Skip to content

The Anime Scripter - Release [2.0.1]

Latest
Compare
Choose a tag to compare
@NevermindNilas NevermindNilas released this 11 Jan 23:58

The Anime Scripter:

New

  • Added rtmosr-directml and rtmosr-tensorrt upscaling method by @umzi2.
  • Added slow_nvenc_av1 encoding preset.

Improved

  • TensorRT and DirectML based depth maps have been significantly improved.
  • Changed TAS.log -> TAS-Log.log for easier debugging.
  • Changed ffmpeg.log -> FFmpeg-Log.log for easier debugging.
  • Generalized some logs.
  • Removed unwanted terminal prints generated by ffmpeg when encoding with x265
  • Imense improvements in the realtime playback, with the amazing tool from mpv realtime preview of the content is now plausible.
  • Switched from FFmpeg to MPV playback for --realtime
  • If --realtime is active encoding will no longer take place in order to save on system resources.
  • Improved the TRT Shape printing when building engines to be a bit more compact.
  • Improved the handling of the ffmpeg encode command.
  • Improved reusability of repeated calls, specifically logging and printing.
  • Added a forced sys.exit() if FFMPEG crashes during encoding for any reason in order to avoid deadlocks.
  • Changed sub from srt -> copy, which should retain metadata from the base input.

Removed

  • --depth_quality has been deprecated for directml and tensorrt, now all models will run internally at 518x518 to better match the model resolution and quality.

Fixed

  • Shape issues with Segment Anime TensorRT
  • Issues with rawvideo and celux
  • Issues with 16bit workflow and nvenc_h264

Adobe Edition

New

  • Added Take a Screenshot to the Toolbox.
  • Added an About Panel which replaces the previous dialog on the top right with more info regarding TAS.
  • Added System Requirements in the About panel.

Improved

  • Improved the Extracting TAS... indicator, the ... should now 'progress' just so that users know it's not actually stuck.
  • Slight adjustment to the backend for ease of reusability, more work needs to be done in order to make TAS slightly more manageable.
  • Defaulted depth quality to low
  • Added error handling when the desired input was an image, TAS should now generate a toast alert letting the user know that it only works with images.
  • More tweaks to how UI elements get generated.
  • Improved descriptions of many features available in TAS.
  • Significant improvements to the UI/UX, adding labels, icons and descriptions.
  • Added direct credits to the Anipy-CLI team in the download an anime section.
  • Added direct credits to the ytdlp team in the youtube downloader section.
  • Changed the update functionality from a dialog to a toast.
  • Refactored even more code.

Fixed

  • The Remove Background segment would not update upon selection of different elements.
  • Fixed a repeated unncessary task which will now happen only on the initial mount of the UI.

Full Changelog: v2.0.0...v2.0.1