Dub a 12-minute video into 4 languages: STT once, TTS four times

The speech part of dubbing a 12-minute video into four languages comes to about $1.84 on Sume at list: one transcription pass of 12 cents, then four voice jobs of roughly 43 cents each. That assumes each translated script is about 9,000 characters, which is an assumption to replace with your own count. Translation itself is a step you bring, and the video swap is a separate small job.

The chain

Dubbing here is a pipeline of separate calls: pull the audio from the video, transcribe it once, translate the text, synthesise one voice track per language, then replace the audio on the timeline. The transcription is done once no matter how many languages you add, so extra languages only add TTS. Sume STT accepts 10 minutes per job, so a 12-minute video is two jobs.

Check the voice for each language

Pick a voice tagged for each target language. If you request Spanish with an English-tagged voice, Sume returns a 409 language mismatch before it creates a job, which saves a charge but costs a retry. The library tags sixteen languages, so check that your fourth market is on the list before you commit.

What can go wrong

Translations run longer or shorter than the source. A dubbed line that is much longer than the original has to be spoken faster, and Sume's speed setting tops out at 1.5, so rewrite long lines rather than rushing them. Music under the original voice is replaced along with the voice when you swap the audio track, so plan a music bed if the original had one. Finally, confirm you have the right to dub and to use the speaker's likeness before publishing in any market.

• Count the characters of each translated script before you budget.
• Measure the dubbed audio length against the video before you render.
• Keep each language's job result so a retake is a small bill.