Welcome. I'm Jordano one of the Directors at DreamCodeFilms. This community exists because most AI video looks the same, same prompts, same outputs, no point of view. We do the opposite. Here we direct AI, we don't just generate with it. The craft comes first; the tools serve the story. So before anything else — introduce yourself in the comments. Tell us three things: 👋 Who you are (and where you're posting from) 🎬 What you're making right now, the project or style you're into 🎯 The one thing you want to be able to create by the time you leave here Everyone in this room started exactly where you are. Drop your intro below and say hi — and reply to one other person while you're here. That's how this place stays alive. Stuck or just want to talk shop? DM me anytime. I'm here. — Jordano

This is one of the videos I made on TapNow. My content focuses on the concept of film cuts, with more emphasis on dialogue. Some friends have given feedback on the lighting and color grading. What do you think? What could be tweaked to make it better?

TL;DR: Your AI video gens have correct lip-sync but bad voices. ElevenLabs has great voices but no lip-sync. Combine them generate with audio on (keep the timing), cast and re-record in ElevenLabs (keep the voices), overlay (mute the original). That's the whole workflow. --- The problem AI video has gotten cinematic. The visuals are real. The motion is real. Then a character speaks and immediately everything feels like AI again. This isn't a tool failure. Higgsfield and Seedance can generate audio. ElevenLabs makes professional voices. The reason your scene still sounds fake: nobody combines them correctly. The fix is a 3-step pipeline. --- Step 1 — Generate your video with audio ON Most creators turn audio off when they generate, because the draft dialogue sounds bad. That's the mistake. The voices are wrong. The lip-sync is right. The lip-sync is the only thing your gen does that ElevenLabs can't. So: generate with audio on. The output has a draft audio track with timing locked to mouth movements. You're going to throw away the voices. You're keeping the timing. --- Step 2 — Cast and re-record in ElevenLabs Two clicks per character: 1. Choose the character voice. Pick from the ElevenLabs voice library or clone your own. Lock in one voice per character — this is your reusable cast. 2. Have it say the same lines. Take the dialogue from your video gen. Paste each line into ElevenLabs. Generate. You now have clean dialogue tracks for every line, in the voices you actually want, that can be reused across every shot in the same movie. Same Goblin voice in shot 6, shot 14, and shot 27. Same King voice anywhere he speaks. Consistency without re-recording. For the scene I used as the example, the two dialogue lines were: > Goblin: *"Sire... the boy... he's almost at the gates."* > > King: *"Release the dragon."* The ellipses on the goblin's line force the stammer for his panic. The period + short sentence on the king's line delivers weight. Punctuation does the acting ElevenLabs respects it.

TL;DR: Stop sending every shot to Seedance. Storyboard your scene first, tag each shot's complexity, then route accordingly — Low → Kling, High → Seedance. Time in pre-production = credits saved in generation. --- The mistake everyone makes When you start making AI movies, you learn fast that Seedance is the best at complex action. Multi-character combat, fast motion, dragons attacking giants — all Seedance territory. Then most creators do the dumb thing: send EVERY shot to Seedance. Even the establishing shots. Even the character intros. Even the slow atmospheric beats. That's how your credits drain in a week. --- The split Half your scenes don't need Seedance horsepower. Here's the rule: Complex shots → Seedance - Multi-character action - Fast motion / combat - Heavy VFX - Anything where the model has to track multiple interacting elements Simpler shots → Kling - Character introductions / portraits - Establishing shots - Atmospheric beats (slow camera, single subject) - Anything that's mostly composition + slow movement For my Warrior project, the climactic combat went through Seedance. The character sheets, the silhouette opener, the wind-up shots all Kling. Same project. Two tools. Half the cost. --- The unlock: storyboard first Most people skip storyboarding because it feels like a chore. Then they generate, hit a budget wall, and start cutting scenes. Storyboard every shot in your scene *before* you generate anything. Tag each: - Low complexity → Kling - Medium → Kling or Seedance (your call) - High complexity→ Seedance Now your storyboard IS a cost spreadsheet. You can see where the budget is going before you commit a single credit. --- The principle > Time in preproduction = credits saved in generation. 30 minutes of storyboarding and tagging saves you hours of regeneration AND a chunk of your credit budget. The ROI is insane. --- Inside AI Film SchooI we teach the full pipeline — storyboarding, tool routing, prompting, editing — at script-to-movie-in-24-hours pace.

Project, scene, the prompt you're tweaking, the cut you're proud of, the one that's wrestling you to the ground — drop it in the comments. Wins and stuck points both welcome. I'll be replying. Reply to each other too.