channels\n', samples, channels);

if channels == 1
    disp('The audio clip is mono');
else
    disp('The audio clip is stereo');
end

duration = samples/Fs;
fprintf('The audio clip duration is %.2f seconds\n', duration);

The audio clip has 950001 samples and 2 channels
The audio clip is stereo
The audio clip duration is 21.54 seconds

Karaokize

Now we are going to write an anonymous function that takes the audio data and reduces the main vocal track. To do this, your anonymous function should accept the audio data, invert the second channel and then combine the channels to convert from stereo to mono. To invert a channel, you can multiply by -1. To convert from stereo to mono, you can do the following (y(:,1)+y(:,2))/2. Once you have the anonymous function written, use it to reduce the vocals in the audio data from above. Feel free to listen to the processed audio data after class using the sound function.

remvocal = @(x) (x(:,1)+(x(:,2)*-1))/2;
yk = remvocal(y);

Image Manipulation

MATLAB is excellent when it comes to handling images. Let's learn how to do some basic image manipulations using the matrix operations we learned in previous lectures. Begin by reading in the emory.jpg image.

img = imread('emory.jpg');

Examining Image Data

Take a moment to understand how the image data is stored in MATLAB. What are the dimensions of the image? Is it grayscale or color? Use the class function to determine the bit depth.

[h, w, c] = size(img);
fprintf('The image has a height of %d and a width of %d\n', h, w);

if c == 1
    disp('The image is grayscale');
else
    disp('The image is color');
end

fprintf('This bit depth is %s\n', class(img));

The image has a height of 1198 and a width of 1597
The image is color
This bit depth is uint8

Cropping

Write an anonymous function that returns the top-left quarter of an image. In other words, your function should use indexing to return a cropped version of the image. Then, use your function to crop the image.

tlcrop = @(x) x(1:floor(end/2), 1:floor(end/2),:);
cropped = tlcrop(img);

Display

Use one of the display functions to inspect the original image and the cropped image.

figure()
imshow(img);

figure()
image(cropped);

Color Swap

Write one last anonymous function that reverses the colors of the image. There are many ways to reverse the colors, but the primary goal is to swap the position of the red channel and the blue channel. Use your function and one of the display functions to visualize the results.

rcolor = @(x) img(:,:,3:-1:1);
figure()
imshow(rcolor(img));

Published with MATLAB® R2016a