USC
About this Article
Written by: James Bunning
Written on: December 7th, 2016
Tags: lifestyle, computer engineering, music, sound engineering
Thumbnail by: London Guitar Academy/London Guitar Academy
About the Author
James Bunning was a student at the University of Southern California at the time this article was written.
Also in this Issue
To Float or Not to Float?Written by: Sabrina Winarko
Cryptography and Communication Security in a Digital AgeWritten by: Patrick Vanderwall
The Computation of Love: Finding Your Soul Mate OnlineWritten by: Melisa Osborne
The Future of SurfingWritten by: Christian Lee
Stay Connected

Volume XVIII Issue II > The Battle of Sound
From the time of vinyl records to modern portable music capabilities on phones, music producers have been fighting for the best (read: the loudest) music releases. Sound engineers have been the primary weapons in this war. Compressing the dynamic range, or the gap between the loudest and softest moments in a song, has armed the music industry with the ability to increase the volume of a song–perhaps at the expense of quality. Even with modern music streaming companies combatting loudness efforts, these “Loudness Wars” are still far from over.

Introduction

Have you ever been scanning the radio or listening to your iPod on shuffle when suddenly you need to crank down the volume because the next song sounds deafening compared to the previous one? Have your ears ever tired after rocking out to your favorite artist’s latest album? If so, you have been caught in the crossfires of the loudness wars, an auditory arms race that has been waging within the music industry for the past 60 years, all in the name of who can make their recordings sound the loudest.
Ever since the early days of audio recording, record companies and sound engineers have tried to create louder-sounding recordings to grab people’s attention and compete commercially with other songs. During the 1950s and 1960s, record labels started releasing louder 7-inch vinyl singles to stand out from records played on public jukeboxes and auditioned for in front of top-40 radio program directors [1]. However, this rather innocuous, sonic shouting match did not escalate into an all-out war until the advent of the compact disc (CD) in the 1980s. Digital recording technology gave music producers the unprecedented power to make songs sound louder than ever before, and as a result, record companies took full advantage of their shiny new toy. Looking at the audio signals of 4,500 best-selling or most well-received songs recorded and produced between 1969 and 2010 (Fig. 1), the average level of recorded music has consistently grown from 1982 to 2005 [2].
https://illumin.usc.​edu/https://illumin.​usc.edu
Figure 1: The average audio levels of 4,500 best-selling or well-received pop songs recorded and produced between 1969 and 2010. The red line tracks the average level for each year. Rectangles indicate distribution: the darker the rectangle, the higher density of songs recorded at that level.

How Do Music Producers Make Recordings Louder?

While experimenting with ways to make recordings stand out on 1950s jukeboxes and radios, sound engineers figured out that in order to make a song sound louder, they actually needed to start by turning the volume down, not up. In every genre of music from classical to jazz to rock and roll, songs have dynamics, so parts of the song are deliberately played at different volumes by the recording artists for effect. Although these loud and quiet sections give music a sense of excitement, movement, and emotional punch, they also presented a problem for early sound engineers. If a song had a large dynamic range, or a wide gap between the softest and loudest moments, a music producer could only push the recording level as high as the loudest transients, such as a drum hit or trumpet blast. Boosting the audio any higher than those peak levels would cause the signal to distort. Therefore, a singer belting out a note in one part of a song would severely limit how much a sound engineer could raise the level of the quieter parts. Sound engineers needed to devise a way to turn up the quieter sections of an audio signal while holding the pesky peaks at bay.
To solve this problem, sound engineers developed a studio trick known as dynamic range compression. While recording music in real time or mixing already recorded tracks, 1950s and 1960s sound engineers fed audio through a piece of electronic hardware called a compressor. By setting the compressor’s threshold control to a certain level, the audio engineer could make the compressor act like a brick ceiling, reducing the level of any mischievous audio peaks that tried to jump above it. After bringing down the loudest parts of the audio signal, the compressor then boosted the entire signal level back up to the original volume, transforming the quieter parts to sound closer in volume to the peaks. By reducing the dynamic range of a song through compression, sound engineers could now effectively increase the average level of music recordings.
James Bunning/Illumin
Figur​e 2: Dynamic range compression applied to "Rich Girl" by Hall & Oates (1976).
Software models of hardware compressors are used to create dynamic range compression (Fig. 2). Figure 2(a) depicts the uncompressed audio of the 1976 hit “Rich Girl” by Hall and Oates. The lighter blue portion in the middle of the waveform represents the audio signal’s average level. The higher the vertical height of the waveform, the louder the sound. Figure 2(b) shows what the audio signal looks like after passing through the compressor. Notice how the compressor greatly reduces the level of the waveform’s loudest parts. Lastly, the painted-on look of Figure 2(c) exhibits the last and most important step of compression: boosting the entire audio signal back to the original peak level. Comparing the compressed version of “Rich Girl” in 2(c) to the uncompressed version in 2(a), we see that the compressor has noticeably reduced the difference between the loudest and quietest parts of the song, increasing the track’s average level.

Limits of Loudness in the Vinyl Era

Although dynamic range compression initially allowed sound engineers to make louder sounding recordings, the physical characteristics of vinyl records put a limit on how loud music producers could push audio. A vinyl record stores audio information via small V-shaped grooves cut into its surface and plays the audio back when a needle runs over those grooves. In order to play louder audio recordings without the needle jumping during playback, sound engineers needed to cut deeper grooves [3]. However, with only a limited amount of useable surface area per disc, increasing the size of the grooves for louder songs meant sacrificing how many songs could fit on the side of a record. As long-play (LP) records became more popular in the 1960s and 1970s and needed to hold more songs per side, duration won over loudness, and the average loudness of songs stayed relatively the same until the early 1980s [3]. The surface of these vinyls were occupied by V-shaped grooves, offering a visual explanation of the pressure to choose between duration and loudness (Fig. 3).
https://illumin.usc.​edu/https://illumin.​usc.edu
Figure 3: V-shaped vinyl grooves.

Why Digital Audio Escalated the Loudness Wars

Similar to the way the machine gun in World War I completely changed the landscape of how wars were fought, the added firepower of the compact disc (CD) and digital audio processing in the 1980s escalated the loudness wars to new, ear-splitting levels. When first introduced, the CD recording format let music producers make recordings louder than vinyl recordings because digital audio overcame the physical limitations of vinyl records, such as needles jumping out of grooves [3]. Digital audio gave sound engineers the technical freedom to record music regardless of the peak loudness of the incoming audio signal [4].
While CDs did not place the same physical limitations on peak levels, the format did have a definitive loudness level over which the digital audio signal could not be boosted without clipping and losing information. This level is known as 0 dBFS, or “full scale.” If 0 dBFS were a card in the game Monopoly, it would likely read: “Do Not Pass Go, Do Not Collect $200.” Music producers began recording CDs far below this level, but the loudness wars picked up again towards the end of the 1980s as multi-disc CD changers started to gain popularity; record companies wanted their CDs to stand out from their competitors’ [3]. By the early 1990s, music producers were elevating the peaks of digitally recorded songs to the point where peak levels started to push the loudness limit of 0 dBFS. Once music producers reached this point of no return, the only way to raise the average level of the audio was to compress the peaks. Yet, unlike analog compressors, which were restricted by how much they could reduce the peak levels of tracks, digital compressors were much more powerful [3]. Sound engineers quickly began hyper-compressing audio in order to squeeze in as much loudness as possible, pushing the average level of songs up to as high as -6dBFS through the 1990s and 2000s [3].

Does Hyper-Compression Hurt Sound Quality?

In-and-of-itself, compression does not diminish audio quality. In fact, music producers often use compression as a creative tool in recorded music. Sound engineers use hyper-compression in a variety of creative ways other than simply increasing loudness, such as shaping the sounds of instruments and helping them stand out from others in a mix [4]. In hip hop, R&B, and dance music, music producers purposefully clip percussion by pushing it over the 0 dBFS point of no return as an aesthetic choice to get the driving “boom-bap” sound behind their beats [5].
However, audiophiles, casual music listeners, and even many music producers have argued that hyper-compression of the loudness wars have degraded the sound quality of modern recordings. Their chief complaint is that, by squashing the dynamic range of audio tracks through hyper-compressing, many modern music recordings sound flat, lifeless, and unexciting [6]. As prominent music producer Bob Katz put it, “You want music that breathes. If the music has stopped breathing, and it's a continuous wall of sound, that will be fatiguing. If you listen to it loudly as well, it will potentially damage your ears before the older music did because the older music had room to breathe” [3].
One famous backlash against the loudness wars occurred in 2008, when the band Metallica released two versions of its album Death Magnetic: one for the video game Guitar Hero and another as an audio CD. The CD version of the album was hyper-compressed and significantly louder than the Guitar Hero version, and as a result over 20,000 fans signed an online petition asking the band to re-master the CD [6]. Waveform analysis of "My Apocalypse" from the audio CD demonstrates the lack of dynamics, which were sucked out of the song as a result of audio compression (Fig. 4). In fact, the recording has an even smaller dynamic range than the song “Down Where the Big Bananas Grow,” primitively recorded in 1909 with an Edison cylinder.
https://illumin.usc.​edu/https://illumin.​usc.edu
Figure 4: Waveform A: "Down Where the Big Bananas Grow" (1909 Edison Cylinder); Waveform B: "My Apocalypse (Metallica, Death Magnetic, 2008).

Are the Loudness Wars Here to Stay?

Although music producers continue to hyper-compress tracks in order to increase the loudness of digital music recordings, the end of the loudness wars may be near due to the growing popularity of Internet music streaming services such as Spotify and Apple Music. Both Spotify and Apple Music use a loudness normalization algorithm that maintains a consistent loudness amongst all songs played back in shuffle or playlist modes [4]. By normalizing all tracks to the same level, streaming services take away the advantage of producing loud CDs or other digital audio downloads from music producers because all songs sound equally loud. In fact, when compared to less-compressed tracks, the hyper-compressed songs sound dull, weak and somewhat muffled, which may encourage sound engineers to started producing more dynamic music.

Conclusion

Even if Spotify’s and Apple Music’s loudness normalization trend catches on, the loudness wars may still be far from over. Hyper-compressed, louder-sounding music often suits the way modern listeners listen to music– on-the-go, via iPods and other portable music players. Louder sounding music blocks out the ambient noise of cars, public transportation, waiting rooms, and coworkers, which can interrupt quieter songs. Furthermore, after 20 years of listening to hyper-compressed tracks, many listeners, the youngest in particular, are simply accustomed to loud recordings and do not know the difference. Only time will tell if a peaceful end to the loudness wars is on the horizon.

References

    • [1] C. Clark. “The Loudness Wars: Why Music Sounds Worse.” NPR: All Things Considered, December 2009.
    • [2] E. Deruty. “Dynamic Range & The Loudness War.” Sound on Sound, September 2011.
    • [3] S. Sreedhar. “The Future of Music: Tearing Down the Wall of Noise.” IEEE Spectrum: Technology, Engineering, and Science News, May, 2007.
    • [4] H. Robjohns. “The End of The Loudness War?” Sound on Sound, February, 2014
    • [5] G. Milner. Perfecting sound forever: An aural history of recorded music. New York: Faber and Faber, 2009.
    • [6] E. Vickers. “The Loudness War: Background, Speculation and Recommendations.” AES 2010: Paper Sessions: Loudness and Dynamics. San Francisco: Audio Engineering Society, July 14, 2011.