PreprintCC BY 4.0Zenodo
Voice cloning sample length: a 30-second default and how to test 60s
Abstract
Default to 30 seconds of clean speech for voice cloning; extend to 60 only when a held-out test shows measurable MOS or speaker similarity gains.
Independent technical note mirroring the canonical version: https://clonemyvoice.io/blog/voice-cloning-sample-length-a-30-second-default-and-how-to-test-60s.php
Related
- Stream Voice Cloning Audio: C# `yield` at 24 kHz, Frame by Frame
- Zero-shot voice cloning reference audio: when to clone, re-record, or skip
- Voice cloning robustness test: Mean Opinion Score (MOS) 4.2 floor — ship or retrain?
- HURDLER: a cloning method to overcome length limits in designed repeat proteins
- Deep-ultraviolet second-harmonic generation in a three-coherence-length SrB ₄ O ₇ microcavity
BibTeX
@misc{cooper2026voice,
title = {{Voice cloning sample length: a 30-second default and how to test 60s}},
author = {Dylan Cooper},
howpublished = {Zenodo},
year = {2026},
month = oct,
doi = {10.5281/zenodo.23431656},
url = {https://zenodo.org/records/23431656},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Voice cloning sample length: a 30-second default and how to test 60s
AU - Dylan Cooper
PY - 2026
DA - 2026/10/16
PB - Zenodo
DO - 10.5281/zenodo.23431656
UR - https://zenodo.org/records/23431656
AB - Default to 30 seconds of clean speech for voice cloning; extend to 60 only when a held-out test shows measurable MOS or speaker similarity gains. Independent technical note mirroring the canonical version: https://clonemyvoice.io/blog/voice-cloning-sample-length-a-30-second-default-and-how-to-test-60s.php
ER -