Voice Recognition Technology Comes of Age with Tantiv4’s ‘ThingaOS™: A Case Study

Voice Recognition Technology Comes of Age with Tantiv4’s ‘ThingaOS™: A Case Study

(Overcoming many difficult challenges in building a top-end SaaS IoT Platform for Sharp’s Blue-ray Disks Recorders in Japan)


Entertainment systems all over the world are continuously evolving and changing at a breath-taking pace. Today, new machines, projectors, and TV sets are  rapidly replacing old ones.


The Problem

However, each individual platform has its very own particular needs. Let us take the example of the interface present on a BDR (blue ray disk recorder). Your average BDR has highly specific requirements such as the activation and deactivation of the ‘like, play, pause and stop’ functions. These functions and their attendant interface are unique to the individual BDR platforms and models.


SHARP a consumer electronics company launched a whole new series of Blu-ray recorders for 2K, HD, Ultra HD Blu-ray™ playback. Sharp and Tantiv4 partnered together to bring several advanced features such as voice controls functions of their 2019 BDR models.

2B-C20BT3 and 2B-C10BT3 BD Recorders

However, they also wanted to make sure that there are no excessive ripples in the waters. This way, they can launch all-new ‘IoT voice capability’ enabled products without destabilizing their current highly successful product line up.


This is a major challenge that many consumer electronics companies are facing today. They want to keep the product experience intact, at least in terms of features and functionality. But at the same time, they also want to boost the capabilities of their cutting-edge products with the help of voice recognition controls that are powered by the IoT (Internet of Things). This way, they will save the product development and marketing expense, not to mention the associated risk of introducing a completely new product in the market.


The ‘Project had two key objectives with regard to the overall direction of the plan.


voice control via Google and Amazon

Implementation of a Modular Voice Recognition Solution

The main objective was to add voice control (via Google and Amazon) for the whole range of commands that might be given to the recording device with the help of its remote-control unit. We, at Tantiv4, have developed a solution, which involves multiple ways through which the user can invoke the desired command.


Addition of a Voice Operated ‘Quick Access’ Option

Our secondary objective was to enable ‘smart voice’ commands with the help of which the user could access certain options and also trigger various actions. Here, he would be able to activate functions that would otherwise take a long time to access the remote control. Some examples include the following:

  • Searching and playing specific content from the recorded list or the electronic program guide (EPG).
  • The ability to skip a few seconds or minutes in a video with the help of a verbal command
  • Search content based on genre
  • Playing a title from the beginning and many other scenarios of a similar nature.

voice control via Google and Amazon

It was Tantiv4’s core responsibility to:

  • Develop a highly advanced voice module which would be able to understand the various verbal commands given by the user
  • Match this module with the existing database of movies, programs, genres
  • Present the most accurate result by selecting the desired option for the end user

A few common issues that the Tantiv 4 voice solution helped to resolve include

A few common issues that the Tantiv 4 voice solution helped to resolve

DVR Function Problem and Its Solution

Suppose your remote-control unit is lost and there is no other way to operate the unit. However, with the help of the Tantiv4 voice recognition solution, it is possible to ‘talk’ to the speaker directly and conduct many of the more basic operations. For example, you can turn the device on and off, change channels and execute many other commands without using the device’s remote control unit.


Searching Through Reams of Heavy Data

Searching Through Reams of Heavy Data

Suppose you want to watch a particular video file or movie, but your device contains over five terabytes of data. Here, you will be forced to use the ‘up and down’ keys of the remote in order to search for the required file. This can be a laborious and time-consuming job when there are literally thousands of files to choose from. But using a voice-controlled ‘smart speaker’ set up means that you have to simply say the name of the movie out loud and the speaker will do the rest.


Understanding the Different Dialects

A very real problem is the translation of the voice into an understandable dialect, i.e., one that the machine can actually recognize well enough to take the desired action. The technology has to be so advanced that it can interpret subtle nuances of the speaker’s dialect.


In the case of the Japanese language, we, at Tantiv 4, had to work with five different dialects of the same language. In fact, even the most powerful search engine generally mixed the different dialects and seemed incapable of producing the right results with any degree of consistency.

Japan Map

There are more than a hundred regions in Japan, and they all have their own EPGs so that it is almost impossible to acquire 100 percent reliable information. Since mixed results are not acceptable, we created and perfected a tailor-made SAAS (software-as-a-service) platform for an effective solution to these Blue-ray Disc recording problems.


Making the Whole Desired Program Search Experience a Breeze

It is not easy to flip channels or search for a particular program or channel with the help of a remote-control unit alone, especially if you are searching amongst thousands of titles and channels, simultaneously.


We, at Tantiv4, were able to successfully develop a highly advanced voice recognition module. The main feature of this voice recognition solution is that it is able to seamlessly understand virtually all of the different dialects and accents in order to present the most accurate results.


Conclusion

Thanks to Tantiv 4’s ThingaOS™, you don’t have to worry about wasting time on tediously long and inefficient searches.


Our IoT solution uses highly advanced machine learning and artificial intelligence tools to consistently provide accurate search results that are also able to adapt to the individual user’s needs.