* Lsmas/vapoursynth based splitting (massive disk savings and MASSIVE stability improvements) [Linux only, for now]
* Time based splitting. The classic from neav1e comes to qencoder! But unlike regular time based splitting this time splitting is slightly smarter, relying on maximum kf distance to ensure it's creating barely/no unneeded keyframes
* ffmpeg based splitting. Safer than pyscenedetect (but slower and possibly slightly less efficient) and built RIGHT into ffmpeg.
* uses a standardized, smart, and public backend - parallel encode. Parallel encode is a tool for ANY developer looking for parallel encoding functionality.
* no more av1an code in qencoder, it's all moved to a separate backend
* fewer deps for both qencoder and for parallel encode - more optional deps as well - good for everyone basically.
* less crap printed to the terminal/console window
* ADD FOLDER TO QUEUE
* QUEUE JOBS - encode multiple queue items at once
* Encoding now completes automatically.
* Target VMAF works with scaling/cropping video.
For Linux/Windows users who want to avoid large binaries- install the latest version with pip: `pip install qencoder==2.0rc3`
This RC contains ALL the features of 2.0 but with less testing than I feel comfortable with to release as a proper 2.0 version. I'm still confident that it is as stable as previous qencoder versions, just not as stable as a proper 2.0 should be, which is 100% stable with 0 failings.
(okay actually it's rc2 but it's basically functionally identical to the version on pip)