Method
A human-led workflow: every track starts with a specific human idea, often drafted in German. These drafts become English lyrics through iteration, rewriting, and a lot of back-and-forth until the phrasing matches my intent. This page exists for transparency about the creative process.
Workflow (high level)
- Impulse: Start with something I genuinely care about (a character, a theme, a moment, a stance).
- Drafting: I write the rough lyrics and structure, often in German or broken English, defining exactly what needs to be said and the emotional arc.
- English lyrics (iterative): The drafts get reworked into singable English through multiple rounds of rewriting until the phrasing, rhythm, and meaning line up with my original intent.
- Music generation (Iterative): Generate many candidates, then narrow down by vibe, structure, and how well the music supports the story.
- Targeted rework: If a track is strong but one section fails, regenerate only that segment and stitch it back in via the editor.
- Artwork and wallpapers: Derive visuals from the lyrical core so the imagery feels like the song.
- Release: Publish the track and attach wallpaper variants.
Creative Control & Authorship
Since I am not a trained singer and English is not my native language, I use technology to bridge the gap between my internal idea and the audible result.
My Role (The Songwriter)
- Original Concepts & Drafts: Every song starts with my specific ideas, often written down in German or rough English notes.
- Lyrics: I decide what the song says, the structure, the rhyme scheme, and the meter. Every line goes through my judgement before it stays.
- The "Final Cut": I generate, cover, and stitch together segments until the track works as a whole. The decision of which take makes it into the final cut is a purely human artistic choice.
The Generative Tools (The Performer)
The tools act as an extensive instrument.
- Vocal Synthesis: Rendering the lyrics with the timbre and emotion I requested.
- Instrumental Execution: Playing the arrangement based on my stylistic instructions (e.g., "play this like a slow 80s synth ballad").
Why lyrics first
Many people enjoy music without understanding a word. For me, lyrics and story are often the main gateway. So this project focuses on tracks that try to say something, not just fill the background.
Iteration is the real work
- Singability beats readability: A lyric can read great but sing badly. If that happens, I rewrite it for rhythm and phrasing.
- Feedback loop: Music generation often forces a return to the lyrics. The song decides what works.
Prompting without naming artists or brands
The goal is to describe musical characteristics and visual properties, not to name existing artists, bands, or trademarked worlds. This helps avoid direct imitation and keeps the song output anchored in the story. Since transparency is key, you can find the primary prompt used for generation on each song's detail page.
If you want to try a similar workflow, the simplest rule is: start with a human reason, iterate until the expression matches your intent, and never lose sight of what the song is about.