← Projects

Audio Paranoia

Made by Lucy Tan, Kaalen Kirrene and brandonx

To play on the effect of audio pareidolia and trick the viewer into creating something from nothing

Created: October 20th, 2015

0

Curatorial Statement

We used audio pareidolia with a visual component to trick the listener into thinking our looped audio sounds like different words even though it's the exact same series of sounds each time.

0

Intention

The big idea behind this project was to try to create and use the audio pareidolia effect in a meaningful way. Audio pardeidolia is usually something that just happens in the wild. Ghost hunters will use the effect to their advantage to make it appear as if spirits are talking, but they don't deliberately add it to their videos. We thought that it would be  interesting to try to generate the effect by ourselves and use it in an artistic way. 

Our goal was to create poem that would be "read" using the same audio track for each line. We attempted to utilize audio pareidolia to trick the listener into believing multiple lines were being read to him or her. The idea for this project arose from listening to Rachel Gu's presentation and Kaalen's desire to feature an audio illusion. This project seemed challenging but incredibly rewarding to attempt.

0
​
0

Product

Kaalen handled all the audio. The audio was created by recording himself saying the two phrases and using the leveling tool to distort them. Then he generated a sine wave in audacity and used the pencil tool to match the two tracks onto the sine wave 10 milliseconds at a time. He also added white noise to give it the "ghostly" effect. The images were drawn in Paint Tool SAI by Lucy and imported into Premiere as .PSDs. Brandon used Adobe Premiere to put together Kaalen's audio and Lucy's drawings into a single video by adding transitions and syncing the audio. Lucy changed the video speed, uploaded to YouTube, and embedded a loopable version onto this page.
0

Perceptual Illusion

Pareidolia is the effect that occurs when people perceive a pattern when in reality, there is none. This phenomenon works because a person's brain searches for patterns upon evaluating a stimulus and tries to relate the stimulus to something it already recognizes. An example of visual paredolia is seeing an animal in a cloud. In this case, we tried to leverage the effects of audio pareidolia. Examples of this include a person misinterpreting spoken words or mistaking random sounds as a person speaking.

0

Context

We drew most of our ideas form Rachel's project as it held a great deal of information about the audio illusion we were using. However in terms of the creation of our project it was largely our own unique ideas. Not much has been done in the ways of art in terms of audio pareidolia so most our inspiration came from our own ideas and trial and error. We simply noticed a phenomena and tried to use it to create something new and interesting. The visual techniques came from trying to emulate the feeling of the phrase we chose and the harshness of the audio file. 
0

Process

We brainstormed for a bit, but none of the ideas we came up with would work with the five second constraint. Plus, Kaalen really wanted to do something with audio. We ended up deciding on audio pareidolia. We then had to decide how we were actually going to implement the project. Initially, we wanted to do a full poem. We tossed around some ideas involving homophones and rap music among other things. We soon realized that creating a complex poem that would work for the project was a bit too ambitious, so we scaled it down to something more manageable. The first phrase that we came up with that seemed to work was "can't eat candy." Eventually, Kaalen came up with "Hear it, said illusion. Fear it, head delusion," which we decided to use. These two lines were in line with the idea of confusion produced by our project. When creating the audio track, a problem we had was how human it sounded. In order to remedy this, Kaalen manually created a track that looked very similar to actually readings of the phrase. This lead to a less human sound. When doing the visual component, Brandon imported Lucy's images into Adobe Premiere. He then set keyframes and made transitions, while also syncing the audio such that the whole thing would loop. The audio track plays twice per loop, once per line. There is no difference between the audio played for each line, thus allowing the brain to do all the work.


0
Top two are the audio recording, the bottom one is the generated one
Wave forms.png.thumb
0
Had to be this zoomed in to use the drawing tool.
Close up.png.thumb
0
initial planning. We didn't end up actually animating anything, but the poses remained.
Idea1.thumb
0
Sketches
Idea2.thumb
0
How it turned out. Retained sketches underneath to give it a wilder feel, particularly in the right image. Redid expression on right for the same reason.
Idea4.thumb
0
Putting everything together in Adobe Premiere
2015 10 26 20 19 40 photos.thumb
0

Critique

The inner critique in Kaalen wishes he had a better understanding of how signals become sound. If he had more time he would have recorded each sound present in each of the phrases to get a better understanding of what kind of signal they would generate and then use that information to create an audio file that was more accurately just the sounds present in each word rather than trying to hopefully capture the sounds by looking at each recorded phrase.

Brandon: I think that while our the individual components of the project were beautiful and interesting in themselves, we fell apart in putting it all together. The inherently hard to grasp nature of the project did not help in this.

Lucy: I think the audio portion sounded rather harsh, and there was a lot of potential for the visual part that we didn't go through with due to time constraints. Overall, it didn't turn out the way I think we wanted it to, but the video is still pretty interesting.

0

Reflection

In terms of Audio Kaalen did not realize how tough creating the audio pareidolia effect would be. It easy to make tones but capturing the intricacies of speech is incredibly difficult. Knowing the depth that he would have to go into he wishes that he used logic pro X rather than audacity to try and create the audio track.  The difficulty with this illusion is that not everyone will hear it and so the effect will be lost to some. Unfortunately it is just a byproduct of the illusion out group chose. 

Brandon: Having worked with the audio, visual, video editing, as well as conceptual aspects of the project, I found that it all was much more difficult that I had thought initially. If I were do do anything different, I would never do a project based heavily around audio illusions. While I found our idea itself interesting, it was also very ambitious. We simply did not have enough expertise or knowledge regarding the implementation of such an illusion.

Lucy: I actually attempted a solo mock-up, but gave up on it due to how much time it was taking for just a proof of concept. Voice processing is difficult. To be honest, it would've been a lot easier and more cohesive if I executed this myself, but differences in vision are bound to occur in groups. With more time to work on this project, or with a longer video time I think we could have made something better. 

x

7 comments · Latest by Ricardo Tucker

This was a very interesting project! This is pretty similar to the McGurk effect, where the same sound can be heard differently based on the visuals provided. There's pretty much only this video (https://www.youtube.com/watch?v=G-lN8vWm3m0) on the subject, so I never thought it could work with anything more than just one-syllable sounds! I like how the entire piece came together, using the distortion both to create the effect as well as to set the mood of the piece. Nice job!

Join the conversation

Share this Project

Found In
About

To play on the effect of audio pareidolia and trick the viewer into creating something from nothing