Day 2. Gave the model a voice and got the first real number

The second day went into the voice. Until then the model was a mute picture, and a mute picture does not tell a story. I compared three lip-sync setups, ran the same text through each and watched one thing: does the mouth hit the sound in close-up and does the face drift.

I picked a setup and assembled the first talking reel, 37 seconds long. In two days it got 60 views. That is small by any measure, but it is several times more than any of my mute reels gets now, and it is the first number I can lean on.

A side conclusion about length. Our project rule says keep reels within 15 seconds, and a talking reel physically does not fit that limit: the hook, the story and the conclusion need time. For now I keep talking reels longer and watch retention rather than the rule.