Collaborative Creation of Digital Tactile Graphics
Jens Bornschein, Denise Prescher, Gerhard Weber · 2015 · ASSETS '15: Proceedings of the 17th International ACM SIGACCESS Conference on Computers & Accessibility · doi:10.1145/2700648.2809869
Summary
This paper presents the Tangram Workstation, a collaborative system enabling sighted transcribers and blind reviewers to work together in real-time when creating tactile graphics. The fundamental insight is that current tactile graphics production involves sighted transcribers working alone, with blind users only seeing the final embossed product—a workflow that often produces suboptimal results requiring costly revisions. By involving blind users during the digital creation phase, quality can be improved before physical production. The system consists of two components: a Tangram Toolbar extension for LibreOffice Draw (used by sighted transcribers) and Tangram Lector software connected to a BrailleDis 7200 pin-matrix device (used by blind reviewers). The 120×60 pin display at 10 dpi resolution allows blind users to explore the graphic tactilely in real-time as it is being created. Both users have independent "focus" indicators—the blind user's Braille-focus appears as a red overlay on the sighted user's screen, enabling awareness of what their partner is examining. The blind reviewer can operate in multiple modes: passively observing, actively consulting (giving verbal feedback), or actively editing (directly manipulating elements using the pin device's cursor keypad). Touch gestures on the pin device trigger audio feedback providing element identifiers, titles, and descriptions. The system follows BANA guidelines for tactile graphics, including preset line styles, textures, and text settings optimized for Braille dimensions.
Key findings
The evaluation involved eight teams (sighted transcriber + blind reviewer) transcribing a complex nitrogen cycle diagram. Four sighted participants were professional tactile graphics transcribers; four were laypeople. All blind participants were trained Braille readers with varying experience levels. The collaborative approach demonstrably improved graphic quality. Average quality ratings (0-10 scale) increased from 5.9 to 8.5 for blind users and from 5.0 to 8.0 for sighted users after collaboration. Laymen teams showed the greatest improvements. Critically, blind reviewers recommended changes even on graphics prepared by professional transcribers—nearly all teams changed arrow heads (transcribers used arrows that were too small), six of eight teams widened line weights, and teams added spacing, labels, and structural changes. Three collaboration types emerged: passive observers (guided by sighted partner), active consultants (giving instructions but not editing directly), and active graphics editors (directly modifying elements). Prior experience with pin-matrix devices predicted which role blind users adopted more than their tactile graphics experience. On average, 2.3 changes were made per graphic; 80% of blind reviewer suggestions were implemented. The complete workflow averaged 131 minutes, with collaborative editing taking 40-57 minutes depending on team expertise. While not faster than traditional workflows (1-2 hours for initial creation), this approach eliminates costly revision cycles after physical production.
Relevance
This research challenges the assumption that tactile graphics production must be a sighted-only activity validated post-hoc by blind users. By enabling real-time collaboration, the Tangram Workstation allows blind users to influence design decisions during creation rather than merely accepting or rejecting finished products. The finding that even professional transcribers' work benefited from blind reviewer input underscores that expertise in production techniques does not substitute for lived experience with tactile reading. For accessibility practitioners, several insights are actionable: arrow heads and thin lines are frequently too small for tactile perception; spacing and structural organization matter as much as content accuracy; and direct user involvement during creation improves outcomes more efficiently than iterative revision cycles. The three collaboration types observed (passive, consultant, active editor) suggest that tools should support varying levels of blind user agency based on experience and preference. The pin-matrix device's 10 dpi resolution proved limiting—fine details visible on screen disappeared on the tactile display, and participants sometimes preferred hardcopy embossed versions for final review. This highlights an important caveat: digital previews on refreshable displays may not fully predict final tactile readability on higher-resolution production media like embossers or microcapsule paper.
Tags: visual impairment · tactile graphics · collaboration · pin-matrix display · refreshable braille display · transcription · assistive technology
Standards referenced: BANA Guidelines for Tactile Graphics