osuKeysoundSplitter

How to

• Download the latest release.
• Generate a voice-only 16bit flac from a song and its instrumental version.
• Also export both the synchronized instrumental and complete song separately (@192kb/s).
• Import the instrumental in osu! and time it.
• Add the voice flac inside the map's folder.
• Make a copy of the diff and make it use the complete song's mp3. The name of diff should be suffixed with the flac name: if your diff is called NM, and the flac voice.flac, call this diff NM -voice.
• Place notes over the voice on this diff.
• Place bookmarks at the end of mapped sections, where the voice stops. They can also be used to cut breathing sounds from notes. See how they are placed in this map:
• 00:10:898 (3) - "ni", will play when the note is hit / 00:11:131 - breathing sound, will play automatically / 00:10:898 (3) - "a", will play when the note is hit.
• 00:21:131 (3) - "de", will play when the note is hit / 00:21:596 - stops the "de" sound, nothing will play here / 00:21:828 - breathing sound, will play automatically / 00:22:293 (1) - "a", will play when the note is hit.
• 01:21:131 - end of last mapped voice sample.
• Place keysoundOgg.bat/keysoundWav.bat and the jar in the mapset's folder and run keysoundOgg.bat or keysoundWav.bat depending on the format you want. You can also use this command:

java -jar "osu!KeysoundSplitter.jar" "mapsetPath" keysoundsOffsetInMilliseconds encodingFormat

FAQTS

• Yes, it works with any audio track, it doesn't have to be voice.
• Yes, you can have multiple keysound sources in the same map by having multiple flac and multiple diffs, each suffixed by one of the flac name: Diffs named NM -voice, NM -guitar with voice.flac and guitar.flac will be combined into NM.
• It works with osu!, ctb and osu!mania.
• No, it doesn't separate different instruments from a complete song for you.
• No, there's no user interface yet.
• No, it's only for .osu files, so no .bme.