POSTHUMAN SOUNDSCAPES by Victor Nazianzeno – Le Jamtel

My artwork focuses on the relationship between human senses and AI. To start with, I wrote a small program using the Selenium library to automate actions on google chrome. The program generated random coordinates (latitude and longitude), input it to google maps and downloaded the first image it could find. This picture would then be send to the imaginary sounsdcape website so that a sound is generated and downloaded. After that, I would pass the pictures through the DeepDream TensoreFlow generator. For my final artwork, I repeated this process ten times. I then edited a video using the ten pictures and sounds, alongside a music I composed some time ago.

‘Posthuman Soundscapes’ relates human’s and AI’s views of our current world. I found it interesting to create an artwork that would transcend the physical boundaries of our world, and add a touch of creativity with the help of the deepDream generator. My goal was to provide a discussion on the how we could define senses for AI. For humans, senses are a way to understand, see, hear and touch our world. For the AI, it is more like a representation tool. Stylizing a picture and adding a generated descriptive sound would therefore provide a whole new representation of an existing location. Another very interesting aspect of this artwork for me was to play with automation and randomness. The first input of the process were random coordinates, and I think it is a very unique starting point that allowed for a systematic but still creative art process. Using Selenium provided automation for the whole pre-processing part of my artwork creation. As I enjoy programming, it was also a good opportunity to progress in this domain.

Finally, I think that this artwork could be extended to provide a full representation of the earth, as an artificial map. Though google maps is used as a practical tool for itineraries and tourism, adding an artistic layer to it would be quite interesting.

Tools used :

DeepDream Tensor Flow Deep neural network
Imaginary soundscape image-to-sound CNN:
Selenium web brower automation library:
OpenShot Video editor:
Github repository of my program :
The music I composed, “Silence”:

Scroll Up