musicube blog image

Music Supervisors and AI - a relationshop made to last

April 28th, 2022

Hours of music listening & search on your back. Entry folders. Non-stop-beeping with new music submissions. Loads of hard drives (and shelves) full of tracks (and, cassettes, vinyl, CDs…) storing hundreds of potentially syncable tracks. Temp love with a track that you know in advance you won´t be able to license and that will be extremely hard to replace. The perfect track found online… without any sort of rightsholders' information. The perfect track found online… whose rights are shared by an endless list of people. The perfect track stored in your head that you have NO CLUE where to find again…

Does any of these statements sound familiar to you? Then you are probably a music supervisor.

In that case let us ask you: have you already started taking advantage of AI to make your job a bit easier? If the answer is no, you should consider it.

There are several ways in which an AI can help you. Most of them are related to two key concepts a music-focused AI engine can get very good at: sound analysis and metadata enrichment. Do you want to know how? Just keep reading!

AI can help to find the type of track you need:

When it comes to sound analysis, AI engines can be trained to analyze the digital sound waves and identify countless aspects related to the music style, mood, main instruments, nature of the sound, etc*.

Once these aspects are identified, you can access them and/or incorporate them as part of the track's metadata. This enriched information about the track can be very useful to:

  1. Easily organize your personal music library to find the tracks you need using the metadata keywords.
  2. Extract the metadata of the tempted tracks to be more specific about the type of music you will need.
  3. Perform better searches within your catalog, external catalogs… or track requests within your community.
AI can help to identify the info you need from the tracks you’d like to use:

There's another aspect of AI and machine learning that not so many people think about, but that can be very helpful too: finding key metadata. There's so much data available online that is not being used because there's no system in place to take advantage of it. That's when the power of AI and machine learning come into place. Making use of computing power, one can design algorithms able to track that data and make use of it. There's a track you want to use without contact information? There's no available spreadsheet showing rightsholders and/or publishers' information? There's no ISWC attached to the ISRC? In most cases, that information is out there. The algorithm can clean up the data and summarize the different spellings of an artist's, rightsholder or a contributor's name and find more information about them. It could also match those names with the ISRC, and get a correct data set enriched, not only with the aforementioned information about the sound wave, but also with alternative titles, rightsholders information, versions of the song, and much more details that could, indeed, allow a more efficient workflow.

AI can help to organize your own music library and the submissions you receive standardizing the searching keywords.

You've probably found yourself listening to a track submitted in response to a particular request, that did not fit the features you needed for the project. Imagine if you could automatically pre-classify the tracks, you receive according to certain musical features of your choice. This way, you wouldn't need to spend time that you don't have, checking submissions that won't do the job. AI can easily do this for you too. Once the AI is trained and running, you could automatically analyze your recordings and have them immediately separated and ready to be found, according to the features you need. By doing this, you could rest assured you will be using your time to focus on the potential winners, and get to the discovery mood when you actually have time for it.

Doesn't it look like a kind of magic?

There's a lot to do implementing AI tools in your normal workflow. Possibilities and combinations are almost endless.

If you want to start experimenting with an AI that could help you do part of the job (for free!), check out our musicube cloud. Our cloud is a web-based system that allows you to upload tracks, automatically tag them, and have them stored, ready to search (based on cool metadata), and available for you (and others) from anywhere and at any time.

If you liked what you’ve read and you started to think about a customized solution that could fit your needs and make your job easier, don't hesitate to contact us. We are always ready to learn more about your needs and we are looking forward to putting all the power of our AI in your hands.

That's all for now. More details coming soon. Remember you can start trying out our musicube cloud, just register on musicu.be.

Thanks so much for being there!

*To know more about some of the categories our AI can detect, check the following posts on:

  1. Basic features: primary mood, pleasantness, engagement, and valence.
  2. Rhythmic features: rhythm affinity, bpm, grooviness, arousal, and tempo class.
  3. Musical Components: main key, scale, dominant instrument, vocals, roughness, and sound generation.
  4. Complex musical features: space, harmony, texture, tonality, and energy.