评估 Google 多说话人 TTS：为什么私有评测很重要
Evaluating Google’s multi-speaker TTS: A case study in why private evaluations matterText-to-speech systems were originally built to read text aloud in a single voice. As voice AI expands into audiobooks, game dialogue, and advertising, these systems are being asked to do more: generate conversations between multiple speakers. That takes more than generating two distinct voices and stitching their lines together. The speakers need to sound like they are responding to one another, with natural timing, changes in tone, and smooth handoffs. Each voice must remain distinct and consistent while contributing to a believable conversation.Sharath Rao, Kimberly Lo, Alice Baird / Sep 29, 2026
Hume AI 用自建评分标准独立评测了 5 个 Gemini TTS 模型的多说话人对话能力，Gemini 3.8 Flash 与 3.8 Flash-Lite 综合得分最高，分别为 4.11 和 4.10。情感控制是各模型最强项，达 4.26 至 4.31；说话人区分度最弱，五个模型中四个垫底，同性别组合尤其困难，男男配对平均仅 2.87。
来源：Hume AI Blog · hume.ai