Exploring the Opportunities of Haptic Technology in the Practice of Visually Impaired and Blind Sound Creatives

: Visually Impaired and Blind (VIB) people as a community face several access barriers when


13
This article explores the access barriers and wider issues salient to the workflows of 14 visually impaired and blind (VIB) sound creatives, which are typically centred around the 15 digital audio workstation (DAW). 16 We use the term 'VIB Sound Creatives' to denote anyone with a visual impairment 17 who works creatively with sound, whether on a professional or hobbyist level. VIB people 18 exist on a broad spectrum of sight loss which includes not only complete blindness but 19 also varying degrees of light perception, peripheral vision, and the refractive issues of 20 myopia and hyperopia. The term sound creative attempts to capture the plethora of diverse 21 and distinct sound-oriented practices that people engage with creatively, such as music 22 production, audio engineering, composing and DJing. 23 VIB sound creatives comprise a community of practitioners for whom existing music-24 technology software, particularly DAWs, present several accessibility barriers. Accessibility 25 tools commonly used by VIB people include screen-reading software, keyboard macros, 26 and braille displays. While these tools effectively provide access to digital materials such 27 as text-based content, significant gaps remain between a sighted person's experience of 28 specialist software, such as a DAW, and that of a VIB person. Our research has highlighted 29 how these gaps arise, primarily affecting how VIB users perceive the content of their DAW 30 and the efficiency with which they can manipulate software features. Through analysis 31 of interviews with twenty VIB sound creatives, we consider whether haptic technology 32 represents an untapped potential for improving VIB sound creatives' access to DAWs. We 33 consider how haptics might integrate with a VIB sound creative's existing workflow, and 34 indeed whether this reflects the priorities of the broader community. 35 The United Nations deem accessibility in all aspects of life to be a human right 1 . The 36 International Standard Organisation (ISO) describes accessibility as the "extent to which 37 In this article, we define accessible music technology (AMT) as any software or hard-48 ware elements which VIB sound creatives incorporate into their workflow. The specificity 49 of this definition is crucial as the practice of VIB sound creatives is often underpinned by a 50 complex artefact ecology (Bødker and Klokmose 2011), including mainstream software and 51 hardware (sometimes adapted), open-source software extensions and scripts. It is essential 52 to capture this reality, regardless of whether the technology was initially designed with 53 accessibility in mind. 54

55
Most VIB sound creatives use a text-to-speech (TTS) engine called a screen reader to 56 access their computers (Payne et al. 2020). The screen reader allows someone to navigate 57 and control a computer system using a combination of keyboard shortcuts and navigation 58 commands and hear the visible elements of the graphical user interface (GUI) spoken aloud. 59 This way, someone can use a computer without needing to interact with a mouse or look 60 at the screen. For MacOS users, the only available screen reader is Apple's VoiceOver 2 . 61 VoiceOver is fully integrated with the MacOS architecture, comes free with every computer, 62 and cannot be replaced with a third-party product. On Windows, there are two primary 63 screen readers commonly in use; JAWS 3 (Job Access With Speech), a commercial product 64 designed by the company Freedom Scientific, and NVDA 4 (Non-Visual Desktop Access), a 65 free and open-source equivalent. Some VIB sound creatives make use of additional software in their workflow. Optical 76 character recognition (OCR) is a method used to scan a GUI, identify any text elements, 77 and make them available to the screen reader for description and navigation. OCR can be 78 beneficial in situations where the GUI has not already been made accessible to a screen 79 reader by the software manufacturer. NVDA offers a free OCR addon 5 . VOCR 6 is an 80 open-source VoiceOver OCR extension popular with the Mac users within our participant 81 group. 82 Some VIB sound creatives use additional scripts with their screen reader and DAW. For 83 example, the script Flo Tools 7 is popular among VIB sound creatives using the DAW Pro 84 Tools 8 on MacOS. Pro Tools already boasts a significant level of screen-reader functionality 9 . 85 However, Flo Tools expands the capabilities of Pro Tools, enabling VIB sound creatives 86 to work more efficiently. Some VIB sound creatives consider Flo Tools a usability rather 87 than an accessibility extension as all of Pro Tools' features are accessible without Flo Tools; 88 however, the latter improves the workflow. The Komplete Kontrol 10 range of keyboards produced by Native Instruments supports 91 the accessible manipulation of plugin parameters via TTS feedback, instigated by the oper-92 ation of hardware controls but delivered via the computer. To this end, Native Instruments 93 developed the Native Kontrol Standard (NKS) 11 protocol. Third-party software companies 94 following this protocol can benefit from the same degree of accessibility via the Komplete 95 Kontrol hardware. 96 Some VIB sound creatives also use braille displays in their accessible workflows. 97 Braille displays are physical devices that manipulate small cells on a tactile interface to 98 simulate braille embossed on paper. These displays can interface with the computer and 99 change cell state to represent the text currently focused on the screen. Most displays also 100 allow someone to enter text in place of a typical QWERTY keyboard. 101 2.6. Industry 102 Several prominent software manufacturers have recently announced commitments to 103 make their technology VIB accessible. A number, such as Native Instruments mentioned 104 above, have released technology featuring components designed specifically with visual 105 accessibility in mind. The industry is taking steps in the right direction, in part thanks to 106 the advocacy of VIB sound creatives, but we would also like to note that these commercial 107 products require a monetary outlay. A thriving community of open-source community 108 developers offer users alternative tools at zero cost. For example, the OSARA 12 accessi-109 bility extension commonly used with the DAW Reaper 13 is open source, free to use, and 110 developed in close communication with the developers of Reaper itself. For completeness, it is vital to acknowledge online communities' role in providing and 113 supporting accessible DAW workflows. Examples of such communities include Reapers 114 without Peepers 14 , the Pro Tools Accessibility Google group 15 , the Logic Accessibility 115 Google group 16 and the MIDI Mag mailing list 17 . Amongst other things, these communities 116 are a place for VIB sound creatives to seek technical advice regarding access solutions and 117 music production techniques. Members of these communities benefit from one another's 118 experience, for example, by checking the degree of accessibility of certain music-technology 119 products before committing to a purchase.

121
HCI researchers in related fields have proposed haptic technologies to address access 122 barriers that VIB people encounter. Commonly, haptic technologies are considered a means 123 of presenting additional sensory information to a technology user, often to replace the 124 visual modality.

125
O'Modhrain et al. (2015) provide an overview of existing technologies with the po-126 tential to represent visual media through refreshable tactile displays and the kinds of data 127 and media which could be displayed with them. Three common tasks proposed for haptic 128 technologies and accessibility are data representation, navigation and communication.

129
Typically, numeric data is represented through visual methods such as graphs; con-130 cepts such as software architecture or algorithms are often presented as flowchart diagrams; 131 and 'concept maps' or 'mind maps' are common ways of organising and representing 132 related concepts. In discussing approaches to non-visual methods of co-design with 133 visually-impaired users, Metatla et al. (2015) propose diagram editing as a domain in which 134 haptic devices could potentially address access barriers. Ramloll et al. (2000) propose a 135 method of blending sonification techniques with haptic technology in order to present line 136 graph data .

137
Researchers have proposed haptic gloves to communicate information about the envi-138 ronment to VIB people in a wearable and portable format. Zelek et al. (2003) and Kilian et al. 139 (2022) propose using haptic gloves to support wayfinding and navigation, incorporating 140 cameras and computer vision software to represent the distance and location of objects in 141 the surrounding environment as haptic information. Gloves and wearables have also been 142 explored as aids to social interaction by converting non-verbal communication cues such 143 as facial expressions as haptic information (Krishna et al. 2010;McDaniel et al. 2008). The Haptic Wave (Tanaka and Parkinson 2016) is a prominent example of haptic 146 technology leveraged to support VIB sound creatives' access to music production soft-147 ware. The Haptic Wave is a custom-built prototype device consisting of a motorised fader 148 mounted on horizontal rails. The hardware couples with a software counterpart. This 149 device allows VIB users to 'feel' the amplitude of a waveform as they scrub across the X 150 axis, translating the waveform's amplitude to the motorised fader's position on the Y axis. 151 This design addresses a key issue which arises from the move from analogue studios and 152 their inherent physicality and tangible nature (for example, large mixing consoles, outboard 153 audio processors, reel-to-reel tape) to digital audio workstations, which rely heavily on 154 visual representations of waveforms. This GUI-based representation allows sighted users 155 to identify peaks in the waveform and silences, providing a reference for edit points such 156 as cutting and splicing between different takes. For VIB sound creatives, the visual rep-157 resentation of the waveform is not generally available; therefore, edit points are typically 158 found by 'scrubbing' through the waveform and listening intently in a linear process. The 159 Haptic Wave was designed over a series of co-design workshops and evaluated in situ as a 160 technology probe (Hutchinson et al. 2003). An interesting approach that the designers took 161 was to map audio information directly to the haptic domain, rather than aim to represent 162 visual information via haptics, to 'avoid translating paradigms of the visual into the haptic 163 modality' (Tanaka and Parkinson 2016).

164
The Moose (O'Modhrain and Gillespie 1997) is an early example of a prototype haptic 165 device targeted towards VIB users of specialist software. O'Modhrain and Gillespie discuss 166 its potential as a device for interacting with DAW software. It comprises a 'puck' that 167 moves on a 2D axis, similar to a computer mouse. The puck's freedom of movement can 168 be affected in two axes using linear motors to simulate force feedback effects and artefacts 169 such as detents and grooves. In this way, the authors propose a method of representing 170 GUI objects such as buttons and sliders via haptic feedback, allowing the user to navigate 171 DAW software.
The HaptEQ (Karp and Pardo 2017) is a low-cost, open-source, tactile device designed 173 for modifying graphical equaliser (EQ) curves via a tangible interface. Using a metal chain 174 on a magnetic sheet, the user can effectively draw and feel an approximation of the desired 175 EQ curve on a 2D plane. A webcam positioned directly above the chain captures images 176 for computer-vision-based software that translates the shape of the chain's curve into 177 parameter settings of an EQ plugin.

178
The touchEQ (Pesek and De Man 2021) is an application designed for the Surface 179 Haptics TanvasTouch 18 . The device utilises a phenomenon called electroadhesion, which 180 can simulate the effects of friction, and therefore different textures, across a smooth touch 181 screen surface. Pesek and De Man describe developing an EQ programme for this device 182 and a subsequent user test with sighted users. While significantly more expensive than the 183 HaptEQ, the touchEQ provides 2-way communication between the user and the device, 184 allowing the perception and manipulation of EQ parameters.

185
Outside the recording studio, haptic devices have addressed access barriers in perfor-186 mance contexts. Where a music performance requires a response to visual cues, for example, 187 an orchestra or choir following a conductor, VIB musicians rely on several methods to 188 work around the lack of visual feedback from the conductor. Pragmatic approaches to this 189 issue can include listening for breathing cues from neighbouring players or relying on a 190 neighbouring player to provide a physical timing cue such as tapping on the shoulder or 191 foot (Baker et al. 2019).

192
Many research projects have explored using haptic technologies to communicate in-193 formation between conductor and musician. Baker et al. (2019) propose translating the 194 movements of a conductor's right hand, via a motion sensor, into a 2D array of vibration 195 motors worn as a vest by the performer. The authors found that some of their participants 196 preferred a simpler vibrating 'pulse' pattern to the complex 2D representation of a con-197 ductor's arm movements. As with the Haptic Wave, this observation again highlights the 198 notion that a one-to-one mapping of visual information to haptic may ultimately be less 199 valuable to a VIB person, who may not be working from a visual reference point to begin 200 with.

201
The Haptic Baton 19 also focuses on translating the movements of a conductor's baton 202 into haptic feedback. The baton features sensors which track movement, which is translated 203 into signals conveyed by two vibration motors worn by the musician, one on each leg. 204 Kawarazaki et al. (2014) describe a similar project which uses a baton with a built-in motion 205 sensor. In their system, the movements of the conductor's baton are translated to 'beat' 206 signals conveyed by haptic devices worn by choir members. They also incorporate a 207 Microsoft Kinect 20 sensor to detect the direction the conductor is facing, allowing members 208 of the choir to perceive when they are being given specific instructions from the conductor, 209 such as to prepare to play one's instrumental part.

210
From related work, it is clear that haptic technology has great potential to address 211 several access barriers that VIB people face when using music technology. However, we 212 have yet to see a widespread uptake of haptic technology in the domain of VIB sound 213 creatives, for whom access tools which leverage the auditory modality (i.e. screen readers) 214 remain the norm. In the next section, we present an interview study with a group of VIB 215 sound creatives to unpack why they choose the tools they use and how new tools, including 216 those incorporating haptics, may fit in people's existing workflows. The study consisted of twenty remote, semi-structured interviews with VIB sound 220 creatives. One common thread connecting each participant is their use of DAW software as 221 a central tool in their creative pursuits. Each interview was two hours in duration. The first 222 three authors conducted the interviews and shared responsibilities for leading them. 223 We interviewed 20 people in total. Most of our participants are in the 35-44 age range, 224 with 3 in the 18-24 range and one in the 65-74 range. Two of our participants are female. 225 Most of our participants answered 'totally blind', or words to that effect, when asked to 226 describe their visual impairment. However, a few reported 'low vision' 'legally blind' or 227 'registered blind'. Half the participants described themselves as 'advanced' when asked to 228 rate their skill or knowledge in using music software, while nine answered 'intermediate', 229 and two answered 'beginner'. None of our participants uses Ableton Live, although several 230 mentioned this DAW. Reaper was the most commonly used DAW, followed by Logic 231 Pro, then Pro Tools. Other less common DAWs include Cakewalk Sonar, FL Studio and 232 Samplitude.

233
Although our research questions are specifically concerned with the experience of 234 VIB people, we acknowledge that some resultant insights may discuss ideas relating more 235 broadly to audio technology. These insights are still valuable, and as they arose naturally 236 from our chosen analysis methodology, we have chosen to include them in this article. 237 Additionally, our findings do not relate specifically to other disabled groups within the 238 wider sound creative community, and we encourage other researchers to explore these 239 avenues in their future work.

240
The first hour of each interview focussed on getting to know the participant. The 241 authors invited participants to complete a short questionnaire beforehand, which served the 242 practical goal of ascertaining their demographics, location, timezone, and choice of music 243 and accessible technology. The interview then delved deeper, attempting to understand 244 the participants' creative priorities, their views on their current choice of technological 245 tools, accessibility and inclusion in education and the music industry, aspects of music-246 technology accessibility to be celebrated, and points of frustration. As the vast majority 247 of participants were approached via online communities, such as mailing lists, the role of 248 these communities also formed a prevalent topic of conversation.

249
While the interviews adhered to a typical semi-structured methodology, we stimulated 250 the discussion further halfway through each interview by switching our approach to that 251 of a Contextual Enquiry. A contextual enquiry is, in essence, an interview in context 252 (Holtzblatt and Beyer 1997). Participants are asked to complete a task related to the 253 domain of interest while interviewers observe and ask questions. It is helpful to frame the 254 relationship between participant and interviewer as similar to that of master and apprentice. 255 The participant is the master, the interviewer the apprentice; the latter holding the objective 256 of learning the participant's approach to the task at hand. We asked each participant to work 257 towards a creative goal relevant to their typical creative pursuits. Between participants 258 working as sound designers, audio engineers, music producers, musicians and composers, 259 we observed tasks associated with various goals 260 The benefit of a contextual enquiry as opposed to a conventional semi-structured 261 interview is that it exposes aspects of a practical task that may be unarticulated, implicit 262 or taken for granted in a verbal response to a posed question. For instance, when asked 263 how they might typically start a DAW project, a participant may respond by stating that 264 they import an audio file. However, in observing a participant, one can see details of their 265 approach; for example, the use of any filename conventions, whether or not a particular 266 file import dialogue is utilised, and if not, the observer can ask why not. The creative goals 267 of our participants included the following examples: Composing a short section of music.

269
• Mixing a selection of pre-recorded audio stems.

270
• Designing a sound to reflect a magic spell being cast in a fantasy game.

271
• Audio editing to remove unwanted incidental sounds from recorded dialogue.

272
It is essential to highlight that a contextual enquiry of one-hour duration only provides 273 enough time to focus on a mere snapshot of a sound creative's workflow. However, this 274 snapshot is rich and deep and arguably maintains a greater ecological validity than a 275 semi-structured interview.

276
Conducting the study online and remotely was a double-edged sword. This approach 277 enabled the researchers to connect with international participants in their studio envi-278 ronment, increasing diversity with a low monetary outlay. The flip side was that using 279 Zoom video conferencing software somewhat disrupted the (at times) complex audio 280 configuration utilised by participants. Occasionally, it was not possible for participants to 281 use their regular audio interface, and new audio configurations introduced latency -the 282 arch-nemesis of the sound creative.

284
For this article, the authors extracted salient quotes from the interview transcripts 285 deemed relevant to applying haptic technology to address access barriers. We identified 286 discussions of haptic devices already incorporated into a participant's artefact ecology 287 and how tangible and tactile interfaces can improve access and ease of task execution. 288 Furthermore, we captured participants' perspectives on the opportunities presented by 289 haptics in their practice as sound creatives. Of the twenty interviews conducted, we 290 identified six as being particularly relevant to the topics covered in this article. We omitted 291 the remaining interviews from the thematic analysis; however, the authors will present a 292 full-scale analysis of all interview data in a forthcoming publication.

293
These quotations formed the source material upon which we performed a reflexive 294 thematic analysis (TA). Braun and Clarke (2021) spearheaded this flavour of TA, arguing 295 that all researchers are inherently situated and subjective. Furthermore, while academics 296 embedded in a positivist research culture may push against subjectivity, it is an asset to be 297 embraced. The reflexive researcher brings value (Ibid.). 298 Braun and Clarke utilise a six-phase process when performing TA, with phase six 299 amounting to writing a research report. It was this process to which we adhered. The first phase involves immersion with the dataset, which, in our case, were quo-302 tations from interview transcripts. Here began each author's immersion and subsequent 303 familiarisation with the dataset. We were fortunate to each attend the six interviews and 304 play an active role in each discussion. The AI-powered natural language processing tool 305 Otter.AI 21 provided initial transcripts of each discussion. However, due to each transcript 306 being laced with minor errors, the authors divided the transcripts between themselves, 307 proofread and corrected each, using the original recordings as a reference. While this was a 308 somewhat arduous task, it contributed significantly to our goal of dataset immersion.

309
Each author selected quotations from the transcripts they personally edited. These 310 extracts captured points salient to this article's research domain: the potential of haptic 311 technology within each participant's ecology of creative practice. For us, salience took 312 precedence over the frequency of ideas expressed.  The open-source web-based TA tool Taguette 22 supported the coding process. While 318 this software is constrained in functionality compared to NVivo 23 , it is largely accessible to 319 screen-reader interaction. This feature is essential as one team member is visually impaired, 320 and collaboration is critical in our work. Another factor is that, by necessity, we worked 321 remotely, a workflow supported by Taguette. We shared a code book that initially captured 322 succinct and somewhat reductive codes such as 'access barrier', 'community support' and 323 'efficiency is important'. However, these codes saw frequent revision, expanding their 324 detail and specificity. The goal was for these codes to stand alone and reduce the need to 325 return to the original text for explanation. We moved to Excel 24 spreadsheet software to generate initial themes from the codes. 328 This platform is also largely accessible to screen-reader interaction and supports remote 329 collaboration. Braun and Clarke argue that themes are not lying within a collection of codes 330 awaiting discovery. Instead, they are generated by the researcher(s) as they reflect on the 331 codes, their knowledge and their experience. While we focused on the six interviews we 332 deemed most pertinent to haptic technology, inherently, our experience and insights from 333 participation in the other fourteen interviews also coloured and informed our perspectives. 334 We grouped codes by categories, such as 'access strategy', 'role of industry' and 'benefits of 335 haptics'. Some codes appeared in multiple categories. From these categories, candidate 336 themes emerged. We spent several hours online discussing, developing and refining themes. At times, 339 the authors returned to the quotations to clarify specific points, a process beneficial in 340 corroborating emergent insights, which ensured we stayed close to the source material. This phase was largely an extension of the previous one. However, we also considered 343 the potential overlap between themes and their differences. Fundamentally we wanted 344 to ensure that the themes depict an accurate and compelling representation of the study 345 participants' thoughts, opinions and experiences.  347 We identified six themes, each split into sub-themes. Alongside each theme, we present 348 illustrative quotes from the interview transcripts. The quotes are presented verbatim but are 349 edited for brevity and readability in some places. We use our participants' names without 350 pseudonyms (with permission) and the names of the three first authors who conducted the 351 interviews.

352
Theme 1: In the absence of support from the mainstream music technology industry, online 353 communities have found it necessary to step in to improve 354 This theme illustrates a common observation among our participants on the role of 355 online communities of VIB sound creatives and the impact of the music-technology industry. 356 Many of our participants are active members of mailing lists, forums and messaging groups 357 relating to access tools and techniques. These online communities are typically DAW-358 centric, for example, the Reapers without Peeper's group, whose members share tips and 359 strategies on using Reaper software and contribute to the ongoing open-source Reaper 360 accessibility initiatives such as the Osara extension. Participation can come in the form of 361 asking/answering questions, providing tutorials and sharing custom scripts to bridge gaps 362 in software access. In the case of Reaper, the developers of the software itself engage with 363 online communities of VIB users in order to improve accessibility and address bugs: All the time, they're, they're fixing stuff for us. They're being proactive with all their 369 new software, you know, when they're releasing new software and stuff like that, they're 370 thinking about us, they're thinking about accessibility, they're thinking about the way 371 things are laid out and stuff like that. So I think that's one of the reasons why Reaper has 372 been super successful".

373
While the presence of these online communities was generally seen by participants 374 as a positive thing, some noted that these communities are required in the absence of full 375 accessibility support from music technology companies: Furthermore, software that may be considered ancillary to a DAW but an essential 394 part of a professional workflow may feature access barriers.  There are often numerous ways to learn how to use new music technology for sighted 402 people, such as through instruction manuals or video demonstrations. These resources may 403 be inaccessible to VIB sound creatives due to the use of graphical diagrams or incompatibil-404 ity with screen-reader software. Guides may focus only on a typical workflow for a sighted 405 person (e.g. using the mouse instead of keyboard-based navigation). It is common for VIB 406 sound creatives to look to online communities for tutorials geared towards VIB workflows. 407 Such resources can be presented in accessible formats, for example, via podcasts or by 408 formatting text-based materials to be screen-reader-friendly. Informal help and advice are 409 often sought through these communities, for example, by seeking recommendations on 410 software: Screen readers rely on relevant information being made available via attributes spec-419 ified in code by software developers. For example, in web browsing, the correct use of 420 HTML tags allows screen-reader software to correctly parse and navigate the information 421 contained in a website, such as heading structures, hyperlinks and image descriptions. 422 Screen readers require similar information from DAWs. For example, by correctly and 423 appropriately labelling plugin parameters in the source code. If this support is lacking, the 424 quality of information the screen reader relays to the user is negatively affected.

425
Justin: "Basically, you have some plugin developers, you look at their -so let's say there's 426 a slider with a four stage switch for an LFO shape, so it'd be sines or square or triangle.

427
On some plugins, you will see that: the screen reader will represent it as sine, square, 428 triangle. In other plugins, it'll be just zero to one, there'll be a slider going from zero to 429 one. So you have to know that you need to bring your slider to 0.25 to make this work,  Generally speaking, accessibility continues to improve for VIB sound creatives. These 435 improvements are partly due to technological developments, leading to computers being 436 powerful enough to run DAWs reliably alongside screen readers and hardware capable of 437 supporting TTS feedback. However, improvements in accessibility are not solely a result of 438 technological developments. Efforts from the VIB community and industry have leveraged 439 these developments for accessibility purposes.

458
VIB people are experts in their tools, workflows and lived experience of disability. Soft-459 ware developers and hardware designers can follow accessibility standards (for example, 460 the Web Content Accessibility Guidelines 28 (WCAG) in the case of website design). How-461 ever, insight into how well a piece of technology meets access requirements can only come 462 through the inclusion of VIB people in the design and development process. An example 463 of this is the way that Reaper developers regularly engage with their VIB users. It was also 464 common for our participants to mention their willingness to engage with developers to 465 report bugs and feedback and their mixed experiences with how well their requests were 466 met.

467
Theme 2: Access comes through a multitude of factors not limited to technology 468 Many of our participants work in broad creative ecosystems. These ecosystems do not 469 merely contain technological artefacts (for example, DAWs, plugin instruments and effects, 470 screen readers, scripts and MIDI controllers -but also include the less tangible online 471 tutorials, documentation, online communities and their collective knowledge experience. 472 Existing within and managing such an ecosystem can often be challenging for VIB sound 473 creatives, especially in times of change, such as major software updates or when introducing 474 new technological artefacts. Some participants voiced frustrations around conflicting or 475 malfunctioning tools, leaving parts of their workflow ineffective. VIB sound creatives often encounter numerous accessibility issues within their com-484 plex ecosystems. Due to the highly specialised nature of many of these issues, and a 485 perceived lack of support from the technology industry, many of our participants demon-486 strated unique workarounds. However, through online research, community input and 487 trial and error, participants found solutions; that they often invested much of their time in 488 creating. Many of our participants' uniquely curated workflows were highly multimodal, com-492 bining auditory and tactile feedback. In particular, we noticed several people using their 493 screen reader in conjunction with a keyboard controller or dedicated control surface, sup-494 porting efficiency in their workflow.

495
Peter: "So normally, the easy way to do that [edit EQ parameters precisely] is with the 496 control surface, which then puts the main bands that I want access to -the high shelf, the We also observed how some participants would deliberately not use their screen 508 reader's TTS feedback, relying solely on the familiar key commands to manipulate the 509 audio without the speech disturbing concentration. Some of our participants demonstrated a reliance on memory when navigating menus 517 or timelines. They remember what they encounter to construct a mental map, which they 518 can reference later. In this way, memory can become an effective tool for increasing usability 519 and efficiency in the workflow. 520 Jacob Harrison: "Is the information that the screen reader is giving you, is that useful?

521
Are you happy for it to read that out, and then wait for it to stop talking?" Learning how to use new technology was a common experience described by our 527 participants. They describe how online tutorials are often inaccessible and designed 528 visually for sighted, mouse-oriented users. Without accessible learning material, VIB sound 529 creatives often need to rely on instruction manuals provided by the developer of their tool. 530 If these materials are not screen-reader compatible, a person might be forced to sacrifice 531 independence and find sighted assistance, or begin using the tool completely unaware of 532 its functionality. It can be argued that one key product of an access barrier is inequality. An access 539 barrier restricts possible options; presenting the VIB sound creative with a user experience 540 which differs to that of their sighted peers; the experience of the two is not equivalent. 541 This problematic seed can grow and cause secondary issues, for instance inequivalences in 542 employment, education, social and creative opportunities.

543
Theme 3.1 VIB sound creatives share common creative and professional goals with sighted 544 users but have different workflows and unequal opportunities.

545
Several of the study's participants expressed creative and professional goals that 546 one might consider common to those of sound creatives more broadly, without visual 547 impairment or blindness. For instance, the goal of forging a career as an audio engineer or 548 music producer in a professional studio is typical of sighted sound creatives. 549 It is apparent that in bringing accessibility to their DAW, VIB sound creatives need 550 to employ a workflow that fundamentally differs from sighted conventions. The screen 551 reader affords an alternative control modality, with information articulated through synthe-552 sised utterances, accessing features and functionality through QWERTY keyboard-based 553 navigation. It would be too bold to claim that VIB sound creatives are essentially operating 554 a different piece of software. However, the experience is undeniably different to that of 555 their sighted peers. The challenge brought by this discrepancy primarily relates to a need 556 for a shared experience with those outside of the VIB sound creative community.

557
Theme 3.2: There are fewer employment opportunities for VIB sound creatives 558 Except for Logic Pro and arguably Pro Tools, DAWs do not typically provide full 559 access to their features and functionality through screen reader interaction. Access comes 560 through bricolage, a combination of hardware and software; each to plug a particular gap 561 in accessibility. Such tools are not typically installed in professional recording studios. 562 Therefore, VIB sound creatives often cannot simply enter such a facility, roll up their sleeves 563 and engage with the task at hand. Therefore, a lack of appropriate tools in this context 564 has a negative impact on employment opportunities. A degree of out-of-the-box DAW 565 screen-reader compatibility will likely go far in addressing this workplace inequality.

566
Peter: "You still can't just rock up at a studio that you've not been to before, and use 567 the equipment. That would be the dream, and it's miles away from being anywhere near 568 feasible. And so it's still very difficult, I think, for a blind person to just get an ordinary 569 job in the mainstream studio." 570 Theme 3.3: VIB sound creatives may not necessarily be at a disadvantage in sound-based 571 creative practice.

572
As with many observations, we cannot sit solely on one side of the fence. It is not 573 simply the case that inequivelances are innately negative, leading to barriers in access. 574 Conversely, the VIB DAW workflow may be preferable in some aspects of sound-based 575 creative practice.

576
Maja Sobiech: "I've heard about sound engineers who purposefully turn (off) their 577 screens, or don't look at them to not be distracted."

578
As stated by Maja above, sometimes visual stimuli can be distracting. Furthermore, 579 we have witnessed incredible efficiency by some VIB sound creatives in completing specific 580 DAW-based tasks, employing an arsenal of keyboard shortcuts and commands.

581
Theme 4: Hardware may bring enhancements to VIB sound creatives 582 Commercially available devices incorporating haptic feedback were rare among our 583 participants. However, one participant had experience with the Haptic Wave prototype 584 (Tanaka and Parkinson 2016), while another had previously used a custom device for 585 peak metering, which incorporated a modified braille cell 29 . Despite haptic technologies 586 featuring less prominently than other tools among our participants, many reported the 587 advantages of hardware in the form of MIDI controllers, control surfaces, analogue mixing 588 desks, and modular synthesisers. While many sighted sound creatives also regularly use 589 commercially-available hardware, the specific benefits of such hardware to VIB sound 590 creatives is relevant to the haptic-audio community. Participants implied that analogue 591 hardware aided usability due to the predictability and consistency a 1:1 mapping between 592 control and effect affords, aiding efficient workflows. It is worth noting that this particular benefit of hardware can be lost in the digital 600 domain due to the ability for devices to map the same knob, slider or button to a wide 601 range of parameters. Some manufacturers have addressed this issue, including Native In-602 struments with their Komplete Kontrol range 30 of MIDI controller keyboards and Arturia's 603 equivalent offering, Keylab 31 .

604
An additional benefit of hardware that is relevant to both analogue and digital devices 605 is the ability to focus on audio that is the subject of an edit or processing operation, as 606 opposed to audio feedback from the screen reader: 607 29 Unfortunately no details of this bespoke device can be found online. 30 https://www.native-instruments.com/en/products/komplete/keyboards/komplete-kontrol-s49-s61/ 31 https://www.arturia.com/ranges/lab Maja: "For example, if I am able to switch to using a braille display with Pro Tools, I 608 think it will be beneficial for my work because I don't have to use VoiceOver that much. 609 And I don't distract my hearing, then it might be useful there" 610 A number of our participants also modified their existing hardware, for example, 611 using a Dremel to make a tactile mark on a knob cap or using embossed braille tape on a 612 MIDI keyboard. Other participants did not make direct reference to haptic technology but discussed 636 methods and tools which could be described as haptic devices, for example, a modified 637 braille cell which could be used for peak metering without the need for auditory feedback: 648 Some of our participants described experiences of using tactile methods in their 649 education as a means of conveying visual information. One participant described how their 650 music teacher helped them visualise how a musical score looked to sighted people. And what these quarter notes look like, and a whole note looks like so at least if somebody 658 described it to you on a page, you can understand it.'" 659 There is potential for tactile and tangible objects to communicate concepts often 660 presented visually. Such an approach presents learning and education opportunities for 661 VIB people, as described by our participant, but also potentially widens the communication 662 bandwidth between VIB and sighted people. Screen readers, which are just one tool in a VIB sound creative's ecosystem, are only a 665 solution to some access barriers. Screen readers are limited in their functionality. Access 666 barriers can arise when software designers and developers do not consider the screen-667 reader-interaction modality.

668
Theme 5.1: Screen readers are an imperfect access strategy due to the limitations of the TTS 669 modality.

670
Screen readers possess one major limitation; they convey information linearly. Screen 671 readers invoke a speech description of most elements they encounter, meaning that it 672 can take a considerable amount of time for a VIB person to parse information required 673 in executing a task. Furthermore, it can be highly challenging to garner information 674 from multiple sources concurrently via speech feedback alone. This issue often leads to 675 VIB preferences for naming conventions that first state the most contextually essential 676 information. Similarly, our participants discussed the problems they encounter when their 677 screen reader is excessively verbose or distracting when working with audio.  Screen readers are complex tools which require technical skills and an understanding 700 of user requirements to implement effectively in software. Additionally, they are specialised 701 tools which means that certain manufacturers, designers and developers may need to be 702 made aware of their existence. With comprehensive support of the screen-reader interaction 703 modality in music software, VIB sound creatives can avoid significant access barriers in 704 their creative pursuits. In software platforms with no screen-reader support, some of our 705 participants told us that this would exclude that particular software as a potential candidate 706 for integration into their artefact ecology.

707
Trey: Yes, there's major disadvantages. . . In my subjective experience from production, 708 yes I do feel disadvantaged. I can't use Ableton for a start. But everyone else is using 709 Ableton. It's a massive disadvantage in a few ways. Number one, it affects your learning. Screen reader compatibility is overlooked when software is designed solely with visual 721 users in mind. This situation can result in unintuitive navigation (through software pa-722 rameters inadvertently grouped in unusual ways), missing information (such as parameter 723 names), and in the case of highly visual interfaces (such as those that depict waveforms 724 graphically), a complete lack of functionality.

738
When access barriers arise, VIB sound creatives often rely on solutions which radically 739 alter their workflow from the typical sighted approach. These approaches are often shared 740 amongst online communities and widely adopted as functional access strategies. For 741 example, one of our participants described how he separates the screen-reader TTS feedback 742 and the audio from Reaper to different sources. can do all of that. We can't. We dive through the presets in alphabetical order very slowly.

748
And there are 1000s of presets. And then we find what we like, and we fiddle with it. And 749 we hope the controls which we would like to fiddle with are exposed. And maybe they are 750 and that's great. And then that's how it goes yeah. There's so much that sighted people 751 can access with the whole Native Instruments deal, and we can't." 752 Theme 6: Accessibility isn't a VIB sound creative's only priority 753 Many of our participants discussed having priorities that were not limited to issues 754 of accessibility. There are many reasons behind a VIB sound creative's choice of tools and 755 workflow. Many of these reasons are common across all music-technology users and not 756 unique to VIB people -for example, financial costs.

757
A common concern among participants was the ability to work as quickly or efficiently 758 as sighted peers. This concern is a particular issue for professional VIB sound creatives, 759 who are at a disadvantage in employment if they cannot compete with their sighted 760 counterparts.

761
Peter: "That's another thing where, however much we'd like to be able to say, 'okay, we 762 can compete on an equal footing with a sighted sound engineer', it's not true, because 763 some of the things that you have to do will simply, they just take longer, they just do. And 764 if you've got something that has to be done in two hours. And it's something a sighted 765 person would be able to do in two hours because they can 'click click click' and it's done - 766 we have to use the workaround" 767 Theme 6.1: VIB sound creatives do not just want things to be possible; they want things to 768 be usable 769 While some of our participants reported being able to use several tools, many high-770 lighted frustration with usability issues, particularly in comparison to sighted people's 771 experience. For example, tools such as custom accessibility scripts and VOCR can compen-772 sate for several usability flaws. However, these workarounds still require additional effort 773 and place an extra strain on VIB people. There are perceived advantages to using the same tools and practices that are popular 788 with sighted sound creatives. The use of common tools allows ease of communication 789 between sighted and non-sighted people, supporting collaboration. It is also important 790 to note that different products come with an attached social capital, for example, Ableton 791 Live software, which is common in electronic music production. While other DAWs may 792 currently be more accessible to VIB people than Live, there is a desire to use this software 793 for its recognisability within a particular genre and the common experiences that this can 794 support:

802
By comparing the themes generated from our interviews with the latest efforts in 803 accessibility and haptic-audio research, we can consider how haptics may play a role in 804 addressing access barriers experienced by VIB sound creatives. It is important to highlight 805 that only one of our participants has used the technology mentioned in our discussion of 806 existing haptics research, and we hypothesise that a general absence of such technology in 807 our findings suggests potentially unexplored opportunities. We now discuss our findings 808 in relation to these opportunities, which could provide a jumping off point for future 809 accessibility related haptic-audio research. An obvious advantage of haptics in the VIB accessibility domain is that haptic tech-812 nologies are inherently tactile and tangible, a common attribute of many access tools and 813 strategies employed by VIB people. The most prominent example of tactile, tangible access 814 is in the use of braille, whether reading braille-embossed text, accessing web content via a 815 braille display, navigating public spaces using braille signage, or using braille music to read 816 scores. Some participants described how teachers incorporated tactile and tangible teaching 817 methods into their education. In contrast, others discussed how hardware, sometimes with 818 modifications, benefited their workflow. Here, we specify some specific areas in which we 819 believe haptic technologies could leverage or support the benefits of tactile and tangible 820 interaction to improve accessibility based on the common access barriers encountered by 821 our participants. Some of our participants with experience in analogue recording studios or who cur-824 rently use external hardware discussed the difference between tangible, physical devices 825 and modern, highly graphics-based software. As vision-centric music software has pro-826 liferated, so too have access barriers for VIB sound creatives. Before music studios were 827 based around a computer running DAW software, analogue equipment provided tactile 828 means of navigating and manipulating channel strips and effects parameters. Some of 829 our participants continue to use analogue equipment, citing this tactility as an advantage 830 over working in the digital domain. Other participants use outboard gear such as MIDI 831 controllers, audio interfaces and modular synthesisers. The tactility and ease of navigation 832 of some hardware was mentioned as something VIB sound creatives found helpful in their 833 workflows, with many examples of modified or DIY hardware being used.

834
The above observation suggests that the predictability and consistency of hardware 835 is something VIB sound creatives value. Despite this, there is a trend towards hardware 836 devices whose physical design is agnostic to the functions they may be assigned, in order 837 to maintain flexibility in mapping. It is common to see hardware which features poten-838 tiometers with no tactile markings, end stops, or detents so that they can be mapped to any 839 number of functions within a DAW or plugin. While this is a valuable feature for sighted 840 users, VIB sound creatives may need a way of determining the control's current state and 841 its associated parameter. VIB sound creatives can value predictability and consistency in 842 the structure and function of interfaces, which is easily lost in generic control surfaces with 843 several modes of operation.

844
A method for working around the limited tactility, predictability and consistency 845 in some commercial hardware is through TTS support, allowing VIB people to query 846 and determine the current mapping and value of a given fader, knob or button. This 847 approach could extend further using haptic feedback. For example, to simulate detents 848 on a continuous potentiometer or to communicate functionality via 'haptons', the use of 849 different haptic textures to delineate different areas of the interface, an approach that has 850 been explored by Pesek and De Man (2021). This approach would bypass the auditory 851 modality, allowing the sound creative to focus on the immediate task. Haptic feedback can 852 help delineate between parameter types by using consistent vibration patterns to signify the 853 assignment of parameters related to one another, indicating the control's function without 854 TTS feedback A hardware project which we believe shows some promise is the Smart Knob, an 856 open-source haptic device developed by Scott Bezek 32 . The Smart Knob incorporates a 857 brushless motor and magnetic encoder to provide 'closed-loop torque feedback control, 858 making it possible to dynamically create and adjust the feel of detents and endstops'. The 859 project is not explicitly aimed at music-technology use cases, but it is possible to imagine a 860 similar implementation of this technology within other hardware. Several participants cited peak metering as a common task for which few options 863 provided a DAW experience equivalent to a sighted user. For sighted people, a glance at an 864 on-screen mixer or channel strip is enough to rapidly determine whether a track is clipping 865 and give an oversight of the relative amplitude levels of a multitrack project. There are 866 few options capable of providing this same overview of volume levels in a project without 867 sight. Our participants gave examples of workarounds for this scenario, for example, via 868 DIY devices or by using 32-bit recording to allow for more headroom to avoid clipping 869 during the recording stage.

870
Incorporating haptic feedback, potentially via haptic displays, might offer VIB people 871 a means of peak metering, which is more equivalent to the 'glance' afforded to sighted 872 people. This approach has already been explored to some extent by the Haptic Wave 873 (Tanaka and Parkinson 2016). However, this relies on the user scrubbing across a single 874 waveform, so it does not provide an overview of multiple tracks concurrently.  We were intrigued by one of our participant's comments on the potential of haptic 877 devices to support work with Dolby Atmos, describing a Rubik's cube-like device which 878 allows the user to work with spatial audio in a 'physical way'. Working with spatial audio 879 currently requires a means of visualising a 3D space on-screen in order to pan audio objects. 880 This scenario could be well served by a tactile device capable of communicating spatial 881 information non-visually. Indeed, haptic rendering of 3D objects and environments is a well-882 explored topic ( Many of the people we have spoken with are professionals or people with significant 897 experience interacting with DAW software via screen readers and software extensions. 898 The response was mixed when discussing if they felt at a disadvantage compared to their 899 sighted peers. We noted that many VIB sound creatives had developed incredibly efficient 900 workflows with the available tools. Some participants stated that they worked more 901 effectively than sighted people in certain scenarios, for example, due to their mastery of 902 keyboard shortcuts for rapid navigation and manipulation within DAW software. When 903 32 https://github.com/scottbez1/smartknob we are discussing introducing new technologies to people's workflows, it is essential 904 to consider that many already have workflows that work well for them. We should be 905 careful to not discard the work done already by people with lived experience of visual 906 impairment or blindness by proposing radically different ways of working as solutions to 907 access barriers.

908
In contrast, while some of these existing workflows and tools have proven to be highly 909 effective for some people, we also noted that this came after a considerable amount of addi-910 tional time and effort in learning. In many cases, this effort also involved the development 911 of bespoke tools or paying for custom keyboard scripts and additional software. The cur-912 rent model for DAW accessibility for VIB people appears to be largely based on DIY efforts 913 and open-sourced, community-led software, although music-technology manufacturers are 914 increasingly demonstrating efforts to improve accessibility out of the box. Either way, we 915 suggest that the onus shouldn't necessarily be on VIB people and unpaid, community-led 916 efforts to work around access barriers.

917
The observations above lead us to a tension between acknowledging the existing 918 efforts of VIB people and communities to develop tools and workflows which work for 919 them, but also aiming to lessen this burden on VIB people in the future; to meet the goal of 920 lowering barriers to access for VIB people who have not yet begun their learning journey 921 with a DAW.

922
An additional factor when considering the use of new, specialist tools within this 923 community is the potential for groups to be siloed or organised around a particular tool or 924 software. Some participants reported the potential social impacts of not having access to a 925 commonly used tool. These inequivalences are particularly stark when a tool is considered 926 'industry standard', affecting employment prospects, or is strongly associated with a 927 particular genre, which could hinder building social connections with people working in 928 the same genre. New tools should consider the importance of compatibility with common 929 tools and workflows, as a means of leveraging shared knowledge and supporting social 930 connections around a particular tool.

932
This article has combined existing research in the haptic-technology domain with 933 insights provided by VIB sound creatives. These insights reflect our participants' unique 934 lived experiences and should be considered valuable when considering future design 935 avenues in the AMT domain. As researchers and authors, we acknowledge that we 936 work subjectively and reflexively. However, that counter to positivist objections, our 937 subjectivity and reflexivity add richness and depth to our findings. We have both named 938 our participants and quoted them extensively to highlight the relevance and agency of their 939 voices, both individually and collectively. 940 We have identified several access barriers which could be partially or fully overcome 941 through the careful implementation and provision of new haptic prototypes. Whilst these 942 design possibilities present an exciting opportunity for developing new technologies, we 943 urge the community to recognise the importance of the socio-political factors that arose 944 from our thematic analysis. VIB sound creatives often self-organise into internationally 945 supported online communities that emphasise inclusive practice and invest significant time 946 and effort into producing free, open-source access strategies. Preserving these values of 947 inclusivity and availability, as well as the existing collection of accessible VIB workflows, 948 should be at the heart of any design proposals moving forward.

950
Jacob Harrison is a middle-class, thirty-one-year-old white British male who identifies 951 as non-disabled. His background is in academic research focusing on accessibility in music 952 making and how the design of new technologies can impact disabled people's access to 953 music, whether positively or negatively. He also uses DAWs as a recording studio Session 954 Leader with learning-disabled young people in his freelance work. James Cunningham is a Christian, visually impaired, white, twenty-five-year-old male 956 from Northern Ireland. He is an experienced composer and improviser. At the time of this 957 paper's publication, he is studying for a PhD at Queen's University Belfast, combining his 958 artistic practice with an exploration of accessible music technology.

959
Alex Lucas is a middle-class, forty-year-old, white English male who identifies as 960 non-disabled. Alex has a background in product design in the music-technology industry. 961 His primary role was the design of user interfaces on hardware synthesisers. His academic 962 research focuses on accessibility and inclusion in music-making practices. Several years ago, 963 this author worked as a part-time support worker for Mencap's children's services. This 964 experience of working with a diverse range of children shaped Alex's view on disability, 965 seeing it as highly situated and unique to the individual. Alex also spent some time as 966 a sound technician working with the now-disbanded Surrey Heath Talking Newspaper 967 for the Blind, recording local news items to be distributed by audio cassette tape. The 968 latter experience helped Alex to appreciate that some cultural forms are unavailable to VIB 969 people and that solutions to such barriers sometimes involve antiquated technology.