Pitch Spelling Jazz Lead Sheets, Solo Transcriptions, Classical Piano and Monophonic Scores

2026-06-19 · Source: cs.CL updates on arXiv.org · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics · Depth: Expert, quick

Summary

An algorithm for pitch spelling and key estimation is presented, designed to convert MIDI-like input (semitones, bar boundaries) into appropriate note names, a global Key Signature, and local scales for each bar. This process involves two optimization stages: a "modal" stage proposes a probable scale per bar using a shortest-path search to minimize accidentals, followed by a "tonal" stage that estimates the overall Key Signature and note names for optimal musical notation. Evaluations were conducted across diverse datasets, including jazz lead sheets from the Real Book, jazz solo/bass line transcriptions, traditional tunes, and classical piano/monophonic scores. Originally developed for music transcription, particularly for building digital jazz solo collections for analysis, teaching, and cultural preservation, the method also defines new distances between jazz scales, relevant for musicological studies.

Key takeaway

For music transcribers and developers of music notation software, this algorithm offers a robust approach to generating accurate and musically intuitive scores. By employing a two-stage optimization process, it effectively handles diverse musical styles from jazz to classical. You should consider integrating this method to enhance the quality of digital music collections, particularly for complex jazz improvisations, and explore its defined jazz scale distances for advanced musicological analysis.

Key insights

A two-stage optimization algorithm jointly estimates pitch spelling and key signature from MIDI-like input for musical notation.

Principles

• Minimize accidentals for better musical notation.
• Jointly evaluate related musical information elements.
• Separate modal and tonal optimization stages.

Method

The "modal" stage proposes a probable scale per bar via shortest-path search, minimizing accidentals. The "tonal" stage then estimates the global Key Signature and note names for optimal notation across the entire piece.

In practice

• Automate music transcription from audio recordings.
• Build digital collections of jazz solos.
• Support music analysis and teaching tools.

Topics

• Pitch Spelling
• Key Estimation
• Music Transcription
• Jazz Lead Sheets
• Classical Scores
• Optimization Algorithms
• Musicology

Best for: AI Scientist, Research Scientist, Software Engineer

Related on AIssential

• Pitch Spelling Jazz Lead Sheets, Solo Transcriptions, Classical Piano and Monophonic Scores — The algorithm optimizes pitch spelling and key estimation using a two-stage modal and tonal process for diverse musical scores.
• Tonal parsimony in chord-sequence analysis: combining modulation cost and tonal vocabulary — Tonal parsimony optimizes chord sequence analysis by minimizing modulations and distinct tonal centers.
• Beyond Artifacts: Towards Generalizable Synthetic Song Detection via Music-Intrinsic Features — Sofia improves synthetic song detection by modeling music-intrinsic features with an adaptive Mixture-of-Experts framework.
• Hybrid Diffusion Transformer for Instruction-Guided Audio Editing via Rectified Flow — The Hybrid Diffusion Transformer uses a two-stage coarse-to-fine strategy for efficient, accurate instruction-guided audio editing.
• Voxtral transcribes at the speed of sound — Mistral's Voxtral Transcribe 2 offers both open-weights and API-based speech-to-text models with real-time and diarization features.
• PASQA: Pitch-Accent-Focused Speech Quality Assessment Model Trained on Synthetic Speech with Accent Errors — PASQA accurately assesses pitch-accent correctness in synthetic speech by leveraging a specialized dataset and architectural enhancements.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by cs.CL updates on arXiv.org.