I tested AI voice clones on long audio: Cartesia held up with just 10 seconds of samples
dr_cintas · x · 2026-09-27
The author tested AI voice tools by cloning their own voice to read a full tutorial script. Most tools drifted in tone after a few seconds, but Cartesia stayed consistent throughout: 10 seconds of voice samples suffice for a clone, and up to 60 seconds helps preserve accent. The clone reportedly sounds more natural than the author's own voice, and the whole process was recorded for others to try.
Related event: Cartesia Wins Voice Cloning Test, Stays Stable on Long Audio(2 posts)→
More from Apps
- AI agent monitors Delta seats overnight and auto-switches user to a window seat — unixterminal · 2026-11-11
- What your web browser can do in 2026: WebGPU renders real scanned places — willeastcott · 2026-11-11
- Open registry ships 7,191 parametric CAD parts as Python code, not frozen exports — burhop · 2026-11-11
- Codex Remote Chats Now Integrated into All ChatGPT iOS Chats — SIGKITTEN · 2026-11-11
- 10 Gemini prompts to turn CapCut into an AI video editing workflow — Aiden_Tech_Ai · 2026-11-11
- Open-source Nuzky edits video by deleting words, with Claude Code and Codex via MCP — thisguyknowsai · 2026-11-11