In recent times, emerging video production AIs are standing out in three main ways: generating videos from text, recreating existing videos with synchronized voiceovers, and animating hyper-realistic digital characters. Which one do you think will become the standard method of the future? When considering this, factors like production speed aren't the only criteria—human/object interaction and realism also matter. Are there differing opinions? If you'd like to elaborate, feel free to simplify your point!
Perspectives on AI Video Generators: Which Approach is Preferred?
👁️ 5 views💬 1 replies❤️ 0 likes
1 Replies
Last month, my friend and I were working on a video trailer for an indie game prototype, and we were so clueless that we were racking our brains just to click on H.264 in After Effects. Anyway, we tried this tool called Elai.io—pasted our script, and in 10 seconds, AI spat out a short that was almost lifelike. When I say speed, I don’t just mean “in that moment,” it actually met the criteria the professors talked about. As for naturalness, we even had AI handle the voiceover, and now you can tweak the intonation. That’s when it hit me: text-to-video, at least on a hobby scale, is cracking open the door to the future. Sure, hyper-realistic characters still feel like a myth, but they require such specific scenarios that they feel like overkill for standard use. Maybe once models like Sora come out, things will balance out.