MidiCaps - A Large-Scale MIDI Dataset With Text Captions



Jan Melechovsky, Abhinaba Roy, Dorien Herremans

Singapore University of Technology and Design, Singapore

[Paper on ArXiv]    [Code on GitHub]    [MidiCaps dataset]


Abstract

Generative models guided by text prompts are increasingly becoming more popular. However, no text-to-MIDI models currently exist, mostly due to the lack of a captioned MIDI dataset. This work aims to enable research that combines LLMs with symbolic music by presenting the first large-scale MIDI dataset with text captions that is openly available: MidiCaps. MIDI (Musical Instrument Digital Interface) files are a widely used format for encoding musical information. Their structured format captures the nuances of musical composition and has practical applications by music producers, composers, musicologists, as well as performers. Inspired by recent advancements in captioning techniques applied to various domains, we present a large-scale curated dataset of over 168k MIDI files accompanied by textual descriptions. Each MIDI caption succinctly describes the musical content, encompassing tempo, chord progression, time signature, instruments present, genre and mood; thereby facilitating multi-modal exploration and analysis. The dataset contains a mix of various genres, styles, and complexities, offering a rich source for training and evaluating models for tasks such as music information retrieval, music understanding and cross-modal translation. We provide detailed statistics about the dataset and have assessed the quality of the captions in an extensive listening study. We anticipate that this resource will stimulate further research in the intersection of music and natural language processing, fostering advancements in both fields.


MidiCaps examples:

Below are some examples of Captions from MidiCaps along with their synthesized audio from MIDI.
A soundtrack song with a classical influence, featuring a string ensemble, church organ, brass section, drums, and violin. The piece evokes a cinematic and epic atmosphere, suitable for an action film scene, while also having a hint of Christmas spirit. It is in the key of G major, with a 4/4 time signature, and revolves around the chords G and D.
A short pop Christmas song with a touch of instrumental pop, featuring piano, acoustic and fretless bass, drums, and electric guitar. The melody is happy and relaxing, played in the key of F minor with a 4/4 time signature. The chord progression follows Bb, Cm7, Abmaj7, Bb, and Cm7.
A melodic and happy rock and pop song featuring a string ensemble, piano, clean electric guitar, slap bass, and drums. The song is in the key of F major with a 4/4 time signature and a tempo of 120 BPM. The chord progression alternates between Bb and F throughout the song, creating a motivational and energetic corporate vibe.
A very short fragment of electronic classical music with a dark and epic feel, featuring a piano playing a single A, F/A chord sequence in the key of F major and a 4/4 time signature.
A happy and melodic electronic and rock song featuring electric guitar, clavinet, drums, fretless bass, and music box. The song is in the key of C major with a 4/4 time signature and a moderate tempo. The chord progression of C, F, and G is repeated throughout the song, creating a positive and uplifting atmosphere that would be suitable for use in games or corporate settings.