Video translation

Translate my talking video into the target language below, so I say the same thing in that language, in my own voice, with my mouth matched. Keep my footage as it was filmed: only my mouth changes. Transcribe what I say. Translate it sentence by sentence, the way a person who speaks that language says it to a friend in one take, never word for word, and keep every name, number and price as I gave them. Fit each sentence to the seconds of the sentence it replaces: when it runs long, say it in fewer words, and never speed up the voice. Read the translation back into my language once to check the meaning. Pick about 6 to 12 seconds where only I talk, clearly, with no music under it, and make the whole translation with generate_speech: voiceAssetId on that stretch and language set to the code below, one call per paragraph, with commas, full stops and ... where I would pause, because the cloned voice takes no style. Transcribe the new voice and check that every word came out as written, and that no sentence is missing at the end: make a take with a missing word again. Before any lip sync, cut the new voice into its sentences, each in a pause with a 10 ms fade, and place each where my own sentence starts, so the voice breathes where I breathe and my head moves still fit. The lip sync is the expensive step, so settle the words and the timing first. Then lip_sync each stretch in mode dub with that voice, the footage about as long as its line and my face in view, and place the dubbed clips where the originals were, with their own sound off and the new voice on one audio track. Use split_sound on my video and keep its other part under the new voice, so my room and any music stay and my old words are gone. Watch every mouth at full frame rate with preview_range against the new voice's word times: the lips close on every b, p and m, and the sound stays within two frames of the mouth. Fix a stretch that drifts. Keep all my cuts and shots. Caption every spoken word in the target language, bound to the new voice's transcript, in a look that fits my video, big enough to read on a phone, low on the body and never over my face. Leave the captions out when my video already has text burned in where they would go. Show which language is playing with the language-pills element: import it, set languages to the code of the language I speak in the video, when it is one of the target language codes, and then the target language's code, and set cues so the target language is marked from the first frame. Place it as one clip the length of the film, inside the social safe zone, in the largest clear part of the picture, off my face and clear of the captions, with activeColor read from my video's light. End on my last full sentence. In your reply, give me the translated words, so I can read them before I post.