Assignment 3: Spectacular Spectrograms
Musical Statements
"Comb-tuned (Da ba dee)"
by anonymous
"Rosie"
by "not shark"
My statement is a spectrogram image. It uses an image of Rosemary Walten from The Walten Files (with inverted colors because it looks better on the spectrogram). The code used to generate it is as follows:
Using the default parameters in librosa after mp3-compressing, we get this:
"Tralie"
by "bentdoug"
Using the default parameters in librosa after mp3-compressing, we get this:
"Avicii"
by Mike E
Using the default parameters in librosa after mp3-compressing, we get this:
"Let's Groove"
by Diane
"Alien Suisei" (Art by kento @imarandomartist)
by Dylan
Using the default parameters in librosa after mp3-compressing, we get this:
"Fallen CS372Guru"
by CS372Guru
Using the default parameters in librosa after mp3-compressing, we get this:
"The Very Best"
by Hannah
Using the default parameters in librosa after mp3-compressing, we get this:
Not 4'33"
by Kacey
Using the default parameters in librosa after mp3-compressing, we get this:
"Monster Call"
by PANDA
"Song of The Robot"
by Anonymous
I thought it'd be fun to mess around with DALL-E, the AI art generator, to make a png to make a tune from. I made a prompt along the lines of "Robot performs for alien planet of robots" and ended up with this image, which I converted to grayscale and passed into the im2sound algorithm. I thought the resulting "song" sounded quite alien/robotic. I like the idea that this is considered a masterpiece to alien/robot ears
Unfortunately, mp3 compression cuts off the head, which is in the high frequencies that are lost
Original image here:
"Usher-ified Cough"
by Abby
"Mystery Beeps"
by Anonymous
"Lucky"
by Anonymous
"Vocoder Vocoded"
by Sean Sharo
I found a clipart model of a vocoder online and had the Phase retrieval use that picture to create a noise for it
Using the default parameters in librosa after mp3-compressing, we get this:
"Dog"
by Anonymous
Using the default parameters in librosa after mp3-compressing, we get this: