Studio Fontana

Challenge

Studio Fontana was looking for a system that would help their clients to find music in a better way. Finding the right music for your clients is the biggest challenge in the synchronization business. Their biggest clients, advertising agencies, broadcasters and production studios do not want to experience with the hassle of music search, “they just want the music” as Martin said.

The normal way of searching has always been to use keywords filters or text search engine. One of the most difficult obstacles to overcome when tagging music is consistency. Studio Fontana gathers tracks from different sources and sub-publishers. What is labeled Electro by a partner might be tagged Electronica by Studio Fontana’s team. With such a big amount of tracks and such a diversity of sources, there was not enough manpower to retag the entire catalog.

Even with a perfect tagging work, it would still remain a user-experience problem for their clients. Most of their users do not have a musicologist background, and are not familiar with music expert taxonomy. What’s the difference between “happy” and “positive” ? There is always a loss when you translate the musical feel in your head into keywords.

Solution

Studio Fontana was looking for an audio search technology for a long time. Building an internal system tailored for the industry was a long and uncertain way to go. This is where Niland comes into the picture.

Every time new tracks are uploaded to Studio Fontana’s system, they are sent to Niland API. Using advanced techniques of machine learning and signal processing, Niland’s technology analyzes the tracks and makes its own acoustic representation of music tracks.

Every time a user wants to perform an audio search uploading an mp3 to Studio Fontana’s website, this track is sent to Niland API. This track is being analyzed on the fly by Niland API and matched tracks are then returned within a few seconds. This integration has been running for a year with no scalability issue since then.

Implementation and Results

Studio Fontana implemented Niland’s technology into its new version of hudebnibanka.cz. In this online search engine for production music, Studio Fontana built a user interface which allows its clients to drag and drop an mp3 or paste a youtube link. Most users already have a reference track in mind when they log into hudebnibanka.cz. The audio search is a big short in the search process, clients get closer to their needs much quicker. Some track searches used to take 10 to 15 minutes and are now reduced to a few seconds.

Studio Fontana attracted new clients with the launch of its renewed platform. Advertising Agencies were particularly sensitive to the Audiosearch and the time they saved when using it. Niland’s technology also helps regular clients to find new alternatives for their projects. A lot of clients use a particular track as a “safe choice” to play in the background of interviews or TV shows, but eventually they become tired of it. They can now easily find similar tracks which leads to repeat business for Studio Fontana.

Smarter Together

Using the Niland's dashboard, Studio Fontana can change the similarity and configure the relevance and the ranking of any search query results. These feedbacks are then returned to Niland and used to make the algorithm deliver even better results. This feedback loop was essential to get better results for Studio Fontana's clients and to tailor the algorithm for this industry.