horror AI音色

The evidence describes an adult male English voice with a US accent designation, tagged English and Public Voice. It lists Speech, Creator, and Voiceover as possible contexts, while the description emphasizes quick speech tests, creator drafts, and production voiceover experiments. The preview's natural and expressive language is generic and offers no genre-specific proof. In particular, nothing establishes fear, suspense, menace, dramatic timing, vocal texture, accent accuracy, sound-design compatibility, or competence in horror material. Genre expectations suggested by the identifier remain hypotheses to test, not attributes to advertise as facts.

Audition with original suspense copy that includes a quiet setup, an ambiguous warning, a sudden factual interruption, and a final unresolved sentence. Compare restrained and heightened direction for intelligibility, silence placement, dynamic control, and whether tension comes from delivery rather than added assumptions. If a passage feels flat, adjust punctuation and performance direction separately, then listen blind to identify which change mattered. Reject if whispered words disappear, intensity becomes unintentionally comic, timing varies too much across takes, or the voice needs unsupported genre claims to justify selection. Keep it only when the actual scene edit confirms the intended effect.

Creators have run 540.0K generations with horror, ranking it #21 of 340 voices in the Eleven AI public library.

Audition against real suspense material; keep only if clarity and tension survive controlled, blind comparisons.