What is silence? Therefore, what is sound?

To an acoustician, “what is sound?” might appear to be a trivial question, one that might be asked by a student taking his or her first course in acoustics or hearing. However, it is a question that has fascinated scholars since almost the beginning of recorded history [e.g., see Boring (1942)]. This opinion essay was motivated by a recent article¹ (Goh , 2023) titled, “The perception of silence.” In this article, the authors did or did not present certain well-described sounds to their subjects. They referred to the times when a sound was not presented as “silence,” but they also inferred that “silence” means the absence of sound. Thus, defining “silence” in these cases appears to require a definition of sound. As a result, this essay expresses my opinion¹ about what sound is and is not, and how the consideration of “what is sound?” influences considerations of “what is silence?.”

Without getting too technical, most often sound is said to occur when a vibrating object causes molecules and their atoms in some medium (often air) to move [e.g., see Rossing (2014)]. This motion can cause a wave of pressure changes in the medium, and this wave is often referred to as a sound wave. Even if no object initiates a sound wave, pressure changes in a medium still exist due to the random movement of atoms and molecules in any medium, and this random fluctuation is sometimes referred to as thermal noise [e.g., see Sivian and White (1933)] or Brownian motion [e.g., see de Vries (1952)]. Only if molecules could not move (e.g., at 0 K) or there were no molecules to move (e.g., in a vacuum) could there be no molecular motion. Thus, if silence were the absence of molecular motion, silence would occur only under conditions like at a temperature of 0 K or in a vacuum.

The simplified operational physical description of sound provided above is, in my opinion, inadequate to explain one's perception of sound. It is certainly the case that a sound wave of changing pressure caused by a vibrating object can lead to the perception of sound, but not always. Pressure changes in a medium (e.g., air) are not always perceived as sound. For instance, slow changes in air pressure that indicate a change from a low-pressure to a high-pressure weather system does not lead to the perception of sound, but rapidly changing air pressure (wind) can. Consequently, weather-related air-pressure changes are not considered sound.

Sound can be perceived when there is no pressure wave. For instance, consider those who suffer from tinnitus [“ringing in the ears,” e.g., see Baguley (2013)], the perception of sound when there is no change in air pressure. Even when the auditory nerve is cut, so that there is no auditory peripheral input to the brain about changes in pressure, tinnitus (perceptible sound) occurs. The majority of cases of tinnitus are due to an, as of now, unknown neural abnormality in the brain that triggers the perception of sound. Another example of the perception of sound when there are no pressure changes is the ability of many listeners with severe-to-profound hearing loss who use cochlear implants (CIs) to perceive sound, sometimes in a similar, if not identical, manner to those with normal hearing. Sound is perceived by CI users when an electric current has the opportunity to stimulate auditory-nerve fibers via implanted electrodes, even when there is no pressure wave. In these cases, the outer, middle, and inner ear are bypassed. As another example, the auditory nerve is also not required in that in some cases electrical stimulation of the brainstem, well above the anatomical level of the auditory nerve, can elicit the perception of sound. That is, neither a pressure wave nor an auditory periphery (an “ear”) are required to perceive sound, but an auditory brain is.

My point is that sound is, in my opinion, not a physical property, but is a perceptual property. In most cases changes in pressure (often treated as a pressure wave) is the physical basis of sound, but not all changes in pressure are perceived as sound, and sound can be perceived when there are no changes in pressure. The auditory periphery in most animals primarily evolved to detect changes in air pressure over a certain range of air-pressure changes and provide a neural code to the brain regarding those changes. If the neural code contains sufficient information, the brain interprets that code as indicating that a “sound” occurred (i.e., “sound” is perceived). How and where that happens in the brain are largely unknown. Moreover, as argued above, there does not have to be a neural code provided by the inner ear for “sound” to be perceived, but there does have to be a brain.

As I have already argued, stating that silence is the absence of sound in the sense that silence means there are no pressure changes is probably not consistent with the common use of the word. What might be a good way to describe silence? Perhaps silence is when pressure changes are not audible. Let us consider the differentiation between audible sound and inaudible sound (e.g., infrasound or ultrasound). Considering inaudible sound as “silence” does not mean that pressure changes did not occur. Inaudible sound most likely means that the changing pressure did not lead to stimulation of the brain circuits used to perceive sound, not that there were no pressure changes. In some circumstances, changing pressures that are inaudible can still be perceived by humans (and probably by other animals). For instance, when objects produce infrasound, the pressure changes can sometimes be “felt” but not “heard.” Defining the difference between “felt” and “heard” is fraught with challenges, especially because pressure changes that are “felt” but not “heard” are often not considered sound. In addition, different animals are sensitive to different frequency ranges of pressure changes. A 50 000 Hz change in air pressure would not be perceived by humans, and would be considered ultrasound. However, many bats or dolphins could not survive if they could not perceive pressure changes at such high frequencies (e.g., they use “ultrasounds” for echolocation). Does that mean that an infrasound that is not “heard” but “felt” is “silent”? Is the existence of ultrasound “silence” for a human but not for a bat?

In peeling back a layer of my opinion, I get to the issue of the perception of sound as opposed to the perception of the sources of sounds. Most of the time when one perceives a sound, one perceives the vibrating source that generated the pressure changes. However, as I have already pointed out, it is not necessary to be able to identify a sound source for perceivable sound to exist. In most cases one's memory allows one to identify sound sources (e.g., a guitar made that sound), but even if one has no memory of what the source of a particular sound may be, one can still indicate that a sound exists, describe perceptual attributes of the sound (maybe its pitch, relative loudness, perceived duration, perceived source location, etc.), and perhaps indicate what object produced the sound.

When one cannot perceive that sound exists, that does not mean that there are no pressure changes. All one has to do is use a sound-level meter to measure a sound level that will always exist. A sound-proof room that meets the American National Standards Institute's (ANSI) criteria for sound-proof rooms [ANSI/ASA S3.1 (ANSI, 2018)] would most likely have an ambient (i.e., there are no known sources producing air-pressure changes) sound level in the vicinity of 25 dBA (an audible level), while in a typical “quiet” room that is not designed to be sound proof, the ambient sound level would likely be in the vicinity of 40 dBA. This value of 40 dBA, as representative of a quiet room, was a motivation for the use of the 40-phon equal loudness contour as the basis of the “A” weighting used for sound-level measurements [see Yost (2020)] originally developed in the 1940s by the American Standards Association and adopted later by American National Standards Institute [e.g., ANSI/ASA S1.4 (ANSI, 2014)]. The sound level in a room when no obvious sound source exists is likely to change over time by several decibels, in some cases more than enough to be perceptible. If sounds are presented via headphones [ones that cover the pinna (e.g., circumaural) or inserts (e.g., ear buds)], they only partially block pressure changes that occur around a listener, and there are pressure changes in the ear canal caused by headphone-diaphragm motion. Thus, pressure changes that may or may not be directly perceived are always present, even when sounds are presented over headphones.

Another layer of my opinion involves the sensitivity of the auditory system to air-pressure changes that result in the perception of sound. Sivian and White (1933) were among the first to measure the sensitivity of the human auditory system to tones of different frequencies and to collate how their measures compared to other such measurements that were made by others at about the same time (Yost, 2021). Sivian and White (1933) showed that the lowest sound pressure that human listeners could detect was approximately 0.0002 dynes/cm² (in today's International System of Units, 20 μPa). This value of 20 μPa was about the lowest sound pressure level human listeners could detect when sound was presented in a calibrated sound field [minimal audible field (MAF) thresholds] or over calibrated headphones [minimal audible pressure (MAP) thresholds]. MAF and MAP thresholds differed by a few decibels, which Sivian and White (1933) attributed to the acoustic properties of the ear canal, a fact that has been verified several times [e.g., see Yost and Killion (1998) for a detailed review]. Thus, pressure changes that occur in a sound field or via headphones are essentially equally detectable when presented at about the same sound-pressure level.

Sivian and White (1933) were also curious about the theoretical limit of aural acuity, i.e., is the ear's sensitivity “limited by its physiological construction, or whether the limit is imposed by the air as a transmitting medium.” Their conclusion was, “For exceptionally good ears, a further increase in physiological sensitivity would be useless in the presence of thermal noise.” Green and Dai (1991) are perhaps the most recent investigators to reach a similar conclusion.² Although there is great diversity in the sound frequency at which different animals are most sensitive, Fay (1988) showed that the sound-pressure level of the most sensitive spectral region across all animals, despite variation in the type of auditory receptors, whose thresholds have been measured is remarkably close to 20 μPa.³ That is, aural acuity throughout the animal kingdom is about as sensitive as it can be, meaning that the lower limit to the auditory system's ability to detect pressure changes generated by a vibrating sound source is that provided by thermal noise (Brownian motion). As such, there may not be, under ideal conditions, any perceived silence as long as thermal noise (Brownian motion) exists. At the very least, the absence of pressure changes is probably not a useful way to determine the perception of silence.

Another aspect of my opinion about sound and silence deals with conditions in which one does not perceive that sound exists (even though pressure changes that are audible do exist), but the sound that does exist alters one's auditory perception. Consider cases in which one is in a room where there are no obvious sound sources that can be identified, and a sound-level meter could indicate a sound level of 25 dBA or more at the position of the listener's ears. Clearly, the room is not “silent,” if silence is the absence of sound. However, one might consider the room to be “silent” because there are no known sound sources. One obvious example of such “silence” leading to perceptual attributes involves the reverberant properties of rooms. If one is in an anechoic (“echo-free”) space or even an “echo-reduced” space (a sound-deadened room) when no sources are obviously producing sound, being in such a room “sounds” different than when one is in a room that has reflections/echoes. The “silence” of the echo-free room is perceptually different from the “silence” in a room with reflections. This is not a subtle effect.⁴ Even patients who use CIs remark that when they are in such an echo-reduced room, the room “sounds different” even if there are no obvious sources producing sound.⁴ This, of course does not mean “silence (defined as the absence of sound)” is perceptible, as “silence” does not reflect off surfaces, like walls, but sound does.

If one is trying to detect over headphones a low-frequency tone (e.g., 200 Hz) masked by a noise, other pressure changes (ones that listeners probably do not directly perceive) in the ear canal will influence the ability to detect the noise-masked tone [e.g., see Yost (1988)]. As the level of the externally produced masking noise decreases, the ability to detect the tone depends more and more on the level of the internal pressure changes (noise) in the ear canals. This can be readily determined if the masking noise is presented to both ears, as the internal noise (presumably that in each ear canal in the absence of any external sound source) is interaurally uncorrelated, and the detection of low-frequency tones is dependent on the interaural correlation of masking noise presented to both ears [see Yost (1988) for a review]. That is, even when there does not appear to be a known sound source, pressure change that might be considered “silence” but is not, can influence auditory perception. If the air pressure changes in the ear canal are considered “silence” because they do not seem associated with a known sound source, then these actual canal air pressure changes can influence the perception of a sound from a known source.

This leads to the final layer of my opinion. In the reverberant cases referenced above and when listening over headphones, what are the air pressure changes that are reflected or influencing the perception of other sounds? When a person is in the room, his or her heart is beating, lungs are expanding and contracting, other parts of the bodies vibrate, the room's lighting structures and air handling systems vibrate, the building sways, etc. These are just a few of the potential vibrating sources producing air pressure changes that reflect off surfaces, but they are usually not obvious sound sources to a person in the room. If the sound-pressure level in a room is measured when no one is in it, and then again with someone in it, the sound level will be greater when the person is in the room, even if the person produces no obvious vibrations. Similarly, the air-pressure changes in the ear canals most likely result from the vibration of the listener's internal organs such as the heart and lungs [e.g., see Yost (1988)].

In conclusion, in my opinion, sound requires a perceptual definition. That is, sound is a perceptual phenomenon that usually occurs when objects vibrate causing changes (a wave) of pressure in a medium, but not always. If “silence” can be perceived, it is not the perception of the absence of sound, but perhaps “silence” is what can influence one's perception when there are no obvious sound sources.

I greatly appreciate discussions with Dr. Judy Dubno and Dr. M. Torben Pastore about my opinions. However, please do not assume that Judy or Torben always agreed with me.

The author has no conflicts of interest to declare.

The data that support the findings of this study are available from the author upon reasonable request.