AI music generator lets users create mashups without musical experience

Introducing Mixboard, an innovative tablet application developed by Professor Gil Weinberg and his team at the Georgia Institute of Technology’s Center for Music Technology. This groundbreaking app allows users to create their dream songs, even without any musical or editing experience. By harnessing the power of artificial intelligence (AI), Mixboard combines songs in unexpected and exciting ways, bringing together artists and genres that users might not naturally pair.

According to Professor Weinberg, the director of the Center for Music Technology, the goal of Mixboard is to make music creation accessible to everyone. While not everyone may have the technical knowledge or musical theory required to compose a piece of music from scratch, almost everyone appreciates music. With Mixboard’s user-friendly interface, individuals can effortlessly place songs in different contexts and create something truly unique.

Credit: Georgia Institute of Technology

The researchers behind Mixboard presented their work at the International Conference of New Interfaces for Musical Expression (NIME 2023) in May, outlining their vision in the paper titled “Mixboard—A Co-Creative Mashup Application for Novices.”

Creating a mashup using Mixboard is a simple and intuitive process. Users can select up to four songs from Spotify or Mixboard’s music library and drag and drop them into different segments such as vocals, bass, chords, and drums, for a duration of up to 32 bars. The app provides a visual experience that allows users to follow along as their chosen songs are mixed, with the final result rendered within seconds. Once completed, users have the option to download and share their unique tracks with others.

A key feature of Mixboard is its AI capabilities, which handle the technical aspects and music theory involved in creating a mashup. The app utilizes open-source music information retrieval libraries to determine the tempo, key, and individual music stems of the songs. This information helps break down the songs into their constituent parts, allowing the algorithm to stretch or transpose them as needed to ensure a cohesive blend. The researchers have also included templates of basic song structures to assist users in crafting appealing tracks.

According to Professor Weinberg, the integration of AI into Mixboard not only simplifies the music creation process but also provides a source of inspiration that humans alone might not have envisioned. By combining the user’s creativity with the AI’s input, the resulting mashups can be truly unique and innovative, offering a musical experience that goes beyond individual capabilities.

Mixboard represents a significant step forward in empowering individuals to explore their musical creativity, regardless of their background or expertise. With its seamless interface, AI-powered capabilities, and emphasis on co-creation, Mixboard opens up a world of possibilities for music enthusiasts and aspiring artists alike.

The researchers behind Mixboard, led by Human-Computer Interaction master’s student Tommy Ottolin, conducted user surveys to ensure the functionality of the app. They recruited 45 participants between the ages of 18 and 27, specifically targeting individuals with limited or no music mixing or composition experience. The users were given 30 minutes to test Mixboard and then engaged in a semi-structured interview with the researchers. Following that, they filled out a 20-question survey using a five-point Likert scale to rate various aspects of the app, such as creative control and user-friendliness.

Qinying Lei, a music technology master’s graduate, expressed her delight in witnessing users interact with Mixboard and the emotions it evoked in her. Seeing users captivated by the app and producing unique music filled her with pride, nervousness, and excitement.

As Mixboard progresses, the researchers are focused on refining its automation and enhancing its functionality on iOS devices. They also recognize the app’s potential for commercialization but acknowledge that licensing constraints pose a limitation. To navigate this aspect, Professor Weinberg sought guidance from Kobi Abayomi, formerly of Warner Music Group and currently heading Science for Gumbel Demand Acceleration. They believe music rights holders should view apps like Mixboard as marketing opportunities rather than solely licensing plays, emphasizing the app’s ability to increase affinity for specific types of music.

However, the primary goal of Mixboard remains the democratization of music production, aiming to demonstrate that anyone can be creative in this realm. Nitin Hugar, a music technology master’s graduate, shared his amazement at hearing the generated mashups improve with each modification to the algorithm. He also found joy in attempting to blend unlikely songs together, realizing that music from different regions of the world can be interconnected in surprising and harmonious ways.

Source: Georgia Institute of Technology

Leave a Reply

Your email address will not be published. Required fields are marked *