The Making of Yvette Loved a Boy on Meth
My music career!
Yvette and I met each other in December of 2015 in Chicago, IL. Yvette will always be the person who called me in the middle of the night on January 10th telling me David Bowie died.
Over the years we've primarily kept in contact various ways.
As a gift, I decided to tribute my first audio-visual project of this type to her experience of dating a meth head.
The creation process of the audio-visual product is as follows:
Lyric Generation:
ChatGPT generated the lyrics with requirements such as including phrases like "meth head skinny," "chase that meth," and "yee haw."
Text File Creation:
With assistance from ChatGPT, I created individual text files for each word in the lyrics using Python code.
Audio File Generation:
Utilizing Amazon Polly interfaced with Python code, I converted each text file into a wave file, scrambling the voice by mixing genders and shuffling ethnicities to create a disorienting effect.
Audio Stitching:
Using ChatGPT generated Python code, I stitched the wave files together, creating a scrambled voice audio file with a spoken word sound.
Piano Note Creation:
The wave file was processed through a MIDI program filter to create jittering piano notes, adding a crashing, chaotic element to the track.
Audio Combination:
The spoken word audio and the piano track were merged, completing the song.
Visual Creation:
The overall intent was to capture Yvette's personal experience and evoke the feeling of being in a dysfunctional relationship with a meth head, both through audio and visual techniques.
Although Yvette and I were never romantically involved, I consider her a close friend. I hope she enjoyed the effort and creativity it took to create this project, which consumed most of three full days. It's my experimental style, and I hope she appreciates my chaotic touch when experiencing
Why this matters to me
Nothing in this track was performed or sampled. No singer, no pianist, no recorded sound. The only human input was my concept and what I wanted the lyrics to say. Everything after that was mechanical: words became text files, text files became Amazon Polly voices with scrambled genders and accents, the voices were stitched together, turned into MIDI piano notes, and fed through a Winamp visualizer. A robot did the “playing.”
But getting it to work was the hard part. I tried IBM Watson before Polly, hit wall after wall of errors, and spent most of three days debugging scripts until the pieces finally fit. AI wrote a lot of the code, but it had to be steered, corrected, and pushed.
That’s how I see any artisan skill now: a machine can paint, but it can’t decide what’s worth painting or refuse to give up when it doesn’t work. The skill was the idea and the tenacity.
And skill alone amounts to little. There are plenty of skilled artisans, and a robot can match most of them. So why would anyone choose the artisan over the robot? It has to be the person: who they are, what they meant by it, the story and the stubbornness behind it. Without that, there’s no reason to pick the human at all.
If you'd like to hear a few more simple spoken word tracks see my SoundCloud.
