The Value of Stem Separation
Recently, there have been countless incredible developments in music technology that change the process of music production and its accessibility. One technology I have been particularly interested in recently is stem-separation. A “stem” is a track of one individual instrument or component throughout an audio piece. For instance, songs can be divided into individual stems for the piano, vocals, guitars, basses, or drums throughout a song. Stems can also be used to define components within these instruments themselves–guitar tracks can be divided into stems that separate chords, leads, and other top melodies into separate files. Stem-separation is the process of automatically dividing an inputted audio file into its different parts, and its value and application spans far beyond remixes and simple amusement.
Websites like LALAL.AI and Splitter.ai–both of which offer free online services–have been popular tools for those interested in stem-separation. While it is entertaining to separate songs into its subsidiary components for personal enjoyment, stem separation can actually be used as a great educational tool for novice music producers. Exposure to different stems of a song not only helps producers understand the complexity in instrumentation and different layers of a song, but–when paired with other technology–can cleverly be used to help producers understand the actual melodies of the stems as well.
Melodyne is a tool that I’ve discussed on this blog previously, and I feel it has an interesting application in this context as well. Melodyne is a piece of software made for vocal tuning and it works by allowing the user to analyze each individual note of a vocal audio file and manipulate it so it is in tune. Here is a picture of Melodyne in use, where each rectangle represents a note of a vocal take, similar to a piano roll.
Although Melodyne is initially intended for vocal takes, because it operates based on an input of any audio file, Melodyne can also be used to analyze any audio stem from a song. This means that, with access to both Melodyne and any free stem-separator, users are able to not only listen to individual components of a song, but also visually see the notes of those components. This gives a deeper understanding of a music composition, which users can use to apply to their own music. For instance, popular YouTuber Aiden Kenway uses Melodyne in the video linked below to analyze the chord progression in the song “SAD!” by XXXTENTACION [watch at 0:33].
Kenway uses Melodyne in the video to not only show the original chord progression from the song, but also modify it how he wants to make the song his own.
Stem-separation is such an interesting technology to me because it has not only been helpful in remixing songs–where I need to remove instrumentals–but it has also helped me understand the actual composition of songs, which I have previously been unable to do. Although Melodyne has existed for over 20 years now, I find it even more interesting how something like pitch correction software can be used in a seemingly unrelated context like this to understand songs more deeply.
Image from djmag.com
