Chapter 6: Testing the benchmark with two immersive displays
Continuing from Chapter 6, the benchmark application is used for another set of trials and its results are presented in this chapter. Having lessons learned of the previous trial, a slight modified setup was used to allow for better observation and evaluation of user behaviour and performance. A detailed description is given in the following sections as well as a reflection on the motivation for the experiment.
Hypothesis-2: Immersive CVEs could support a seamless flow of collaboration and result in a performance efficiency increase
Previously in Chapter 6, user trials were mainly conducted between asymmetric (immersive vs. non-immersive) display setups. It demonstrated that immersive users take a leadership role and that desktop users struggle with their interface during complex tasks such as moving and positioning an object in the 3D environment. This has an impact on the workflow and performance of a closely-coupled task. The results also showed that two immersive users outperformed a third connected desktop user. Therefore the hypothesis is that users connected with only immersive (symmetric) displays can work seamlessly and increase performance of a closely-coupled task. To test this hypothesis a new user trial was conducted which is now described, assessed and discussed.
Discussion
For the benefit of better user evaluation the task had been modified for the CAVE-CAVE trial compared to the trial discussed in chapter 5. In addition, all users went through a 15min practice session to familiarise with interface and application. This makes it difficult to directly compare the time performance of both setups. However, if users were not hampered by their understanding of the application or usage of interface, the performance was comparable to expert users. If interaction was hampered it came mainly from difficulties while grasping objects, which in turn had different causes such as loosing 3D vision when shutter-glasses lost the infrared signal, using the wrong button on the joypad or misjudging the distance to an object. Other problems can arise from the tracking system where cables could limit free movements within the CAVE-like display. Such difficulties with the interface can also have significantly negative impact on user’s perception of performance and collaboration. This was in particular a problem for users of the Salford IPT where a “richer” yet “more constraining” tracking system (due to more cables) had a greater negative impact than for their Reading collaborators.
Compared to the trials of Chapter 5 observations could be made that showed a seamless flow of interaction and communication. This increased the perception of collaborative performance as questionnaire results show equally high or even higher values as Chapter 6 results. Yet performance differences were observed and categorised into fast – average – slow sessions. The main reason for slower sessions were due to problems of understanding the task sequence or difficulties with the interface in grasping objects or navigating. This resulted in giving significant different opinions during the questionnaire for contribution, performance and collaboration. Nevertheless, all participants had a high perception of social activity and presence in the environment as well as with others. In addition, transcripts show that at the beginning and end people focus on social interaction, while interaction via technical interfaces and the virtual representations play a more important role for task focused collaboration.
Interestingly, a significant difference could be found in the perception of emotions where negative emotions from slow session users were better perceived than positive emotions from users of faster sessions. In addition, as important as verbal communication is for closely-coupled collaboration the use of non-verbal communication should not be underestimated. Although non-verbal support by the IPT was limited and participants did not judge its contribution very high, observations show that gestures such as pointing, gazing, etc. were frequently used to support verbal cues like agreeing or moaning. Furthermore, observations show that participants respect each others personal space (see Chapter 2) by avoiding avatar collisions.
The previous chapter 5 concluded that immersive users take a leading role in closely-coupled collaboration and that they outperform desktop users within the same virtual environment. The workflow was interrupted by application and interface problems as well as orientation issues of desktop users. This raised the question of what would happen in a pure immersive setup. The hypothesis was that immersive CVEs could support a seamless flow of collaboration and could result in a performance increase. The results presented in this chapter can support this hypothesis.
Summary
In conclusion it can be said that maintaining the flow of collaboration is likely to be important in supporting group interaction and that immersive CVEs support a flow of object focussed conversation that is hard to distinguish from the real world, however object manipulation, at least without haptics and the current IPT interface, presently interferes with the flow of the task. However, users interact with objects in a spatially natural way with important aspects of their body movements represented remotely with respect to the object. Immersive CVEs appear promising for the support of distributed group collaboration and seem to be suitable for studying people’s behaviour during such collaboration. Yet how much does the interface compared to collaboration contribute to the success or failure of a closely-coupled task? To answer this question some aspects have to be singled out to increase the focus and the next chapter will next chapter will try to address some of these issues.