Vietnamese audio to text

Mostly. Our Vietnamese test transcript used ă, â, đ, ê, ô, ơ, ư and the tone marks throughout. Many of its slips were a single tone or vowel mark: “sốc” became “sóc”, “chuỗi” became “chui” and “sưu tầm” became “siêu tầm”.

Each of those is a real Vietnamese word, so a spell-checker won’t catch it. Read the transcript with the audio for anything you plan to quote or publish, paying most attention to rare and technical words.

Because some pieces come back without full stops or commas. In our test, about a quarter of the pieces had no punctuation at all, and a few put a capital letter in the middle of a sentence where the speaker paused, as in “hóa chất Bạn có thể”.

The time at the start of each paragraph makes it quick to replay a passage and fix the breaks. Numbers were easier: counts, dates and measurements came back as digits, as in “tháng 8 năm 2011” and “15m”.

The page treats them all as Vietnamese; it has one setting. Southern speech merges the hỏi and ngã tones, which are written differently, so in a recording from the south check words carrying those two marks. For a strong regional accent, try a short clip first.

They are written the way a Vietnamese ear hears them. In our test “T. rex” became “tí bếp”, the Springboks became “Spring Book”, Cuomo became “Kumho” and Jerry became “Cherry”. Expect to correct every foreign name by hand.

For a recording with whole passages in English, use the main audio to text converter instead: it works out the language of each piece, so English stretches are written in English.

Yes. NoteOwl understands Vietnamese among the languages it transcribes on your Mac. It transcribes Vietnamese recordings and meetings with unlimited transcription, labels each speaker and makes every conversation searchable, with nothing uploaded; its dictation is free.