musicube blog image

Complex musical features

March 29th, 2022

This is the last article dedicated to the musical categories that musicube’s AI can identify and tag. It will come in handy soon when we introduce our automatic tagging system. An easy-to-use online tool that will allow you to upload your mp3 and get it tagged by our AI.

We talked about basic musical features, vocals, main instruments, rhythmic features, and much more. In this last week dedicated to our categories, we will be talking about complex musical features our AI can identify. We consider them complex features because they consider particularities of the track that imply several dimensions: tempo, loudness, scale, etc.

Are you ready to fully master musicube’s categories? Then, let's get started!

1. Energy

The energy in musicube’s AI relates to the loudness of the track. The loudness is related to the track's volume, and, in audio, this is normally measured in decibels. In the sound systems, we can set up the volume manually. In contrast, the loudness is imprinted in the song file and undergoes standardization rules. This is very important in music production because it lets us compare one song to another independently of the sound system we are using to listen to the track. There is normally an established “ceiling”, above which the sound systems will start to distort the sound. The closest the sound file is to that ceiling, the more energy our AI identifies. Depending on that, we classify the songs as very quiet, quiet, moderate, loud, very loud, and dynamic.

These tags talk about how close the average loudness of the track is to the distortion ceiling. The exception are the dynamic tracks. Our AI considers that a track is dynamic, when it alternates loud sections, with quiet or very quiet ones, making the track rich in “dynamic range”.

  1. Quiet: Happiest Year by Jaymes Young
  2. Moderate: Time by Hans Zimmer
  3. Loud: High Hopes by Panic at the Disco
2. Tonality

The tonality talks about how melodious the song is regarding to how many things are happening on the track melody-wise. When the melody of a track does not present many variations, we consider it monotonous. When the melody has variations in note range within the scale, we consider the track melodious. Thus, depending on the tonality, our tracks get tagged as:

  1. Monotonous: Cooler than Me by Lucky Luke
  2. Moderate melodiousness: Stunning by Curtis Waters
  3. Melodious: Here comes the Sun by The Beatles
3. Harmony

Some notes fit together nicely. Either they flow seamlessly from one to another, or they sound great when played together. This normally depends on the relative distance between those notes. When the notes present on the track sound great together, the AI considers the song harmonious or very harmonious. If they do not sound great together, the track gets tagged as dissonant. As mentioned previously, dissonancy is not necessarily a bad feature. In many cases, dissonances serve to surprise, or purposefully create an unpleasant feeling such as tension, confusion, or mess.

  1. Dissonant: Blade Runner by Hans Zimmer
  2. Moderate Harmonies: Hammer to Fall by Queen
  3. Harmonious: Unintended by Muse
4. Space

The track's space talks about how wide a song sounds. It relates to the distance between the notes the song is using. When a track uses only a few notes that keep a short relative distance in the song's scale, the song is considered compact. In contrast, if the track features notes that are very far away in terms of frequency (very low and very high), the track is considered wide.

  1. Compact: Go Crazy by Chris Brown
  2. Moderate Space: Living in a Ghost Town by The Rolling Stones
  3. Wide: Intro from Wembley Stadium by Muse
5. Texture

This is probably the category that considers more variables at once. The texture talks about how much it is happening in the song on harmonic and rhythmic terms. If a song has a rather monotonous tonality and steady rhythm it is tagged as thin. But when a song has big dynamic range, groove and rhythm, it is tagged as full.

  1. Thin: The Problem with the big picture is that it is hard to see by Mayday Parade
  2. Moderate Texture: You did good by Alan Silvestri
  3. Full: Starlight by Muse

We have documented, post after post, all the categories our AI can analyze and apply to the tracks. We hope we have provided enough details about the type of information you can rapidly identify and store with the help of our engine. Having a fast, automatized, and objective way to tag the songs is useful to keep your music catalog organized. Adding the different tags as part of the track’s metadata descriptions allows you to rapidly search music matching the desired criteria, and generate a list of ready-to-go tracks in any situation. Shortly, we will inaugurate a section on our website that will allow you to upload an mp3 of your choice, analyze it through our engine, and retrieve the same mp3 with the categories information attached as part of the mp3’s metadata. We are giving the online tool its final touches, and we are looking forward to sharing it with you, so: stay tuned for updates!

In the meantime, keep getting used to the new technologies and enjoying great music!

Did you miss any of the other blog posts about musicube's categories? Here they are: