Text Control and The Rise in Voice Recognition

Voice recognition has aggressively proliferated over the last few years, but one thing keeps it from being the preferred mainstream choice for input | May 4, 2019

Srivats Shankar (@SrivatsShankar)

Voice recognition software has become more common than ever. Everything from smart phones to applications have some amount of integrated voice recognition capabilities. Many people were first exposed to the tech with the introduction of Siri, which made voice recognition more accessible than ever before. Yet, they continued to remain evasive of mainstream user interaction. People continue to choose the keyboard over voice recognition – why is that?

Although, perceptions of accuracy (that are mostly misplaced), comfort in usage, and awareness have a significant contribution to this end, there is another factor that placed significantly into the situation – text control. Users of Dragon NaturallySpeaking would already familiar with this. To the uninitiated, text control means having some method through which without interacting with your mouse or keyboard, a user can simply tell voice recognition to correct a mistake and it offers intelligible choices on how to rectify mistakes.

Wait a second? That sounds really confusing. The easiest way of understanding text control is by offering an example. For example, say that you enunciate "higher", but you actually meant to say "hire" (in the hypothetical situation that you are planning on creating a job posting for your company). Normally, voice recognition would require you to either pick up your keyboard and clear the mistake and re-type every single word. Well text control would cover you in that regard. You can simply give a command and the software will give you better suggestions on how to fix what you just dictated.

 

Full Text Control Example

 

Out of all the software we tried, only a handful support some degree of text control. For voice recognition to go mainstream it really needs to support fulltext control, otherwise it is just another nicety. Developers should consider and integrate voice recognition to meet daily requirements. Transcription itself is not always the solution. Undoubtedly, the ability for a system to process the transcription forms the basis of any voice recognition application, but this forms the next logical step.

Technology

Voice Recognition