A voice recognition display system that visualizes speech in real time, developed and sold by JVC Kenwood Public Industrial Systems.
We spoke to Mei Sato of the Audio Solutions Promotion Department, Marketing Division, JVC Kenwood Public Industrial Systems Corporation, the company that provides the service, about how they adopted AmiVoice and the benefits it has had.

■ Voice recognition display software product page: https://jkpi.jvckenwood.com/solution/stt_display_software/

Background and Challenges in Using Speech Recognition

Promoting efforts toward an "open parliament" by displaying voices in text

Our company has been developing, manufacturing and selling audiovisual systems that support parliamentary operations for many years, but we have been exploring ways to provide added value in terms of information security and efforts to address the recent demand for "open parliaments." As an audio manufacturer, it is natural for us to "deliver sound clearly," but we wondered if we could deliver information to more people by using voice recognition technology to "display voice as text."

There are many voice recognition engines available these days, so we compared them, but the deciding factors were the high recognition rate and the fact that they provided an engine that runs locally without using the cloud. Our software is not designed for post-editing, so real-time performance and accuracy are required, and we think that such high recognition accuracy in real time is amazing.

Usage purpose

Making parliamentary statements visible and promoting universal design

We provide speech recognition display software (TZ-TRACER) that can display the content of statements made in local assemblies and other events in text on monitor screens in the auditorium, and can also display subtitles on video footage of speakers broadcast over the Internet or within the city government.

With this software and minimal video and audio equipment, two types of screen display, suitable for "real" and "remote" views, can be achieved, making it possible to implement at low cost. By visualizing what is being said in parliament, it supports the promotion of universal design, such as support for the hearing impaired in parliamentary operations.

Benefits

Local operation type with fast display speed and no impact from communication failures

With support from Advanced Media from planning to launch, we were able to provide the service to our customers quickly. The local governments that have adopted the system have praised it for its accuracy in character conversion and its local operation. The display speed is fast, and there is no need to worry about communication failures.
We have also received positive feedback, such as, "Now we can check the audio in real time and anyone can join in the hearing at any time."

Future use

I want to contribute to the realization of a more universal society.

The "voice recognition display software" we have developed can be used in applications other than parliamentary meetings, so we hope to expand it to a variety of markets. Furthermore, by linking it with our audio products and audio-related technologies, we hope to provide the experience of "being able to convert sounds into text and understand them," thereby contributing to the realization of a more universal society.