eNTERFACE ‘08: Project4 Design and Usability Issues for multimodal cues in Interface Design/ Virtual Environments eNTERFACE ‘08| Project 4

eNTERFACE ‘08: Project4 Project Team Team Leader: Catherine Guastavino, McGill University, Canada Team Members: Emma Murphy, McGill University, Canada Charles Verron, France Telecom R&D and Laboratoire de Mecanique et d'Acoustique de Marseille (LMA), France Camille Mousette, Ume Institute of Design, Sweden

eNTERFACE ‘08: Project4 Project Proposal The main aim of this project was to run usability tests to investigate issues of integration and effectiveness of information delivery between audio and haptic modalities. It is proposed that an interface involving a target finding task with audio and haptic (touch) feedback is relevant to this aim Furthermore target finding using multimodal cues is relevant to the field of New Instrument Design and also the wider field of Human Computer Interaction

eNTERFACE ‘08: Project4 Literature Various studies have indicated that the use of non-speech audio and haptics can help improve access to graphical user interfaces (Mynatt and Weber, 1994; Ramstein et al., 1996), by reducing the burden on other senses, such as vision and speech. Studies have specifically investigated the use of audio and haptics to convey object location in a spatial structure (Wood et al., 2003; Lahav and Mioduser, 2004; Murphy et al., 2007) Previous studies have also investigated the use of 3D audio with gesture for target finding in virtual environments (Marentakis and Brewster, 2005 )

eNTERFACE ‘08: Project4 Interface Design Audio Feedback Non-Speech Auditory Cues Freesound MAX/MSP Ircam Spat Object Haptic Feedback PHANTOM OMNI H3D API

eNTERFACE ‘08: Project4 Audio-Haptic Design The proposed idea was to implement a target finding task with haptic feedback using the PHANTOM and non-speech audio cues. A virtual environment composed of a number of parallel planes was created with a target located randomly on one the the planes Haptics: A magnetic effect was used to create a rigid surface for the planes and also on the target. Audio: Auditory cues were designed based on a string instrument (a cello) utilizing 3D spatial audio

eNTERFACE ‘08: Project4 Audio-Haptic Design Horizontal and Vertical conditions Audio: Crossing the planes Target Location Cue Target Found Cue Haptics: Magnetic Effect on the surface of planes and target

eNTERFACE ‘08: Project4 Audio-Haptic Design We used the IRCAM SPAT object for 3D sound spatialization over headphones using binaural rendering with HRTF database by (Martin et al., 1994; Gardner et al, 1994) The virtual sound source (the bowed cello sound) is spatialized using the “ears in hand metaphor” The virtual sound source is played only when the target and the stylus are located on the same plane (horizontal or vertical, according to the configuration).

eNTERFACE ‘08: Project4 Experimental Design Independent variables Feedback = audio only or audio-haptic Orientation = vertical or horizontal Resulting on 4 experimental conditions Audio-Haptic Vertical Audio-Haptic Horizontal Audio-Only Vertical Audio-Only Horizontal Dependent variables completion times trajectories perceived effectiveness and ease of use cognitive strategies

eNTERFACE ‘08: Project4 Experimental Hypotheses Hypotheses: 1. Users would find the audio only condition more difficult to navigate without the support of the haptic planes 2. Investigate the effect of flipping the planes from vertical to horizontal orientations in both audio only and audio-haptic conditions.

eNTERFACE ‘08: Project4 Demo Video

eNTERFACE ‘08: Project4 Experiment: Target Finding Task 23 Participants Training Introduction: Users became familiar with audio-haptic cues using a visual representation of the planes. Users were asked to navigate the planes, firstly find the plane with the target and then locate that target. Trial Experiment: Users were presented with 8 trials, 2 of each condition. Users were not give any information about the 3D audio mappings or the haptic feedback Main Experiment - 44 Trials (11 per condition) Condition, Target position within and across planes were randomised

eNTERFACE ‘08: Project4 Initial Results: Completion Times Initial completion time analyses confirm hypothesis; audio completion times are significantly longer than audio-haptic condition. Furthermore the vertical condition took significantly longer than the horizontal condition

eNTERFACE ‘08: Project4 Initial Results: Interaction effects Factorial Anova: Significant effect of feedback Audio vs. audio-haptic (p=>.001) Orientation: vertical vs. horizontal (p=>.01)

eNTERFACE ‘08: Project4 Analysis Strategies From observation the most efficient users were those who immediately grasped the structure of the virtual environment and understood the 3D audio cues. Spatial audio - elevation cues more difficult to perceive without individualized HRTFs Gestures Interesting gestural use of the haptic device Some participants changed hand movement according to the position of the planes One participant had an interesting gestural strategy of recreating the haptic planes with his free hand in the audio-only condition Further Analysis Post-task Questionnaires Trajectories - to further analys gestural control and to investigate random identifications of the target source

eNTERFACE ‘08: Project4 Demo: Recreating User Trajectories

eNTERFACE ‘08: Project4 Future Work Further Analysis Post Task Questionnaires Trajectories Further quantitative results Further Evaluations Develop cues Implement experiment using other haptic devices Applications Visually impaired users Small screen devices

eNTERFACE ‘08: Project4 Summary The aim of this project was to highlight usability issues for audio haptic cues by conducting user evaluations of a multimodal interface. Experiments confirmed our initial hypothesis that the audio-only condition would be most difficult for users to navigate Further analysis will focus on the qualitative comments from post task questionnaires and also trajectory analysis and gestural movements We intend to develop this study in terms of the audio-haptic cues, using other haptic devices and also extend the perceptual evaluation to investigate other aspects of multimodal integration