Media Management & Delivery
Fraunhofer Institute for Digital Media Technology
- © istock.com/Michael Kempf
Semantic Music Technologies
The Semantic Music Technologies research group develops technologies and solutions for automatic music analysis. These are used in music search and recommendation engines, but also in music games, learning software and in music production. The core expertise of the group is in signal processing and pattern recognition.
Main Research Topics
Music Search and Recommendation
The developed technologies make finding up new titles in vast music databases easier. The user can find titles similar to favorites either exploratively or targeted according to musical characteristics he enters. Using different technologies from the Music Information Retrieval research, describing characteristics are derived automatically from audio signals. Even suspected plagiarism can be technically checked this way.
Music Games and Learning Software
Karaoke and Music Learning Games are full of state-of-the-art algorithms for automatic music transcription. Using the transcription, notes played or sung are automatically recognized. This way, these applications can show the user in real time, whether some tune is sung or played correctly. In addition, the software helps to transform a favorite tune into a game or into practise notes. The tune search engine Query by Humming is based on this principle, too.
Today, it is standard in music production to support the creative process using software solutions. Mixing can be significantly simplifyed using rhythm and key recognition.