Skip to content
Review Open access

Technical Evaluations in the Field of AI Music Composition

Nov 2026 · International Journal of Advanced Research in Science, Communication and Technology · pp. 179 · 0 citations · 3 references

TL;DR

A practical, multidimensional framework for evaluating AI music composition systems, which proposes a staged evaluation protocol combining listening tests, task-based assessment, technical checks, and documentation review.

Abstract

AI-based music composition has progressed from symbolic melody generation to systems that produce audio from natural-language prompts, melodic references, or combinations of both. However, the perceived usefulness of a generated track cannot be captured by audio realism alone. A system may sound convincing yet ignore the requested mood, lose musical structure over time, offer little editing control, or provide insufficient information about the provenance of its training data. This paper develops a practical, multidimensional framework for evaluating AI music composition systems. The framework organizes evaluation into five dimensions: audio and musical quality, prompt and structure alignment, controllability and editability, workflow usefulness, and responsible-use transparency. It proposes a staged evaluation protocol combining listening tests, task-based assessment, technical checks, and documentation review. The framework is a conceptual synthesis informed by published work on text-to-music generation, controllable music models, synthetic-music detection, and the supplied systematic review of AI music composition. No new experimental dataset or performance results are claimed. The paper's contribution is an evaluation blueprint that researchers, developers, educators, and music practitioners can adapt before conducting empirical validation

Read PDF

Similar papers

#artificial intelligence Review Oct 2026

A Tutorial on Generative Models for Music Audio

Music audio generation spans instrumental clips, full songs with vocals, edits of existing recordings, and systems that play alongside a musician in real time. This monograph explains the methods behind these systems for readers familiar with neural-network training who are new to music generation. It introduces neural...

M. Modrzejewski, Joanna Kurkowska, Dorien Herremans · 0 citations
Open access Oct 2026

AI-Based Voice to Music Generation System

The rapid advancement of Artificial Intelligence (AI), Machine Learning (ML), Natural Language Processing (NLP), and generative models has created new opportunities for automating creative processes, particularly in music composition and audio production. Traditional music creation often requires musical knowledge, ins...

Abdul Razzaq, Arun Kumar Kh, Bhaskar V et al. · 0 citations
Conference Open access Oct 2026

How Accurately and Creatively Can Generative AI Produce Music Across Genres? A Mixed-Methods Evaluation of ACE-Step 1.5

Generative artificial intelligence music tools such as Suno, Udio, and ACE-Step can now create full songs with vocals, instrumentation, and production quality convincing enough to sound like human-made music. However, musical "quality" is not a single measurable property, and treating "AI music" as one category by itse...

Flora Wu · 0 citations

YuE2: Unifying Symbolic and Audio Music Generation at Frontier Quality

Symbolic models make melody, harmony, rhythm, and form explicit but typically stop before a finished recording; audio models produce complete songs while leaving composition implicit. We introduce YuE2, which unifies symbolic and audio music generation at frontier quality through symbolic planning. A single AR-NAR Mixt...

Rui-Bin Yuan, Jia-Hao Pan, Jun-Yan Jiang et al. · 1 citation
Nov 2026

A Composer’s Workflow for Deriving a Music Score from Nonmusical Audio

“Mixed music” has long served as a means for investigating the evolving relationship between electronic sound and acoustic compositional practice. While a number of composers have explored the extraction of notatable materials from recorded nonmusical audio, such approaches remain relatively underdeveloped and are of...

Yin-Tao Wang · 0 citations
Preprint Oct 2026

SMEBENCH: Benchmarking Neural Music Editing with Synthetized MIDI Edits

Neural music editing requires a system to perform a requested change while preserving the musical content that should remain unchanged. Existing evaluations often lack a paired target edit, making these requirements difficult to measure separately. We introduce SMEBench, a large-scale benchmark of controlled music edit...

Matthieu Cervera, F. Paissan, Micael Antunes et al. · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.