TE科技与人工智能ElevenLabs’ new v4 speech model supports more expression control and 90 languagesElevenLabs v4 can clone voices with a 10 second clipTechCrunch · 2026年9月28日 14:00