@dreamingtulpa If your cognitive control can be knocked out by a Bombardilo krokodilo video, then maybe, damn it, there's something wrong with you, not the AI slop. ๐ค ๐
@ballistic_coda I hope the Chinese Communist Party appreciates this and raises my social credit rating. ๐จ๐ณ ๐ซก Just in case, who knows how things will turn out ๐
@ctair1968 Right now, Black Forest Labs charges 17 cents per second for HD video with audio. Third-party aggregators might be cheaper or pricier, but itโs harder to tell due to their generation credit systems and internal promotions or discounts.
Asked Flux 3 to drop visuals from 15-sec video of a boy singing, keep the audio, and generate a new character with lip sync matching original track.
Maybe my prompt wasn't clear enough, but Flux used original track as inspiration and mostly just captured the vibe of ref sound. ๐ค
How it was done:
> Started with a base image in Midjourney v8.2.
> Then fixed and refined it using: Photoshop, ChatGPT, GPT Image 2, Nano Banana Pro, and Lightroom.
> Enhanced fragments and upscaled with Topaz Labs Wonder 2 + Realism.
@theflapslacks Still, itโs a really cool workflow idea. ๐ค When I was playing with Flux, I needed one phrase to have a very specific tone and tempo. And itโs simply impossible to explain this to AI through text, even one tailored for music and vocals. Just an endless roulette of wasting tokens.
Red has waited 40 years for parole.
Fortunately, Trump is running the parole board. ๐บ๐ธ
This should go well.
๐ฌ The Shawshank Redemption - Animated Parody
๐๏ธ All voices performed by me. No AI voices.
#Parody#ShawshankRedemption#Trump
Shot reframe test with @Hailuo_AI MiniMax H3. Do you have a closeup shot that you wish had some wider coverage? Well, run the clip though Hailuo (and include your original input image[s] and character sheet[s]) and ask it to recreate it as a wide shot and maintain the source audio and performances.
#MiniMaxH3 on #MiniMaxDesign