VASH | Perceptual Hash project for Videos | Media library

VASH is a C library typically used in Media applications. VASH has no bugs, it has no vulnerabilities and it has low support. However VASH has a Non-SPDX License. You can download it from GitHub.

With the advent of filesharing more and more people enjoy watching movies of tv shows on their computers, on top of that more and more people prefer buying digital copies over physical disks. The obvious advantage being that the media is always readily available on the harddrive for local consumption or streaming to other devices. Just as many people now store thousands of music files on their computer this may very well be the case for full-length movie files. And just as with music there will be a need organize and index this large number of videos in one way or another. This trend is underlined by increasingly popular HTPC and Media Center software like XBMC, Plex, Servio, MythTV and myriad of similar software with the sole purpose of organizing, indexing and playing videos, music and images into beautiful easy-to-navigate libraries. For music files ID3 tags are generally available as part of the file itself so the HTPC software so music libraries can be organized automatically and easily navigated by artist, album, year, genere, or any other type of information. On top of that music identification based on Audio Fingerprint technology has become commonplace through services like Apples commercial iTunes MusicMatch or the open-source MusicBrainz AcoustID database. So even music that has not been tagged at all can easily be identified and associated with metadata online. For video the situation is dire. Not only do the most popular video formats not even support any kind of tagging mechanism, even the newer formats like mp4 that would support tagging are generally never tagged. HTPC software can therefore only rely on the filepath to infere at least some pieces of information, and in many cases even that is impossible due to bad naming. The de-facto standard in HTPC software is that media already needs to be perfectly named and organized so that the HTPC software can look up additional metadata via webservices like TheMovieDB or TheTVDB. The process of manually organizing thousands of files is extremely time-consuming and even then there will be a risk of misidentified files if the names don’t match the online database entries verbatim. This project proposes a perceptual video hash algorithm to identify videos based by looking at the video content much like a human being would. Using this hash HTPC software would then be able to quickly and accurately identify video files and fetch additional metadata from online databases without any manual input greatly enhancing the HTPC experience.