# aliutkus/speechmetrics

A wrapper around speech quality metrics MOSNet, BSSEval, STOI, PESQ, SRMR, SISDR

Repository: https://github.com/aliutkus/speechmetrics
Canonical: https://ross.abutalabs.com/products/speechmetrics
Language: Python
License: MIT
License Family: permissive
Last push: 2023-07-05T13:16:29+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 2542, "days_push": 1155, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1051, forks 172 (observed 2026-08-28T04:03:23.236206+00:00)

## What it is
A Python library that wraps several objective speech quality metrics (MOSNet, BSSEval, STOI, PESQ, SRMR, SISDR) behind a unified API. It supports both absolute (no-reference) and relative (reference-based) metrics with optional windowed scoring.

## Use cases
- evaluate quality of speech enhancement output
- compute PESQ and STOI scores for synthesized speech
- estimate MOS of tts audio without human listeners
- compare denoised audio against a clean reference
- score speech separation results with bsseval and sisdr
- batch evaluate speech quality metrics in python

## When to choose
- you need multiple standard speech quality metrics behind one consistent API
- you want both reference-based and no-reference metrics in one tool
- you are benchmarking speech enhancement, separation, or TTS systems

## When to avoid
- you need actively maintained code - some metrics break with numpy >= 1.24
- you need metrics not in the bundled set or non-Python implementations
- you need production-grade Windows/macOS support - installation is smoothest on Ubuntu

## Facets
- artifact type: library
- maturity: maintenance
- function: audio-processing, machine-learning, benchmarking
- domain: speech-processing, machine-learning
- platform: python, cli
- tags: speech-quality, mos, pesq, stoi, srmr, mosnet, sisdr, bsseval, objective-metrics, audio-enhancement-evaluation, audio, linux

## Member repositories
- aliutkus/speechmetrics (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:23.236206+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:59:48.314417+00:00, confidence not recorded.
  - readme: https://github.com/aliutkus/speechmetrics (fetched 2026-08-28T04:03:23.236206+00:00, sha 4d87446249d3)
- Data as of 2026-08-30T08:39:29.467469+00:00.