SoftwarePlus.ai

VideoStories+ help

Transcripts and caption lines

The transcript is every word the storyteller said, each with the moment it was said. It becomes the captions on the video — and, in step 4, each of its lines becomes a beat of the plan.

Step 3 (Transcript) in VideoStories+ has three parts.

Step 1 — Transcribe the video

  • List the names and places first. Speech recognition spells what it hears, so a name it does not know comes out wrong ("Mayaka" for Myakka). Type them in Names and places in the story, separated by commas, before you transcribe.
  • On this computer (the usual way): Whisper, a speech-recognition program, listens to the recording. It is free and the recording never leaves the computer. The first time, Whisper's model (a large file) is downloaded once; after that it works without the internet and takes about a quarter of the recording's length.
  • In the cloud (in Settings): OpenAI or Mistral AI writes the words, using your own key in AI Manager. Only the sound is sent. It costs about one cent for a two-minute story.
  • Already have captions? If a phone or YouTube Studio made captions for this recording, bring in its .srt file instead.

Step 2 — Break into lines

  • A line is one caption: one or two rows of about 42 letters, on screen for up to about seven seconds. Good lines follow the way the story is told — one thought each, broken where the storyteller pauses or a sentence ends.
  • A line is also a beat. In step 4, each line becomes a beat of the video, so a setup or a punchline on a line of its own can have its own picture, or land on the storyteller's face.
  • Move a break yourself: Split cuts a line where the cursor is; Join ↓ puts it together with the next.
  • Or let the AI suggest the breaks (For step 2: break into lines with the AI, at the bottom of the page). Only the breaks move: your words and your corrections stay exactly as they are, and each line keeps the moments its words were said. You see the new lines before you use them, and Ctrl+Z puts the old ones back.

Step 3 — Review the lines

  • Read every line and correct what was heard wrong: click in the words and type. These words are written onto the video exactly as they are here.
  • ▶ plays a line, so you can hear what was really said.
  • Comes and goes are when each caption comes up and goes, in minutes:seconds. Type a time, or click in one and press the up or down arrow to move it a tenth of a second.
  • Correct a word everywhere fixes a name heard wrong in many lines at once. At the bottom of the page, the AI can find names and places that were heard wrong; you tick the ones to correct.

Every change can be undone with Ctrl+Z. While the plan in step 4 is still the one made from the lines (and you have not changed its beats), it follows the lines as you split, join or re-time them.