Home Creators Posts Import Register Favorites Logout

Content

This update includes a 1-minute-20-second short story with a very simple plot (two versions - Chinese and English dialogue), plus 33 AI videos totaling over 30 minutes. Because this little story is also my first attempt, I’ll share a watermarked, lower-resolution version directly on Twitter and see what more people think. I’ll also pick a few clips from the other 33 videos to share first.

这次的更新包含一个1分20秒的剧情十分简单的小故事(两个版本,中文和英文语音对白),以及33段AI视频,合计超过30分钟。这个小故事因为也是我的第一次尝试,我会直接在推特分享带水印的低清晰度版本,想看看更多人的反馈,也会从其他33段视频中挑选一些先分享一下。

The links can still be obtained from the reply to the first comment under this post.

As I’ve gotten more familiar with the video models, production takes a bit less effort, so I can use the models to try out more ideas. I’ve also realized that with AI video, I don’t have to only send out the carefully picked “good” results. A lot of the test outputs — even the failures — can look pretty interesting. So I’ll share as many of these videos as I can, though I still filtered out most of what I generated.

As for this little story, dialogue is still a very weak part right now — the voices aren’t consistent, and the tone doesn’t sound natural enough. So I don’t plan to spend much time polishing that for now. I’ll mainly try different shots and effects, and wait for the models to improve further.

As I said before, comics are still my main focus, but I remain very interested in generating AI video. Right now there are two main directions for the AI videos: one is bringing the comics to life, and the other is these realistic-style everyday shorts about skinsuits.

You can think of all these videos as the daily skin-wearing play of a little couple who love skinsuits.

Feel free to leave feedback on the videos, and also tell me effects you’d like to see — I’ll judge whether they’re doable.

因为对视频模型逐渐熟悉,所以制作所需要耗费的精力也少了一些,可以把一些想法利用模型进行尝试,而且我意识到对于AI视频,其实我不必完全精挑细选出优秀的结果来发给大家,测试中产生的许多结果,甚至是失败的结果有的看起来也会很有趣,所以我会尽可能多的给大家分享这些视频,虽然其实还是筛选掉了大部分我生成的结果。

至于这个小故事,对话依然是目前很薄弱的部分,音色不统一,语气也不够自然,所以目前我也不打算花太多时间去打磨,而是主要尝试一些不同的镜头和效果,等待模型进一步进化。

正如我之前所说,我的重心还是会在漫画上,但是我还是对生成AI视频有很强烈的兴趣,目前AI视频就是两个主要路线,一个是让漫画动起来,另一个就是这种真实系的皮物日常短片,你们可以把这些视频都当作一对喜爱皮物的小情侣的穿皮玩耍日常。

欢迎大家对视频提出意见,也欢迎大家把自己想要实现的效果告诉我,我来判断一下能不能实现。