The rationale and motivation for displaying information using sound (rather than a visual presentation, etc.) have been discussed extensively in the literature for a long time (e.g., Bly et al., 1985; Jeon, Walker, & Barrass, 2018, 2019; Kramer, 1994; Nees & Walker, 2009; Peres et al., 2008; Sanderson, 2006; Supper, 2014, 2015; Walker & Nees, 2011). Briefly, though, it has long been known that auditory displays exploit the superior ability of the human auditory system to recognize temporal changes and patterns (Bregman, 1990; Flowers, Buhman, & Turnage, 1997; Flowers & Hauer, 1995; Moore, 2013). In many instances, response times for auditory stimuli are faster than those for visual stimuli (Spence & Driver, 1997). As a result, auditory displays may be the most appropriate modality when the information being displayed has complex patterns, changes in time, includes warnings, or calls for immediate action.

Additionally, it has long been known that in practical work environments the operator is often unable to look at, or unable to see, a visual display. The visual system might be busy with another task (Fitch & Kramer, 1994; Wickens & Liu, 1988), or the perceiver might be visually impaired, either physically or as a result of environmental factors such as smoke in a burning building or line-of-sight obstructions (Fitch & Kramer, 1994; Wickens, Gordon, & Liu, 1998); or the visual system may be overtaxed with information (see Brewster, 1997; Brown, Newsome, & Glinert, 1989).

In some cases, auditory and voice modalities have been shown to be most compatible when systems require the processing or input of verbal-categorical information (Wickens & Liu, 1988). Other features of auditory perception that suggest sound as an effective data representation technique include our ability to monitor and process multiple auditory data sets (parallel listening) (Fitch & Kramer, 1994).

Finally, advances in technology for the past several decades have simultaneously expanded visual information displays toward opposite extremes in physical size. Portable devices (e.g., the latest «smart» wristwatches) continue the trend toward smaller physical dimensions, thereby leaving appreciably less space (or perhaps even no space) for a visual display (see an early recognition of this, Brewster, 2002). Fixed work stations, on the other hand, have become characterized by multiple visual displays with increasingly large physical sizes, due in part to increases not only in the affordability of displays but also in the expanded computing power to support multiple concurrent displays. This extends to modern immersive virtual reality (VR) contexts with massive pixel counts. As a result, visually intensive workstations and other multitasking situations may overburden the visual modality (see Grudin, 2001, for another early recognition of this problem). Thus, the inclusion of nonspeech audio in interfaces can promote universal design principles such as flexibility in use and perceptible information (see Connell et al., 1997; McGuire, Scott, & Shaw, 2006).

Although investigations of audio as an intentional ¹ information display modality date back over 75 years (see Frysinger, 2005; and see Worrall, 2018 for a «pre-history»), it was the advent of digital computing technology that really enabled sonification to gain the potential for ubiquity. Near the beginning of what might be considered the sonification era (about 1994 onwards), Edworthy (1998) even argued that the advent of auditory displays and audio interfaces was practically inevitable given the ease and cost efficiency with which electronic devices can now produce sound. A quarter century later, we may finally be approaching that level; however, much remains to be done to truly unlock the power and potential of sonification and other auditory display technologies.

The formation of the International Community for Auditory Display (ICAD ² ), and the first of its now-annual conference the International Conference on Auditory Display (sharing the ICAD acronym) in 1992, was a seminal point in the rise of sonification as a systematic scientific tool and a flexible expressive medium (see Kramer, 1994). Members of the nascent ICAD community (many of whom are still active in the field today, more than thirty years later) produced the collaborative Sonification Report (Kramer et al., 1999) as a starting point for a more structured discussion of the theory of sonification by identifying four issues that should be addressed in a theoretical description of sonification. These included: (1) taxonomic descriptions of sonification techniques based on psychological principles or display applications; (2) descriptions of the types of data and user tasks amenable to sonification; (3) a treatment of the mapping of data to acoustic signals; and (4) a discussion of the factors limiting the use of sonification.

Since then, research into the where, when, and how of sonification has blossomed, encompassing researchers from such diverse fields as audio engineering, audiology, computer science, informatics, linguistics, mathematics, music, psychology, and telecommunications, to name but a few. Sonification began to be implemented in fields as disparate as STEM education (Bonebright et al., 2001), astrophysics (Candey, Schertenleib, & Diaz Merced, 2006), and rowing (Schaffert et al., 2009). An array of conferences, venues, and publications have further showcased sonification in science, education, and entertainment. Sonification tools have appeared regularly (see later discussion) and design guidelines (see later discussion) have continually evolved, to support all kinds of users and sonification use cases.

Progressively, an understanding of the theory underlying sonification has evolved. This started with an amalgam of important insights and generalizations drawn from the convergence of these many diverse fields (e.g., Barrass, 1997; Brazil, 2010; de Campo, 2007; Frauenberger & Stockman, 2009; Hermann, 2008; Nees & Walker, 2007; Neuhoff & Heller, 2005; Walker, 2002, 2007), followed by more systematic statements of sonification principles) (see, e.g., Brazil & Fernstrom, 2009; de Campo, 2007; Frauenberger, Stockman, & Bourguet, 2007; Nees & Walker, 2007). The field marked milestones with encyclopedia entries (e.g., Walker & Kramer, 2006; Nees & Walker, 2009), and with the publication of The Sonification Handbook (Hermann, Hunt, & Neuhoff, 2011). Interest in, and need for, periodic summaries of the theory, methods, and practices for sonification has continued to the present (see, e.g., Jeon, Walker, & Barrass, 2018, 2019; Walker, 2021; and Walker & Nees, 2011, as just some examples). Nevertheless, there remains concern about whether there is such a thing as a theory of sonification. Nees (2019) recently summarizes the weaknesses of the field of sonification, as it relates to «theory», and discusses, instead, the potential for a «design theory» of sonification. Regardless of the state of theorizing, it is still possible, and useful, to describe the state of play in sonification, as a point of departure for discussing where the field may be (or may need to be) heading. This is congruent with the recent comprehensive overview by Andreopoulou and Goudarzi (2021) of 30 years of ICAD conference papers, showing a growth in the focus on «sonification», the popularity of the concept of «design», and an increase in interest for / use of more rigorous evaluation methods in relation to sonification.

Alerts and notifications are sounds that indicate that something has occurred, or is about to (see seminal work by Buxton, 1989; Sanders & McCormick, 1993; Sorkin, 1987). Alerts and notifications tend to be simple and particularly overt. The message conveyed is information-poor. For example, a beep is often used to indicate that the cooking time on a microwave oven has expired. There is generally little information as to the details of the event-the microwave beep merely indicates that the time has expired, not necessarily that the food is fully cooked.

Alarms and warnings are alert or notification sounds that convey the occurrence of a constrained class of events, usually adverse, that carry particular urgency in that they require immediate response or attention (Haas & Edworthy, 2006). However, the specificity of the information about the event is generally limited. Fire alarms identify an adverse event (a fire) that requires immediate action (evacuation), but the alarm does not indicate the location of the fire. More complex (and modern) kinds of alarms encode more information into the auditory signal, such as medical information (e.g., Anderson & Sanderson, 2004; Sanderson, Liu, & Jenkins, 2009; and see the development of the international standard for hospital alarms, Edworthy et al., 2018).

Object, item, and status indicators . Sounds such as earcons (e.g., Blattner et al., 1989; Bonebright & Nees, 2007; Brewster, Wright, & Edwards, 1993; McGookin & Brewster, 2004), auditory icons (e.g., Bonebright & Nees, 2007; Gaver, 1989; Keller & Stevens, 2004), and spearcons (Palladino & Walker, 2007; Walker, Nance, & Lindsay, 2006) provide information about the nature of the underlying action or event. These sounds are often used to facilitate user interface tasks (e.g., Brewster, Wright, & Edward, 1994; Winberg & Hellstrom, 2003). Earcons are abstract, artificial sounds that bear no ecological relationship to the represented process or event (e.g., beeps, chimes, abstract sound motives, etc., see Blattner et al., 1989). They can, however, be designed with a hierarchical structure or grammar, thereby enhancing their communicative power (e.g., McGookin & Brewster, 2011). Auditory icons are more natural sounds that have some real world relationship with their referent process or event. One simple example is the sound of a camera shutter being used in a (shutter-less) digital camera to indicate when a picture has been taken (see, e.g., Gaver, 1989). As an alternative to earcons and auditory icons, spearcons (and their «cousins», nearcons ⁴ ) use temporally compressed speech to represent objects, items, or processes with sound (Palladino & Walker, 2007; Walker et al., 2006). Spearcons have been shown to outperform both earcons and auditory icons (Walker et al., 2006) and may be especially useful in the design of flexible auditory menus (see Palladino & Walker, 2007) or for representing a large number of items.

Auditory menus are speech-based hierarchical lists (aka, menus) that present a set of options. Such menus can be simple (i.e., just presenting text-to-speech versions of the menu and sub-menu items), or may be more sophisticated constructions involving multiple voices, louder and softer speech, whispers, and additional elements representing scrollbars and spoken indexes. Menus may be navigated actively by the user («pull» menus) or more passively as the options are presented serially to the user («push» menus). For an overview of auditory menus see Yalla and Walker (2007), and Jeon et al. (2015).

Status and progress indicators convey the state of an ongoing process, such as downloading a file. In these instances, sound takes advantage of «the listener’s ability to detect small changes in auditory events or the user’s need to have their eyes free for other tasks» (Kramer et al., 1999, p. 3). Soundscapes have been designed to mimic natural sounds (e.g., a thunderstorm with rain), and parameters of the soundscape are mapped to variables in a multidimensional data set (e.g., Mauney & Walker, 2004). While the listener may not necessarily act upon every change in the soundscape, the display allows for on-going monitoring and awareness of a changing situation.

Art, entertainment, sports, and leisure-based auditory displays have long been provided for simple, traditional games like the Towers of Hanoi (Winberg & Hellstrom, 2001) and Tic-Tac-Toe (Targett & Fernstrom, 2003), and more complex game genres such as arcade games (e.g., space invaders, see McCrindle & Symons, 2000) and role-playing games (Liljedahl, Papworth, & Lindberg, 2007). There are many audio-only games, too, of course (see, e.g., the Survive the Wild audio game ⁵ ; and the Audio Games website ⁶ currently lists over 850 titles). Auditory displays also have been used to facilitate sports «watching» (e.g., Savery et al., 2019) and sports playing (e.g., soccer, Stockman et al., 2007; rowing, Schaffert et al., 2009; speed skating, Godbout & Boyd, 2010). Auditory displays have also been used as a means to bring some of the experience and excitement of dynamic exhibits to the visually impaired (e.g., sonified soundscapes to convey dynamic movement of fish in an «accessible aquarium», Walker et al., 2006; Walker, Kim, & Pendse, 2007; and more recently to sonified planetarium exhibits, e.g., Quinton, McGregor, & Benyon, 2016; Tomlinson et al., 2017).

Wayfinding and navigation can also be supported by auditory displays and sonification, often leveraging virtual spatial audio. Historical examples include the System for Wearable Audio Navigation (Wilson et al., 2007; updated as SWAN2.0, see Walker & Wilson, 2021), the Personal Guidance System (PGS, Golledge et al., 1991; Loomis, Golledge & Klatzky, 1993; Loomis et al., 2005), and computer-vision and navigation systems by Revuelta Sanz and colleagues (Revuelta Sanz et al., 2014a, 2014b, 2014c). More recently, Microsoft has developed Soundscape ⁷ and XRNavigation has developed AUDIOM ⁸ ; both are available for audio-based navigation and wayfinding.

Data exploration interfaces are what is generally meant by the term «sonification», and are usually intended to encode and convey information about an entire data set or relevant aspects of the data set. Sonifications designed for data exploration differ from status or process indicators in that they use sound to offer a more holistic portrait of the data in the system rather than condensing information to capture a momentary state such as with alerts and process indicators, though some auditory displays, like soundscapes, blend status indicator and data exploration functions. Auditory graphs are a common approach to basic data exploration sonifications, and most commonly use changes in auditory frequency to correspond to changes in data values along the visual Y axis, while time corresponds to the visual X axis. Nees and Walker (2007) proposed a conceptual psychological model of auditory graph comprehension. There have been auditory versions of numerous traditional display formats, including auditory scatterplots (e.g., Bonebright et al., 2001; Flowers, Buhman, & Turnage, 1997), box-whisker plots (Flowers & Hauer, 1992; Peres & Lane, 2003, 2005), histograms (Flowers & Hauer, 1993), multidimensional data sets (see Hermann & Hunt, 2005), and tabular data (Stockman, Hind, & Frauenberger, 2005).

As a bit of an aside, some organizations, such as NASA, are producing sonifications of data using many of the methods that a scientist might use to explore their data; and then releasing the sonifications to the public as a form of outreach. Recent such outreach examples come from the Chandra X-ray Observatory ⁹ and the James Webb Space Telescope ¹⁰ . The point is that the sonifications are not, in those cases, really intended to be used for scientific discovery. Rather, many listeners simply enjoy the novel sounds as a sort of “astronomical artwork” and (hopefully) become excited about the activities of the scientists. This highlights that while one might list out a taxonomy of sonification types, the categories are really much less distinct, the boundaries less clear…and it probably matters relatively little what, exactly, a sound is called, compared to the ultimate utility it has in conveying information or achieving some other purpose.

Another way to organize and define sonifications is to describe them according to the sonification technique or approach. De Campo (2007) offered a sonification design map that featured three broad categorizations of sonification approaches: (1) event-based; (2) model-based; and (3) continuous.

Parameter mapping sonification represents changes in some data dimension with changes in an acoustic dimension to produce a sonification. Auditory graphs and many sonifications fall into this category. Sound has a multitude of changeable dimensions (see Kramer, 1994; Levitin, 1999) that allow for a large design space when mapping data to audio. These approaches to sonification have typically employed a somewhat passive mode of interaction, in that the sonification is «played» and the listener attempts to understand what is happening in the data set.

Model-based sonification (e.g., Hermann, 2002; Hermann & Ritter, 1999) involve a virtual model whose sonic responses to user input are derived from data. A model, then, is a virtual object or instrument with which the user can interact, and the user’s input drives the sonification such that «the sonification is the reaction of the data-driven model to the actions of the user» (Hermann, 2002, p. 40). The user comes to understand the structure of the data based on the acoustic responses of the model during interactive probing of the virtual object. These types of sonifications tend to involve high data dimensionality and large numbers of data points.

Audification is the (nearly) direct conversion of data into sound: waveforms of periodic data are translated into sound (Kramer, 1994). For example, seismic data have been audified in order to facilitate the categorization of seismic events with accuracies of over 90% (see Dombois, 2002; Speeth, 1961). This approach may require that the waveforms be frequency- or time-shifted into the range of audible waveforms for humans.

An auditory display is useless if the listener cannot hear the sounds in the system’s environment of operation. To ensure detection, a consideration of the acoustic spectra of both the sonification (the «signal») and the environmental sounds (the «noise») is critical. Considerations of detection thresholds (e.g., Hartmann, 1997) and masking theories may help (for a discussion, see Watson & Kidd, 1994). And ecologically valid evaluation is important (Brewster, 2002; also see Walker & Kramer, 2004). A second consideration is the discriminability of sounds with distinct meanings, with a long-standing literature of perception research available for guidance on the psychology of hearing (e.g., Moore, 2013), pitch (e.g., Stevens, Volkmann, & Newman, 1937; Turnbull, 1944), loudness (e.g., Stevens, 1936), tempo (e.g., Boltz, 1998), and duration (e.g., Jeon & Fricke, 1997), to name but a few.

Sounds that annoy the user may be ignored or turned off, even when the sounds are beneficial. Aesthetic considerations intersect with performance concerns. Some recommend musical sounds (Brown et al., 2003; Childs, 2005; Ramloll et al., 2001), though that, in itself, will not guarantee a pleasant experience for all users, tasks, and environments. Clearly, developing an auditory interface is, in all regards, a design task, with all the inherent difficulties associated with design (and, as noted above, see Nees, 2019).

Data-to-display mapping refers to the attribute of sound that is used to represent changes in data. Walker has studied the appropriate acoustic dimension for a given type of data by examining mappings between numerous conceptual data dimensions (e.g., temperature, pressure, danger) and three acoustic dimensions (pitch, tempo, and spectral brightness; Walker, 2002, 2007). This is complicated by the fact that many acoustic dimensions (e.g., pitch and loudness) interact with one another (see, e.g., Moore, 2013). Nees and Walker (2007) discuss the convention of mapping data values onto changes in pitch in auditory graphs. Sonification designers should note that not all acoustic mappings are equally effective, and best designs will arise from an awareness of both the historical literature and pilot testing of any displays.

The polarity of the data-to-display relationship refers to whether increases in a given acoustic dimension (e.g., pitch, tempo, etc.) represent increases in the data represented (a positive mapping polarity, Walker, 2002, 2007), or decreases in the data (a negative polarity). Listeners might agree that increasing pitch suggests increasing temperature, yet the same group of listeners may feel that increasing pitch offers a more intuitive representation of decreasing size. Walker and Lane (2001; see, also, Mauney & Walker, 2010) showed early on that some polarity mappings were reversed for visually impaired as compared to sighted listeners.

The scaling refers to the amount of change in an acoustic dimension that will be used to represent a unit of change in the data. Magnitude estimation has been employed to describe the intuitive slopes for scaling frequency to a number of conceptual data dimensions (Walker, 2002, 2007), and the conceptual data dimension being represented impacts the choice of scaling factor in the display. A match between the listener’s preferred or intuitive internal scaling function and the display’s scaling function may improve comprehension, though usability testing will help determine the best scaling for a given situation.

Some data analysis tasks require the comparison of values from different data streams, whereas in other cases it is preferable to fuse streams into a perceptual whole. Bregman (1990) discusses what acoustic properties support or inhibit stream segregation, with key dimensions being timbre, spatial location (or stereo panning), pitch/frequency, and onset/offset of sounds.

Context refers to the purposeful addition of non-signal information to a display (Smith & Walker, 2005; Walker & Nees, 2005). Sonifications need to include contextual cues equivalent to axes, tick marks and labels, so the listener can perform the interpretation tasks. For example, adding a series of clicks to the display can help the listener keep track of the time better, which aids in their interpretation of the graph values (see, e.g., Smith & Walker, 2005).

The perceptual and cognitive capabilities, limitations, and experiences of listeners, as well as transient states (like mood and level of fatigue) will all impact performance outcomes with auditory displays. By understanding ranges in individual difference variables, a designer can build a display that accommodates most users in a given context (e.g., universal design, see Iwarsson & Stahl, 2003). It is interesting to note that for many years researchers predicted and anticipated that musicians would outperform non-musicians on tasks involving sonifications. However, research has very rarely found any correlations between musical experience and performance (e.g., Lacherez, Seah, & Sanderson, 2007; Neuhoff & Wayand, 2002; Sandor & Lane, 2003). One explanation for the lack of relationship is the crude nature of oft-used self-report metrics of musical experience. Indeed, in a systematic investigation Schuett (2019) determined that a more sophisticated measure of musical sophistication, leaning largely on engagement with music, can be predictive of performance. Visual impairment also has been shown to have a potentially profound impact on the perception of sonifications. As mentioned, it has been shown (Mauney & Walker, 2010; Walker & Lane, 2001) that blind and sighted listeners can have opposing intuitions about the polarity of the pairing of some acoustic dimensions with conceptual data dimensions. Individual differences between visually-impaired and sighted listeners require more research and a careful testing of auditory displays with the intended user population.

As Nees (2019) discusses, there have been a (long) series of largely one-off software tools to create sonifications (too many to list exhaustively, here). They have varied greatly in terms of the platform and programming language, the approach to creating sounds, the process for defining mappings, context, and other attributes. Examples range from sonification toolkits focusing on a specific domain (e.g., xSonify: Candey, Schertenleib, & Díaz Merced, 2006) to more general frameworks (e.g., SoniPy: Worrall et al., 2007). The Sonification Sandbox ¹¹ was, for many years, a toolkit that was intended to serve the needs of diverse STEM fields (Davison & Walker, 2007), though it still had the limitation of being optimized for basic auditory graphs, and not a broader range of sonification methods. The recently released Highcharts Sonification Studio (HSS) ¹² is a re-implementation of the Sonification Sandbox as a web application, supported by the power of the Highcharts visualization engine that incorporates extensive sonification capabilities. Backed formally by a mainstream data visualization company (HighSoft), the HSS seems to represent the first corporate mainstreaming of sonification tools (Cantrell, Walker, & Moseng., 2021), and is notably built to be accessible to screen reader users.

Historically, the «last mile», or the actual output of sound was often a challenge. Systems might not be able to produce sound, or if so, might need speakers or headphones as an additional piece of equipment. A sonification designer could never know what the actual listening equipment would be. Now, however, nearly all modern digital devices, from phones to tablets to laptops to smartwatches to smart speakers, are capable of producing high-fidelity sound, with most now including speakers, even if small. An output jack (e.g., audio only or HDMI) or Bluetooth capability is largely standard. As such, it is generally straightforward to play a sonification.

The Technology Acceptance Model (TAM; Davis, 1989) posits that the adoption of a technology depends largely on two main factors: perceived usefulness or perceived utility (PU) and perceived ease-of-use (PEU). In practice, these two factors are interconnected, of course. Perceived usefulness (PU) is defined by Davis as how much a person believes that using a particular technology or technical system would enhance performance on their task. One can extend Davis’ thinking about technology acceptance to sonification by considering some of the specific contributors to usefulness, such as theoretical grounding for the use of the technology, scientific utility, scientific validity, replicability of results when using the technology, standardization of the use of a technology, educational utility, and accessibility or inclusivity for accomplishing a task.

Perceived ease-of-use (PEOU) is defined by Davis as how much a person believes that a technology would be user-friendly. Again, one can extend and update Davis’ thinking by considering factors such as the availability and prevalence of the (software) tools required to use a technology, the usability of the tools, the training that is available, the standardization of the technological solution, the existence of a community of users (and support), portability of the technology, integration of the technology into the work/school/science ecosystem, and the accessibility of the system.

With the TAM framework, and an updated consideration of sonification, it is interesting to assess the current state of sonification, and provide a bit of a “report card”.

Perceived Utility «Report Card» for Sonification. In terms of theoretical grounding, there is a solid body of published research examining the core components of sonification, even though there may still not be an actual “theory of sonification” (and see Nees, 2019, for a discussion of design theory for sonification). There is, for example, a Sonification Handbook (Hermann, Hunt, & Neuhoff, 2011), including a chapter on the Theory of Sonification (Walker & Nees, 2011), providing a grounding for applications of sonification. There have been many findings made possible through sonification, ranging from early space science (e.g., NASA Cassini Mission ¹³ ) to recent cancer diagnosis (Dascalu et al., 2021), demonstrating scientific utility of sonification. The scientific validity has been less-thoroughly investigated, to date, with small sample sizes being typical and many generally unreplicated results. In terms of standardization, the basic concept of an auditory graph (x-y plot mapped onto time and pitch) has become a de facto standard, though there are many design differences amongst the countless implementations. Beyond simple auditory graphs, however, there is little standardization in sonification (in terms of the tools used, the designs, the deployment, and so on). The educational utility has started to emerge, especially as sonification is beginning to be used as part of the assistive technology used in Science, Technology, Engineering, and Mathematics (STEM) education. More needs to be done in this regard, since accessibility could be a huge area of impact for sonification.

Perceived Usability «Report Card» for Sonification. The sonification tools are quite readily available, with many options at various levels of sophistication, and using various underlying technology «stacks». It can be a challenge to know where to look, and there may be technical expertise required (e.g., programming in a particular programming language) in order to actually use some of the tools. The usability of the software tools varies greatly, from walk-up-and-use to experts-only, especially since many tools were not designed or developed for widespread deployment-they are often just built to assist a particular researcher to investigate a particular type of data. As previously discussed, more tools for sonification are becoming available with ease of use in mind (e.g., Highcharts Sonification Studio). Training has often not been very available for sonification tools, beyond the «readme» files that come with software packages, plus the limited details that can be gleaned from academic papers or technical reports. Recently there is an online course on the Coursera platform about the design of sonification (Moore, Tomlinson, & Walker, n.d.), and countless emerging YouTube videos and channels ¹⁴ and websites to support novices in getting started with sonification. Standardization is minimal, as is portability, which makes it a challenge to use sonification. The emergence of file formats that can be exchanged across software applications (and shared from person to person) will help with portability, but this remains a fledgling concept. Thankfully, the community of users and developers in the field of sonification has been around for a few decades, but it remains relatively small and unfortunately a bit on the margins of many other fields (education, STEM, computer science). There is little or no financial support for sustainability of the community. Since sonification is often used by researchers in another field (e.g., astronomy), the sonification tools are often integrated into the ecosystem of data collection and analysis, though this also remains a work in progress, across the field. In education, however, sonification tools, when available, are still largely separate from the ecosystem of other educational technology and assistive technology. This is changing as sonification tools are now being built to play nice with file formats and data transfer protocols that are common. Finally, the heterogeneity of sonification tools also means that there is a range of accessibility and compliance. It is encouraging that many of the more recent sonification tools are «born accessible», often due to the involvement on the development team of a designer or developer or scientist with a disability.

Summary «Grades» for Sonification Acceptance. Overall, it seems fair to conclude that the field of sonification is doing well on perceived utility, though with room for improvement. In terms of the perceived usability, there is more work to be done. Sonification is already seeing adoption in science, perhaps because the «proof» of utility and usefulness has been delivered; and the technical sophistication of the typical users is higher, leading to a greater tolerance of usability and ease-of-use challenges. In STEM education, the promise of sonification is likely understood (especially in the accessibility domain), but the real and perceived challenges in usability hamper further adoption. The field of sonification researchers and developers need to work closely with the end users to make the case for utility, and build usability and accessibility into any and all new tools and methods for sonification.