AI-generated media
Images, video,
and music
I use ComfyUI to explore different ways of generating media. These tea-inspired examples sit alongside human ratings of the workflows.

01
Images
4 selected images




Model / workflow02.505.00Rating / 5
01Krea 2 Raw + realism + 2x
4.388 reviews
01Z Image Turbo + 2x
4.388 reviews
03Krea 2 Turbo
4.1435 reviews
04Boogu 0.1 Turbo
3.9729 reviews
05Ideogram 4
3.6825 reviews
02
Video
3 selected loops · 5 seconds



Model / workflow06251250Preference score
01MiniMax H3 T2V
11968 reviews
02Wan 2.2 A14B quality
10918 reviews
03Boogu → MiniMax H3
10548 reviews
04Krea 2 → MiniMax H3
10258 reviews
05LTX 2.3 T2V
10038 reviews
03
Music
AI compositions inspired by tea settings—not recordings of traditional ceremonies
01Sencha roomShakuhachi and koto-inspired music
02Gongfu tableGuqin and xiao-inspired music
03Mint tea hospitalityOud, qanun, and bendir-inspired music
Model / workflow06001200Preference score
01ACE-Step 1.5 XL SFT
11311 review
02ACE-Step 1.5 XL Turbo
8691 review
These are early human ratings, not a definitive ranking. Images are rated out of five; video and music scores come from pairwise comparisons, where higher means preferred more often. Review counts are shown beside each score.
