Beyond Ray-Casting: Evaluating Controller, Free-Hand, and Virtual-Touch Modalities for Immersive Text Entry

Abstract

Efficient text entry remains a primary bottleneck preventing Virtual Reality (VR) from evolving into a viable productivity platform. To address this, we conducted an empirical comparison of six physical input systems across three interaction styles Controller Driven, Free Hand, and Virtual Touch evaluating both discrete tap typing and continuous gesture typing (swiping), alongside a speech to text (Voice) condition as a non physical reference modality. Results from 21 participants show that the Controller Driven Tap Gesture Combo (CD TGC) delivers the best productivity performance, achieving speeds 2.25 times higher than the slowest system and 30% faster than the current industry standard, while reducing error rates by up to 68%. A clear trade off emerged between performance and perceived usability: although controller based gesture input led on speed and accuracy, participants rated Virtual Touch Tap Typing highest in subjective experience, scoring 80% higher on the System Usability Scale (SUS) than the lowest rated alternative. We further observe that Free Hand interaction remains limited by tracking stability and physical fatigue, whereas Voice input introduces practical constraints related to privacy, editing control, and immersive engagement. Together, these findings characterize the tension between throughput and natural interaction in immersive text entry and provide data driven guidance for future VR interface design.