Frequency Stacking: Listening to Music Differently

Matthew Cembrano, University of Southern California, mattcembrano@gmail.com

Abstract: You could probably be listening to music better. This article details a simple yet thorough process that allows you to hear every sound in a song, understand its place in the mix, and better understand the artist’s intentions: Frequency stacking. Whether you’re a three-time Grammy nominee or someone who just listens to the radio in the car, there are benefits from this methodology for you.

About the Author: Matt Cembrano is a graduating senior at the University of Southern California who, despite majoring in Computer Science Games, finds his passion lies in music production. He releases music under the alias “MATT’S GONE MISSING.”

Keywords: Frequency Stacking, musical analysis, audio engineering, genre exploration, music production, frequency spectrum

Are you aware of how you listen to music? Take a second to think about it. What do you hear when you listen to a song? What stands out to you? Do you hear the lyrics first? The rhythm of the drums? The melody? You may understand the structure of the track, able to sense the chorus coming, or what the next phrase has in store. Maybe you’re a pretty musically oriented person, and you can break down the track into each of its individual parts, instrument by instrument. I used to pick apart a song by hearing the bassline first, then the drums, then the vocals, then the rest of the track as a whole, but that doesn’t quite work for songs that don’t have basslines, drums, or vocals. After eight years of producing music and months of DJ-ing, I find that I now listen to music completely differently, in a way nobody else has ever told me they related to. This method of listening can help casual music enjoyers piece apart new songs and develop a deeper understanding and appreciation for music of any genre while also helping producers, artists, engineers, and DJs make, master, and play higher-quality music. It’s already helped me immensely in all four of those roles and has also allowed me to enjoy music more thoroughly as a listener. The method is simple and applies to all music: You scan the song you’re listening to from the lowest frequencies to the highest frequencies, and each time you encounter a new sound, you add it to a stack of sounds in your head. For lack of a cooler name, I like to call it “frequency stacking.”

Some Science:

To understand frequency stacking, we must first understand the role that frequencies play in sound and music. Frequency, usually measured in Hertz (hz), is “the number of cycles or vibrations undergone during one unit of time by a body in periodic motion” [1]. Sound, as we may know, travels in waves. When a sound wave is long, it has a lower frequency since it will take more time for one cycle of the wave to pass. When a sound wave is short, it will have a high frequency, since many cycles of the wave will pass in a shorter period of time. Hertz is the measure of how many cycles of the wave passes per second, so if a sound has a frequency of 5000hz, it means that the wave will oscillate (go through cycles) at a rate of 5000 cycles per second. An easy way to visualize this is to imagine you’re a surfer sitting out in the ocean and waves are coming in at a steady pace. Each time a wave passes, you go up and then back down. The number of full up-and-down cycles you experience as the surfer per second is the frequency, in hertz, of the waves.

The range of human hearing is somewhat limited. We can only hear frequencies from around 20hz to 20000hz [2]. In comparison, elephants can hear as low as 16hz and dogs can hear as high as 45000hz [3]. This is why we can’t hear dog whistles, but dogs can- the whistle is somewhere between 20000hz and 45000hz. In music, the deepest, lowest sounds are closer to 20hz, and the highest sounds are closer to 20000hz, to match up with the range of human hearing. Generally, modern music that is mixed well will balance almost all the frequencies between the low-end and the high-end so songs don’t sound overwhelmingly boomy (too much low-end) or tinny (too much high-end).

For reference, this is what the frequencies from 20hz to 20000hz (human hearing range) sound like:

Keep in mind that you may want to listen with headphones or large, high-quality speakers, otherwise, you will likely not hear much of the low-end frequencies. Notice how the sound goes up in pitch as the video progresses. That is the relationship between frequency and pitch- higher frequencies have higher pitches. A D-note has a higher frequency than a C-note (assuming they’re in the same octave). A snare drum has a higher frequency than a bass drum. A trumpet plays higher frequencies than a tuba. If you listen to music at all, you likely already intuitively understand the relationship that frequencies have with music. It’s just pitch!

Though it may be easy to understand, it can be difficult to determine the frequencies of the sounds you hear in music. It might take some (or a lot of) practice. The best place to start is by figuring out where sounds lie on the frequency spectrum relative to one another. Try listening to a somewhat minimalistic song like Joji’s “SLOW DANCING IN THE DARK.”

At the beginning, do you hear the kick drum below the harp-like sound? Do you hear that harp-like sound below the synth that’s just playing backing chords? Do you hear the clap below that synth? Do you hear the hi-hats above the clap? When Joji starts singing, do you hear his vocals somewhere above the harp but below the hi-hats? Do you hear the bass that plays below everything else? If you can hear the relative positions of each sound, you’re ready to build a stack!

The Stack:

Now that we understand the role that frequencies play in music, we can start to visualize a frequency stack. Imagine each sound or instrument you hear is represented by a rectangle. We will be stacking these rectangles on top of one another, ordered by frequency. The highest-frequency sounds are going to be on top of the stack, and the lowest sounds will be at the bottom. The larger the range of frequencies a sound covers, the taller the rectangle. For example, a piano may play very low and very high notes, which will cover a larger part of the sound spectrum than the hi-hat of a drum kit, which will cover a narrow band of the high-end. The piano’s rectangle will be quite tall, and the hi-hat’s rectangle will be quite short. In the stack, the hi-hat may even overlap the piano’s rectangle, which is okay! I also like to color the sounds based on how loud they are. The louder that sound is, the more red the rectangle, while the quieter sounds will be more blue. Medium-volume sounds are yellow or green. You can also use brightness to represent loudness- loud sounds will produce bright rectangles, and quiet ones will be dark. Whatever works for you.

To begin visualizing the stack, I like to start with the lowest frequencies in a song and move upwards, but going top-down works too. What matters is that you scan the frequency spectrum continuously in either ascending or descending order. Let’s use “SLOW DANCING IN THE DARK” as an example again. Focus on the part between 21 seconds and 42 seconds in the song. What noises do we hear at the very bottom of the spectrum? The bass, and above it, the kick drum. Let’s put those in the stack. The bass is quite present, but not overpowering, so it’ll look yellow-orangish. The kick sounds like it’s above the bass, and it’s pretty loud, so we’ll make it red. Moving upwards, we hear the harp arpeggiating up and down, so it covers some lower frequencies but also some mid-to-high ones. This means the rectangle will be tall. The sound is in the background and not the focus at the moment, so the rectangle is greenish. Add it to the stack. Next, we encounter Joji’s vocals. There’s some high-end crispiness to the vocals, but the meat of the vocals feels like they’re in the mids-to-highs, in the same range as the upper parts of the harp. This is okay! We can just overlap the bottom of the vocal’s rectangle and the top part of the harp’s rectangle. The vocals are quite prominent, so we’ll make it red. Next, we encounter the clap, which doesn’t cover a big part of the spectrum, and sounds like it might be in the same range as the upper part of the vocals. We can just smack a short rectangle down, overlapping the vocal rectangle in the upper half where we think it belongs. It isn’t super loud, so let’s make it yellow. Lastly, on top of everything else, we have the hi-hats. They sound about as loud as the clap, so we’ll make them yellow as well. I don’t hear anything higher than that, so I think we covered every sound. The stack should be complete, and it should look something like this:

By scanning upwards slowly and deliberately, seeking out each subsequent sound in the frequency spectrum, we were able to pick out every instrument used in that segment of “SLOW DANCING IN THE DARK” to build a frequency stack. Notice how there are loud sounds occupying the lows, mids, and highs of the spectrum. This signifies that the track is fairly balanced, which is something that you want as an artist or producer.

What’s extra cool about frequency stacking is that it works for music of any genre. Go listen to Benny Benassi’s “Satisfaction”.

Do you hear the kick, the vocals, the clap, then the hi-hats? On the drop, can you tell that the lead falls above the kick and bass but below everything else? Take a stab at Joe Hisaishi’s “Merry-Go-Round of Life”.

Notice how the lower end is filled with bass (the big instrument that looks like an oversized cello), followed by more stringed instruments like cellos and violins and violas layered to cover the low-mids, the mids, the high-mids, and the highs? Do you hear the piano on top of the strings in the mid-to-high frequency range? Listen to Lizzy McAlpine’s “Pancakes for Dinner”.

Do you hear the bass guitar under the acoustic guitar? Then Lizzy’s crisp vocals on top of those layers? When the drums come in, do you hear the kick drum under the bass guitar, and the hi-hats behind the vocals (similar frequencies, but quieter)? It might take another scan to determine the order, and you might even be wrong- but that’s okay. The point of listening to music with the intent to build a frequency stack isn’t about perfectly identifying the frequencies at which sounds are located. Rather, it’s a method of listening that helps isolate every single element in a song and find its relative location on the frequency spectrum. Listening this way also helps identify balance (or lack thereof) in a song. Remember- music doesn’t tend to sound great with too much or too little emphasis on the lows, mids, or highs, and your favorite artist is certainly aware of this. See if you can figure out why your favorite artist added each element that they did in your favorite songs, paying careful attention to where in the frequency stack each sound ends up.

What am I supposed to do with this?:

For the average, non-musician, music-enjoyer, frequency-stacking can allow one to develop a deeper understanding and appreciation of music of any genre. It’s cool to be able to hear every sound that composes a song, and a lot of appreciation emerges from the realization that each and every sound heard was a deliberate decision made by the artist or producer. Frequency stacking is also a useful tool that allows one to quickly understand and appreciate new and unfamiliar songs and genres. By following the principle of frequency balance- the idea that there will be a strong yet balanced presence of low, mid, and high-frequency sounds in a song- a listener can understand the role that a sound plays in a new song or genre by categorizing it by frequency range and comparing it to the sounds present in the same frequency ranges of more familiar genres. For example, I listen to a lot of house music but I don’t tend to listen to much jazz. But when I do, I can make comparisons between the genres based on the sounds that occupy the same frequency ranges and thus play similar roles in the song. This helps me better understand the unfamiliar genre.

For example, if I were to listen to Duke Ellington and John Coltrane’s “In A Sentimental Mood,” which I discovered conveniently located on a Spotify playlist called “Jazz Classics,” I’d find that there is a plucked bass being played to fill in the low-end, which fills the same role that bass synths usually play in house music. The piano and saxophone fill the mid-range frequencies, but the sax is louder than the piano. It can be argued that the sax plays a similar role to a lead synth in a house track, and the piano plays a similar role to a counter-melody played on a softer synth. The ride cymbal, snare, and hi-hats in “In A Sentimental Mood” are all drum sounds that are already often found in house music, though played in a very different pattern. By making these comparisons with a focus on frequency between a jazz song that’s new to me, and a genre I’m more familiar with, I can begin to understand what each instrument does for the song, how these instruments work together, and thus allows me to better understand jazz from the perspective of another genre.

Though the benefits of frequency stacking stop somewhat short of musical ascendance for non-musicians, there are massive tangible benefits for those who make or play music. DJs, for example, need to know what sounds will get cut or boosted when they turn the low, mid, and high-frequency EQ knobs on a mixer. Though this can be achieved through memorization, the practice of frequently frequency stacking makes this second nature. One of the primary jobs of a music engineer is to balance the frequencies of a track to make it sound good. It sounds easy- surely you can just boost and cut the frequencies that are too weak or too strong, right? Unfortunately, it’s rarely that easy. An engineer who doesn’t have a strong intuition for what sounds fall into what frequency ranges may not be the most successful engineer. Music producers have this same problem- but they have the creative ability to add or remove entire elements of the track. Suppose a producer thinks the mid-range frequencies are lacking in their project. An experienced producer will likely know exactly what sound to add to the track to fill that void- a skill that can be developed through the practice of frequency stacking. The benefits aren’t limited to technical ear-training skills, either. Remember how we can draw comparisons between songs and genres based on the instrumentation found in each frequency range? Swapping out elements in the same range from one genre to another is a great way to bring some fresh out-of-genre influence into a track to make something new. Reusing our house and jazz example from earlier, it wouldn’t be absurd to replace the synth lead on a house track with a jazzy saxophone- in fact, it might sound pretty good. This is because the synth lead and the jazzy saxophone play the same roles in their respective genres and fall into the same frequency ranges.

So go! Listen carefully and deliberately and analytically. Build frequency stacks. Explore new genres. Appreciate music. It’s one of the greatest things humanity has ever produced, and it’s one of the most accessible joys around the world. It would be a shame for it to go unheard.

Works Cited

[1] Britannica, “frequency,” Encyclopædia Britannica. 2019. Available: https://www.britannica.com/science/frequency-physics

[1] D. Purves et al., “The Audible Spectrum,” Nih.gov, 2001. https://www.ncbi.nlm.nih.gov/books/NBK10924/

[1] “Frequency Hearing Ranges in Dogs and Other Species,” Lsu.edu, 2017. https://www.lsu.edu/deafness/HearingRange.html