Trendpop often processes tens of thousands of new TikTok sounds per day with our Music Reconciliation algorithm, the part of our system responsible for linking TikTok audios to the real music tracks that they are part of. Music reconciliation is a very difficult challenge for a technical system. Even a human can have a tough time hearing or understanding exactly what musical track a TikTok sound came from! Nevertheless, the Trendpop team is dedicated to constantly improving the effectiveness of our matching system.
Here are a few common reasons that our system may miss a match:
Note: This is an actively developed system at Trendpop. These common cases are documented as of May 2022, but subject to change due to improvements we're making every day.
The TikTok sound is an "official" release, but the song or artists are named something differently on TikTok than they are on a digital streaming provider.
Sometimes, a sound will end up named something completely different on TikTok than it is on DSPs. For example, an official sound on TikTok might be named "Song (TikTok release)", but on Spotify it's just named "Song". This can make it difficult for our algorithm to confirm a match, especially if the name is generic (like "Song") and can match many different music tracks.
The TikTok sound is a user-uploaded sound, but the sound is sped-up / tweaked / slowed-down / reverbed excessively.
For user-generated sounds, Trendpop will run a subprocess we call audio fingerprinting on the track where we try to pick up signals that a track matches audio. Trendpop uses state-of-the-art tech and a variety of third-party providers to match as much as we can, but our algorithms are not perfect and we do not always catch it. This is an active area of development for us, and we're always exploring new ways to improve our algorithm here.
The TikTok sound is a "pre-release", or not yet released on digital streaming providers.
In certain scenarios, artists or managers may "tease" tracks on TikTok before they release them on DSPs. This makes a great marketing tactic, but also very difficult for Trendpop to make a sound match because we rely heavily on DSP metadata. That being said, our algorithms do pick up the match when the sound gets released on DSP's, so typically you can wait until then for the system to stabilize. Alternatively, reach out to support@trendpop.social and we can manually add the prerelease metadata for you before we get it from DSPs automatically.
Comments
0 comments
Please sign in to leave a comment.