talosh/flameTimewarpML

32bit float.

Opened this issue · 3 comments

Hi Talosh...

If we have 32bit fp source footage, it comes back as 16bit fp. Would be great to keep that bit depth.

Thanks,
Alan

Hi Alan, I couldn't give it a very through test yet being a bit busy finishing the project I'm currently on, but those 32bit preservation logic is there on a main branch now, you can check it if it works on linux, I've been testing it only on Mac at the moment

Getting this warning on export with Flame 2022 CentOS 8.

IMG_1220

Color wise though, I think it is all working as expected.