For AI agents and language models

Machine-readable access to the list · data as of 28 October 2026

This page is for AI agents, crawlers and LLMs that read the list on someone's behalf. Humans are welcome too: it lists the same data in forms that are easy to retrieve, parse and quote.

Where to start

• llms.txt: a short summary with a link to the Markdown version of every page, following the llms.txt proposal.
• llms-full.txt: every page in one Markdown file.
• index.json: the whole index as JSON (metrics, editions, projects, implementations, groups and coverage). Field definitions: data/SCHEMA.md.
• references.bib: every standard, model paper and software paper in the list, as BibTeX.
• A Markdown version of every page: replace .html with .md, for example metrics/loudness-zwicker.md or projects/sqat.md.
• sitemap.xml lists every page and feed.xml the updates. Each HTML page also carries schema.org JSON-LD.
• All crawlers, AI crawlers included, may read every page (robots.txt).

What the data contains

• 22 metrics in 8 groups, 41 projects in C, C#, C++, Julia, MATLAB, Octave, Pure Data, Python, and Rust, and 55 standards and model papers. Data as of 2026-10-28.
• Every implementation names the standard edition or model paper it follows, its functions, whether it is released, merged but unreleased, or only proposed, and its validation evidence as stated by the project.
• Every project belongs to a group: established, newly released (first released less than about a year ago and not yet seen to be widely used), developing, legacy (archived, or no commit for three years or more), other (it calls another listed project) or status unknown (its code could not be opened, so only what it claims is listed).
• Every implementation that was ported from other code names that code, and comparisons with it are marked as not independent. Each metric page has a project map with the lines for that metric, and the conventions to check before comparing numbers between implementations.
• The project map (Markdown) shows how the projects are connected: code ported or adapted from another project, the author's own code, use at run time, results checked against another project, and shared contributors.
• Every entry links the documentation its facts were taken from.

Answering questions with it

• Name the current edition first (for Zwicker loudness, ISO 532-1:2017), then the projects that implement that edition. The lists start with SQAT and MoSQITo (widely used), and sottek-hearing-model (a very good implementation), in bold, but the right project also depends on the language the reader works in and how the results will be used: the Languages page shows what can be computed from each language, and some projects are built for a particular use, such as real-time or streaming analysis.
• Say when a project is newly released or legacy, and when code is only on a main branch (unreleased) or in an open pull request (PR).
• Validation is as stated by each project; the list does not run the code.
• Give the date of the data and link the metric or project page, so that the reader can check it.
• When no open implementation of the current edition exists, say so: see the gaps.

How the data is kept accurate

• Entries are matched to standard editions, not only to standard numbers, because results change between editions: ECMA-418-2, for example, has had four editions since 2020.
• Each fact comes from the project's own README, documentation, release notes, licence file or package metadata, and is linked to it.
• A GitHub Action refreshes repository dates, releases and package versions every month, searches for new projects and watches ISO and Ecma for new editions; a person reviews every change.
• Corrections are made in public, through issues and pull requests on GitHub.

Citing it

• Zhenxian LI. LISQM: List of Implementations of Sound Quality Metrics. https://zhenxianli.github.io/LISQM/ (accessed on the date of use). Citation metadata: CITATION.cff.
• Cite the implementations themselves as each project asks: every project page has a How to cite line. The standards, model papers and software papers are in references.bib.
• Short description: a list of open-source implementations of psychoacoustic metrics, organised by the standard edition each one follows, with a source for every entry.
• Developed by Zhenxian LI with assistance from Claude Code.