EAS TOM (1990-2016) AI Voice

Urgent-bulletin language is the strongest signal for this adult male English voice, accompanied by old, narration, deep, authoritative, serious, announcer, and radio tags. The preview includes specific places, timing, and hazard claims, but the vague description contributes no vocal specification. The sample therefore supports testing a formal alert register, not assuming broadcast credentials, public-agency affiliation, meteorological expertise, factual accuracy, or authorization to issue warnings. No accent is listed. Written tags also cannot prove depth, authority, intelligibility, or reliability with place names, and the title's date-like wording supplies no safe basis for identity inference.

Use a clearly fictional safety drill that opens with an explicit test notice, includes invented locations, spells out one action, and ends with an all-clear. Compare intelligibility, urgency without panic, pronunciation consistency, and whether the test status remains unmistakable. Revise sentence length and phonetic guidance, then rerun the hardest location to check repeatability. Never evaluate with a live or plausible emergency message in a context where it could be mistaken as real. Reject it if the voice implies official provenance, obscures qualifiers, mishandles instructions, or creates confusion about the drill's fictional nature.

EAS TOM (1990-2016) leads with an older, weathered grain, layered with a narrator discipline that carries long passages, a deep, chest-resonant register, and grounded in an authoritative weight that lands statements as facts.

Creators have run 55.3K generations with EAS TOM (1990-2016), ranking it #102 of 340 voices in the Eleven AI public library.

Formal alert color is plausible, but only clearly fictional drills with precise instructions and no implied authority are appropriate.