Skip to main content

Model selection for Transcription

Via our Klangio API for transcription we offer you a range of models to select from, to optimize the transcription process. Here you find further details and guidance on which model to choose for your use-case.

Overview

Model IdentifierDescription
detectUtilizes the instrument detection capabilities of the Klangio AI to select the best suitable model.
universalA general-purpose transcription model that will extract hearable notes from the song. Specialized models yield better results.
pianoOptimized for audio consisting mainly of piano.
guitarOptimized for audio consisting mainly of guitar.
bassOptimized for audio consisting mainly of bass guitar.
vocalOptimized for audio consisting mainly of vocal.
leadWill extract the overall melody and chords played in the audio.
drums - newOptimized for audio consisting mainly of drums
wind - newOptimized for audio consisting mainly of wind instruments like sax, flute, recorder, trumpet, etc.
string - newOptimized for audio consisting mainly of string instruments like violin, viola or cello
piano_arrangement - newCreates a piano arrangement based on the provided audio