VoiceInteraction will release the fourth generation of its Automatic Closed Captioning system (AUDIMUS.MEDIA) and the Broadcast Edition – devoted to helping broadcasters to comply with legal (and other) obligations and monitor competing channels programming – of its MMS at the NAB SHOW 2015, to be held in Las Vegas, NV, April 13-16, 2015.

AUDIMUS.MEDIA – Fourth Generation

Thanks to the capability of VoiceInteraction’s speech transcription technology to operate faster than real time with speaker-independent models, the company developed a  fully automated system to produce closed captions – in real time – for the (live or recorded) programmes broadcast (or webcast) on TV channels. The fourth generation of the system came up with several major improvements, such as the integration of deep-learning techniques to optimize the performance of acoustic models derived from the ever-growing speech repositories.
The community’s increasing awareness about the need for policies to ensure everyone equal opportunities and access to information and knowledge has led to the creation of rules imposing requirements on TV broadcasters to provide their contents in alternative ways, such as sign language and/or closed captions. Through the combination of innovative acoustic classification techniques with a modelling of the language structure – making use of  statistical language models – and a highly efficient search algorithm, VoiceInteraction created the Fourth Generation of AUDIMUS.MEDIA, aimed at addressing the broadcasters’ need for high-quality, cost-limited closed caption generation approaches raised by the aforesaid legal requirements.

MMS – Broadcast Edition

MMS is an integrated solution for video recording, storage and retrieval, comprising real-time tools for automatic transcription, segmentation and topic indexing of TV contents. In addition to the broad range of audiovisual media operations provided by MMS, that was first developed to automate the processing pipeline of clipping service providers, the Broadcast Edition comprises newly-developed modules to account for the specific needs of TV broadcasters and Government departments (regulatory agencies), such as the legal recording – with generation of loudness history graphs – for auditing purposes, and the monitoring and competitive analysis of rival channel contents, displayed on top of audience measurement data supplied by the customer.

Solutions Portfolio

In addition to the broadcast media solutions to be launched at the NAB SHOW, VoiceInteraction will use the event to showcase a broader range of proven speech processing solutions, such as Speech Transcription Systems for court, medical, parliament and council proceedings, TTS and ASR Engines for IVRs and Speech Analytics in Call Centres, and Embodied Conversational Agents for multi-modal applications.

Getting into the U.S. market and strengthening the market position in Brazil

According to Professor João Neto, co-founder and CEO of VoiceInteraction, the two novel solutions released at the NABSHOW will play a major role in the implementation of the company’s business strategy to start operating in the U.S. broadcast media market, in addition to strengthening its position in the Brazilian market. “We will use our participation in this event to launch a brand new generation of our automatic live subtitling system and a broadcast business oriented version of our media monitoring solution.” said Prof. João Neto. “Our R&D department developed a set of innovative approaches to acoustic modelling with a significant impact on the speech recognition performance that, in turn, improves our automatic live subtitling and media monitoring solutions. We will showcase our systems at the booth SU1621 and would be glad to have your opinion about them, so visit us and enjoy our demos.”

About the NAB SHOW

NAB SHOW is the world’s biggest annual conference and exhibition in the area of production, management and delivery of audiovisual contents. With more than 98,000 attendees from 150 countries and 1,700+ exhibitors, the event will be held at the Convention Center of Las Vegas, from April 13 to April 16, 2015, and is organized by National Association of Broadcasters.

About VoiceInteraction

VoiceInteraction was created in 2008 as a technological spin-off from the INESC-ID, Lisbon, providing cutting-edge speech recognition technology, such as an online speaker-independent speech transcription system that has led to the creation of fully automatic live subtitling systems for TV broadcasters and media monitoring solutions for clipping service providers,  in addition to speech synthesis technology for building customized voices to accommodate the branding needs of its clients.