Skip to content

gkim297/AI-NewJeans

Repository files navigation

AI NewJeans

Based on 100+ textual corpus of kpop albums published since 2017, AI NewJeans' ML engine first generates its new lyric through RNN & LSTM modelling, then creates a corresponding mp3 audio file with the voice cover of NewJeans' vocals through TTS (Text-to-Speech) generation, and Aubio audio processing & alignment tool.

  • Data Source: Utilizing a diverse textual corpus from 100+ kpop albums released since 2017.
  • Lyric Generation: Powered by RNN and LSTM models, the AI generates original kpop-style lyrics.
  • Vocal Synthesis: Employing Text-to-Speech (TTS) technology, the lyrics are converted into NewJeans' synthetic vocals.
  • Voice Authenticity: The TTS-generated vocals are meticulously fine-tuned to capture NewJeans' unique vocal character.
  • Audio Processing: The integration of the Aubio audio processing and alignment tool ensures seamless vocal-track synchronization.
  • Final Output: The result is a high-quality MP3 audio file where NewJeans' voice brings to life the AI-generated lyrics.
  • Creative Fusion: This process showcases an innovative blend of AI and human-like creativity, pushing the boundaries of kpop music production.
  • Usage: The generated audio files can be used for demos, experimentation, or creative inspiration in music production.

To start program type the following command in terminal (make sure node.js & npm environments are pre-installed):

cd path\to\ai-lyrics-generator

Then,

npm start

newjeans_ditto_album_cover__1_by_lealbum_dflo1i4-pre

@image credit to LEAlbum