Section: New Results

Understanding and modeling users

Participants : Géry Casiez [correspondent] , Stéphane Huot, Edward Lank, Justin Dan Mathew, Mathieu Nancel, Sylvain Malacria, Nicolas Roussel, Marcelo Wanderley.

3D audio production tools vary from low-level programming libraries to higher-level user interfaces that are used across a wide range of applications. However, many of the user interfaces for authoring 3D audio parameters are underdeveloped, forcing users to resort to ad hoc solutions with other tools or programming languages, which limits creativity and productivity. Even though there is a significant increase of interest in this problem, usability issues with the manipulation of spatial parameters in 3D audio tools are still not well identified. We have thus conducted an on-line survey with practitioners to gather ethnographic information on their tools, methods, and assessments [15]. Our goal was to identify limitations and custom methods to circumvent them, in order to inform the development of better user interfaces for 3D audio production. Results of the survey revealed specific methods and limitations with regards to Audio Rendering & Recording, Visual Feedback, Functionality, and Workflow Integration. We also identified three basic but important tasks that have to be performed interactively with 3D audio production tools: Defining the Rendering Space, Creation and Manipulation of Audio Objects, and Monitoring with Audio/Visual Feedback. As part of Justin Mathew's PhD [11], this classification helped to identify the needs and to experiment new 3D audio tools that address issues with low-level functionality of 3D audio production systems such as the specification of 3D audio trajectories.

Play and playfulness compose an essential part of our lives as human beings. From childhood to adulthood, playfulness is often associated with remarkable positive experiences related to fun, pleasure, intimate social activities, imagination, and creativity. Not surprisingly, playfulness has been recurrently considered an important criterion in the design of interfaces for Digital Musical Instruments. It is supposed to engage people, often non-expert, in short term musical activities. Yet, designing for playfulness remains a challenging task, as little knowledge is available for designers to support their decisions. To address this issue, we have followed a design rationale approach using the context of Live Looping as a case study [21]. We first surveyed 101 Live Looping tools and summarized our analysis into a new design space. We used this design space to discuss potential guidelines to address playfulness in a design process: (i) advanced looping capacity, so as to extend the musical possibilities of Live Looping; (ii) low input capacity and direct mappings, in order to help getting familiar with the instrument and to favor direct control; and (iii) transparent and intense visual feedback to help infer what is happening inside the device when an action is performed. These guidelines were implemented and discussed in a new prototype of Live Looping instrument, the Voice Reaping Machine.

BCIs are presumably supposed to require the full attention of their users, and to lose accuracy if users pay attention to another task. This assertion has been verified with several BCI paradigms (e.g. P300). But the cognitive demand of the promising SSVEP paradigm had never been specifically assessed. We measured the accuracy of an SSVEP-based BCI used by 26 participants in various conditions of mental workload [23]. Our analysis revealed that surprisingly, for this type of BCI, little attention is actually needed from participants to reach optimal accuracy: participants were able to successfully perform a complex secondary task (N-back) without degrading the BCI accuracy. The same observation was made whether visual or auditive attention was solicited. These results indicate that SSVEP is a low-demanding paradigm in terms of cognitive resources, and are encouraging for its use in complex interaction settings. Last we did a survey on the use of BCIs in augmented reality. The results of the survey show that most of the previous work made use of P300 or SSVEP paradigms with EEG in Video See-Through systems, and that robotics is a main field of application with the highest number of existing systems [34].

In 2D and 3D interfaces, a “steering task” consists in following a predefined path of arbitrary shape, with a given and possibly varying tolerance (or “width”), using the finger or the cursor. While less common than target acquisition, this family of pointing tasks is a distinct component of modern interaction, e.g. selecting items in a hierarchical linear menus, or deep-etching an image in Photoshop. Previous work have essentially modeled straight or circular paths of constant width, and argued that more complex paths can be modeled using combinations of these primitives. We demonstrated that existing models actually fail to correctly model constrained paths of varying, arbitrary curvature [30]. We proposed a new model that integrates instantaneous curvature and width into its formulation, and validated it empirically for direct touch.