Synthetic, expressive, multilingual speech for cross-lingual dubbing research. Each example pairs a style-annotated text with generated audio that clones an English - ~1.9M examples in the onespeaker config - 17 languages - ~900 distinct reference speakers Each sample shows the generated audio followed by the English reference voice that conditioned it. All nine subsets below are registered as loadable configurations with a train split. File counts reflect the repository snapshot inspected on 2026-09-26; they are not row counts. Some generation runs were stopped before exhausting their source data, so Step 3 and ambientv4 should not be assumed to cover all rows in earlier versions.…