Replying to @Iancu_ai @higgsfield
A fox cub walking home as a series, we would watch every episode
1
8
Replying to @dipakbarkale @PJaccetturo
Reference stills per face is smart, how many rerolls per shot do you budget
4
Et toi, tu veux vivre ça ?
Generathon 4 · 5 et 6 décembre à Paris
Il reste quelques places, les sélections commencent lundi.
👉 luma.com/d1dx34nn?utm_source…
1
33
Ce juré vient de voir un film fait avec l'IA en 24h.
Sa tête dit tout.
Vidéo complète ce soir, 19h30.
Replying to @nonameoasis
Faces changing between cuts breaks it for us, your palette rule fixed that
4
Replying to @cluelesslili56
Li, 90% usable would save so many rerolls, will you test it on real shots?
4
Replying to @dipakbarkale @PJaccetturo
Agreed, consistency across shots eats the hours, how do you keep faces stable?
1
9
Replying to @kraqvoid
Snowflakes holding their shape is the proof, did you grade it after the upscale?
10
Replying to @rioaigc @higgsfield
Picking such an unusual matchup for a live action duel is a smart hook
45
Replying to @kraqvoid
We like the 8K first workflow, did the upscale change how the motion feels?
1
1
21
Roland | Generathon retweeted
Day2 of exploring Character consistency in Ai film making.
Frame 1 is the character u build. Frame2 if my character profile( style guide). Frame 3 is lit night view of my character.
I will be sharing my learning experience here.
🤖 Made with AI





