The Impact of Word, Multiple Word, and Sentence Input on Virtual Keyboard Decoding Performance

Abstract

Entering text on non-desktop computing devices is often done via an onscreen virtual keyboard. Input on such keyboards normally consists of a sequence of noisy tap events that specify some amount of text, most commonly a single word. But is single word-at-a-time entry the best choice? This paper compares user performance and recognition accuracy of wordat-a-time, phrase-at-a-time, and sentence-at-a-time text entry on a smartwatch keyboard. We evaluate the impact of differing amounts of input in both text copy and free composition tasks. We found providing input of an entire sentence significantly improved entry rates from 26 wpm to 32 wpm while keeping character error rates below 4%. In offline experiments with more processing power and memory, sentence input was recognized with a much lower 2.0% error rate. Our findings suggest virtual keyboards can enhance performance by encouraging users to provide more input per recognition event.

Comparing Smartphone Speech Recognition and Touchscreen Typing for Composition and Transcription

“…Our analysis demonstrates that transcription trials were, on average, faster, than composition trials. In comparison, Vertanen and Kristensson [28] did not find any difference between transcription and composition, Vertanen et al [24] found transcription was faster, and Yeo et al [31] and Karat et alfound composition was faster. Similar to Ruan et al, we found that speech had a lower Corrected Error Rate and a (slightly) higher Uncorrected Error Rate when both composition and transcription are combined, but not when only considering transcription.…”

“…They found no difference in text entry speed between the tasks, and only a modest difference in phrase length. Later studies by Yeo et al [31] and Vertanen, Fletcher, et al [24] used tasks from Vertanen and Kristensson to evaluate novel mobile text entry methods, on a smartphone and a smartwatch respectively. They found composition to be faster than transcription.…”

“…Most previous studies using composition tasks focus on text entry rate, like words-per-minute. Only four also report time measures [3,10,21,24] that may also include additional composition overhead for preparing to enter text. Since our interest is in this overhead, we further decompose trial times into measures like preparation time to better understand differences between transcribing and composing.…”

Mining, analyzing, and modeling text written on mobile devices

“…To add even more challenging data, we also force-aligned the data from Vertanen et al . (2018). In this study, participants typed sentences on a Sony SmartWatch 3.…”

Accelerating Text Communication via Abbreviated Sentence Input

“…Here we investigate accelerating input by allowing users to skip typing spaces and mid-word vowels. We decided to abbreviate in this manner based on past results on touchscreen text input without spaces (Vertanen et al, 2015(Vertanen et al, , 2018, and a study we present here in which 200 people abbreviated email messages. Our interaction approach of abbreviation is similar to features in commercial assistive interfaces (e.g.…”