* wip
* fix songs again
* don't use lists of video and audio input files
* test concat command output
* move concat pipeline
* start to use ffmpeg state
* add ffmpeg state and audio state
* audio state is required
* attach input options directly to input files
* move filters to input files
* add watermark support
* generate blurhash for all local artwork
* use blurhash song background if available
* only write blur hash to disk once
* use multiple blur hashes
* update changelog
* fix song detail outline
* reset song metadata (artwork)