Voice comparison
Same passage, several synthetic voices — the “sample speech outputs using available voice options” the SOW asks for (3.13.2). Voice selection is per production; the engine adapter is swappable.
Voices
Loading…
Passage (optional, ≤ 600 chars)
Indicative cost per 100k-word title (~550k chars)
generative$30 / M chars
long-form$100 / M chars
neural$16 / M chars
kokorono usage fee (Apache-2.0, self-hosted; ~$1/GPU-hour)
piperno usage fee (MIT, self-hosted CPU)
Billed to the Library’s own AWS account; the software carries no subscription (SOW §3.12).
Results
Pick voices and synthesize.