# H3 ref2v MV "I Know What You Are" — retained record

> Retained reddit record from reddit:StableDiffusion.

Canonical: https://highsignal.app/data/records/1d5bd852059082d7

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:StableDiffusion)  published 2026-10-02 05:50 ingested 2026-10-03 02:38

[original source ↗](https://www.reddit.com/r/StableDiffusion/comments/1wvly4n/h3_ref2v_mv_i_know_what_you_are/)

## retained text



```
Finally finished, this is the full song from my [post](https://www.reddit.com/r/StableDiffusion/comments/1wpmy2n/h3_micro_expressions/) a week ago.

So interestingly enough, I unknowingly had been running the ref2v workflow with the fl2av model and the ref2av turbo 768 Lora this entire time.   After realizing and switching to the ref2av model the quality and consistency actually dropped.  So I ended up making this entire video with the fl2av model instead.

Enjoy!

EDIT: Uploaded to [youtube](https://www.youtube.com/watch?v=CDs82ZWjbCc)
```



document fetched 2026-10-03 00:32 · canonical https://reddit.com/r/StableDiffusion/comments/1wvly4n/h3_ref2v_mv_i_know_what_you_are
