In the early ’90s, Cedat 85 began actively working with IBM to conduct research into speaker dependent automatic voice recognition (re-speaking). As a result, it patented an entirely digital, direct reporting system that has eliminated the need for all the equipment and supports used in the past: stenotype machines, dictaphones, tape decks, audio cassettes, reels, etc.
Voice recognition is now an integral, preferential part of speech processing. The use of high-performance Speech To Text (STT) systems and the working method developed by Cedat 85 have led to the creation of a new professional role: the parliamentary speech-writer with voice recognition techniques. Cedat 85 has designed and provided technical/teaching support for the only training course in Italy, co-financed by the European Union, which lets you achieve the corresponding qualification recognized throughout the EU. The Chamber of Deputies deemed this to be the most modern and efficient method, so much so that it introduced the professional role of parliamentary speech-writer with voice recognition techniques, under comprehensive consultancy from Cedat 85, for House reporting from the beginning of the 14th parliamentary term (May 2001).
Furthermore, in 2005, Cedat 85 launched its research into speaker independent automatic language processing. This led to the development of natural language automatic recognition systems for Italian as well as applications which, thanks to automatic transcription, allow automatic indexing of audio and text, audio/video content searches, call center and help desk quality monitoring, and human-machine interaction.
Exploiting the potential of Cedat 85’s automatic voice transcription system, +VOCE is the most comprehensive and sophisticated system which can make content from any spoken multimedia sequence intelligible and immediately usable.
The transcribed text forms the basis for the accurate indexation of the entire audio and video (or just the audio), thereby letting you conduct a high-performance full text search.
This search guides you to the exact frame you searched for and to the exact point when the word, name or relevant subject was mentioned.
In short, you can now search multimedia content as though it were written text in all respects. Search methods are therefore changing: you no longer browse by file label or title, but use a new way of scanning and browsing around digital content so it can be used point-to-point thanks to a Point & Click system that links each word to the millisecond in the video when it was mentioned.
Cedat 85’s natural speech automatic processing system automatically decodes the audio and transcribes the content so it can then be indexed. The result is a user-friendly app that can make worthwhile savings in time.
For extensive video archives, +VOCE can generate a structure directly from the content index. This produces a dynamic interface that narrows down the content so users can easily browse through the archives (audio/video), even particularly large ones.
In fact, thanks to +VOCE’s capability to structure intrinsically unstructured content, you can skim through audio and videos, based on their actual content, by using ad hoc filters according to the needs of the client.
Videos on a particular subject, i.e. those from a specific period or of a certain duration set by the user, will then be viewable for video archives from major publishers; in turn, the user will be able to select the author, the broadcast time slot, the total listening figures and so on, all with a simple click of a button.
You can use the potential of this system either in “Full Outsourcing” mode or by implementing the entire hardware and software architecture of the “In-House” system. Whether you choose one solution or the other is purely down to the volume of audio and video documents you have to index and the availability of your staff to be trained how to use the system.