How to serve a million songs in-game for Billions of users

As the music game market expands, companies are actively seeking innovative solutions to streamline thedelivering music content. Amanotes, a leading interactive music company, stands out as a prime example of effectively addressing this challenge. With a staggering demand to cater to 12 million active users every day, Amanotes has developed an exceptional Music Solution to handle the intricate process. This solution involves taking raw input data from labels and content creators, processing them according to Amanotes’ standardized format, and preparing them for end users of our games.

In this post, we are going to discuss our suite of solutions to store, produce, distribute, and analyze a large amount of musical content and music data for multiple apps.

1. Music data storage

Storing millions of song data poses significant challenges for Amanotes, ranging from sourcing data from diverse providers to handling duplicate entries and dealing with corrupted music metadata. Let’s delve into how Amanotes addresses these problems.

• Uncategorized song data from different labels: Receiving music data from different labels requires Amanotes Music Database to create data pipelines, data transformation, and mapping processes to ensure that the data from different sources aligns with Amanotes’ existing database structure. This will ensure data accuracy across multiple internal Amanotes Music Databases. Additionally, a music data dashboard will also be set up to monitor metadata SLA (Service Level Agreement) (e.g. data coverage).
• Duplicated song data: Amanotes developed robust mechanisms to identify and manage duplicates effectively. By implementing techniques that allow for the deletion and merging of duplicated data, Amanotes ensures that its Music Database remains as clean as possible. This not only improves the user experience but also streamlines internal processes.
• Corrupted music metadata: Amanotes works closely with music providers to resolve any suspected incorrect data. However, even with these efforts, some degree of manual maintenance is still required. We have developed web-based interfaces that provide intuitive music data management and administration solutions. These interfaces facilitate efficient handling of corrupted metadata, allowing for timely resolution and maintenance.

2. Audio processing:

Processing music on a large scale also requires efficient and effective tools and strategies. Amanotes has developed an auto-content production system that can automatically generate a level design file using machine-learning technique to make it compatible with various Amanotes gaming applications. This technology can handle diverse file formats, converting and transcoding raw audio to a standardized format of different quality levels and lengths. Besides, to ensure security, the audio files are encrypted and can only be decrypted by Amanotes applications. Additionally, we also provide a toolkit that turns each audio file into a gameplay-adaptable version based on musical elements so that it fits different gameplay styles and difficulty levels.

3. Song Distribution

After ingestion and processing, the songs are added to Amanotes Music Catalog and are made available to millions of Amanotes app users. Users can search for and play songs on-demand through the free or premium versions of the different games.

User can search for their favorite songs in the Tiles Hop game

• Our music catalog is always up-to-date with new releases from publishers and labels. We promptly publish newly released songs to end-users. End-users automatically see the newest songs that Amanotes has licensed for their region in nearly real-time.
• Additionally, our automated pipeline continuously receives and processes daily updates to be ready for urgent takedowns, if necessary.
• We also ensure the delivery pipeline is stable and meets SLO (Service Level Objective) with our monitoring system.

4. Analysis

After delivering files to end users, we move on to examine the usage data collected by game players. Establishing metrics to assess the quality of the content delivered to end users is an essential aspect of enhancing your product and elevating the overall gaming experience. The analysis can come from 2 sources:

• Primary data collected from users: We analyze usage data from game players by choosing metrics to measure content quality, such as Song Performance Score. We collect data from player interactions, also identify patterns & trends to better understand quality and player behavior. This information helps to make continuous updates and improvements in new versions.
• Secondary data from trusted third-party metadata providers: We use the additional data in various ways, such as improving recommendation systems, enhancing playlist curation, or gaining insights into music popularity trends. It’s our daily job to make sure that the data is appropriately utilized and integrated into Amanotes product or service offerings.

In conclusion, our Music Platform is designed to create a smooth and efficient system for delivering music from different providers from Amanotes gameplay files to game users. This process ensures that the files are delivered on time, while also maintaining high quality and scalability. Managing huge music song data is challenging, however, with the support of advanced technology in music content management, Amanotes is able to deliver the best service for both internal and end users.