In https://github.com/halfuwu/FortnitePorting/pull/42, I included a bit to scale all bone locations according to the scale of the faceAttach bone. This resulted in shape keys closest to a model I had from a very old export which I used as reference.
Unfortunately, my reference model was probably incorrectly imported since, upon closer inspection, the facial poses are not accurate to what I see in-game today.
Summary
In https://github.com/halfuwu/FortnitePorting/pull/42, I included a bit to scale all bone locations according to the scale of the
faceAttach
bone. This resulted in shape keys closest to a model I had from a very old export which I used as reference.Unfortunately, my reference model was probably incorrectly imported since, upon closer inspection, the facial poses are not accurate to what I see in-game today.
With that said, someone identified a case where the code I've removed as part of this PR was making face poses less accurate. See https://discord.com/channels/866821077769781249/1259962686649794623/1259962686649794623. This PR should fix it.
Note Reminder to @halfuwu, this commit and https://github.com/halfuwu/FortnitePorting/commit/ecee2770e7af813234cd23efcac1cd1e769950ce should be cherry-picked to
v3
.