MiniMax H3 reenacts Chinese classical allusions with faux archival footage
MiniMax Design (H3) posted that ancient Chinese stories and idioms are often hard to explain, but H3 has found "archival footage." The post cited user 0xMomo's tutorial on making a "Sima Guang smashing the vat surveillance recording" with MiniMax H3 (ComfyUI) on RunningHub, which covered generating the video in segments, describing a fixed camera position, and the workflow of creating low-resolution dot-matrix timestamps with VCR OSD Mono and GNU Unifont, then overlaying them with FFmpeg.
View original postI also made a version of the "Sima Guang smashing the vat surveillance recording" Ancient surveillance videos have been really popular lately, very interesting I made it with MiniMax H3 (ComfyUI) on RunningHub Steps: First make reference images of the characters and the courtyard Then generate the falling into the water, thinking of a plan, smashing the vat, and rescuing people in segments The video generation needs to describe a fixed camera position The surveillance time and location text can use VCR OSD Mono + GNU Unifont; this part can be handed directly to AI to draw low-resolution dot-matrix timestamps, slightly blur them, then enlarge them, and overlay them onto the video with FFmpeg to simulate the subtitle texture of old surveillance footage The logic in what I generated still has slight flaws; patiently reroll a few more times and the effect may be a bit better
View original
我也做了一版「司马光砸缸监控录像」 最近古代监控视频非常的火，非常有意思 我是用 RunningHub 上的 MiniMax H3 (ComfyUI) 做的 步骤： 先做人物和庭院参考图 再把落水、想办法、砸缸、救人分段生成 视频生成要描述固定机位 监控时间和地点字样可以用 VCR OSD Mono + GNU Unifont，这部分可以直接交给 AI ，绘制低分辨率点阵时间戳，轻微模糊后放大，再用 FFmpeg 叠到视频上，模拟老监控的字幕质感 我生成的逻辑性还略微有点瑕疵，耐心的多抽几回卡，可能效果更加好一点
Body
Ancient Chinese stories and idioms can be hard to explain. But luckily, H3 found the archival footage. 📹