Search engine giant Google has announced a new Android app, named Live Transcribe that is designed to make real-world conversations more accessible for hard of hearing people. Powered by Google Cloud, Live Transcribe automatically captions conversations in real-time as well as it supports over 70 languages and more than 80 percent of the world’s population.
Google’s new Android app leverages the outcomes of extensive user experience (UX) research with sustainable connectivity to speech processing servers. To make certain that connectivity to these servers doesn’t cause excessive data usage, the team utilized Cloud ASR (Automated Speech Recognition) for higher precision. Also, to lessen the network data utilization needed by Live Transcribe, an on-device neural network-based speech detector was deployed. The on-device neural network-based speech detector is developed utilizing Google’s dataset for audio event research, named AudioSet that was introduced last year. It is an image-like model which is able to detect speech, automatically managing network connections to the Cloud ASR engine, and minimizing data usage over long terms of utilization.
To make Live Transcribe intuitively, the team at Google collaborated with Gallaudet University, with the help of user experience research collaborations. The search engine giant considered an array of devices ranging from smartphones, tablets, computers, and small projectors, and so on, to efficiently show auditory information and captions. However, after a thorough study, Google decided to select smartphones because of its sheer ubiquity and improved capabilities. The tech giant cited that they faced a challenge regarding displaying transcription confidence while developing Live Transcribe. Using earlier UX research, researchers have found that a transcript is easiest to read when it is not layered and concentrates on the superior presentation of the text, so supplementing it with other auditory signals apart from speech signals.