PreprintCC BY 4.0Zenodo
Voice cloning robustness test: Mean Opinion Score (MOS) 4.2 floor — ship or retrain?
Abstract
Ship a 4.2-MOS voice clone only if worst-case MOS is at least 4.0 and speaker similarity is at least 0.80 across three degraded conditions; otherwise retrain.
Independent technical note mirroring the canonical version: https://clonemyvoice.io/blog/voice-cloning-robustness-test-mean-opinion-score-mos-42-floor-ship-or-retrain.php
Related
- Stream Voice Cloning Audio: C# `yield` at 24 kHz, Frame by Frame
- Voice cloning sample length: a 30-second default and how to test 60s
- Zero-shot voice cloning reference audio: when to clone, re-record, or skip
- Prognostic Significance and Clinical Robustness of the HALP Score in Super-Elderly Patients Undergoing Curative Gastrectomy for Gastric Cancer: A Multicenter Retrospective Cohort Study.
- Churn Model Drift: Population Stability Index (PSI) Over 0.2 Retrain vs Recalibrate 2026
BibTeX
@misc{cooper2026voice,
title = {{Voice cloning robustness test: Mean Opinion Score (MOS) 4.2 floor — ship or retrain?}},
author = {Dylan Cooper},
howpublished = {Zenodo},
year = {2026},
month = oct,
doi = {10.5281/zenodo.23587801},
url = {https://zenodo.org/records/23587801},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Voice cloning robustness test: Mean Opinion Score (MOS) 4.2 floor — ship or retrain?
AU - Dylan Cooper
PY - 2026
DA - 2026/10/22
PB - Zenodo
DO - 10.5281/zenodo.23587801
UR - https://zenodo.org/records/23587801
AB - Ship a 4.2-MOS voice clone only if worst-case MOS is at least 4.0 and speaker similarity is at least 0.80 across three degraded conditions; otherwise retrain. Independent technical note mirroring the canonical version: https://clonemyvoice.io/blog/voice-cloning-robustness-test-mean-opinion-score-mos-42-floor-ship-or-retrain.php
ER -