TechnologiesPerVoice always looks to the future by focusing on the continuous improvement of technological excellence in the voice recognition field
PerVoice transcription solution transform the human voice into audio-synchronized documents quickly and effectively ensuring the best state-of-the-art accuracy.
Over the years PerVoice has consolidated its international competitiveness thanks to a complete set of modular and multilingual solutions, developed for different markets and application contexts.
PerVoice voice transcription is available as a ready-to-use solution (all-in-one platform), client-server system to be integrated on third-party platforms, via API to allow a wider integration spectrum.
In collaboration with Bruno Kessler Foundation, PerVoice has implemented translation models based on the most innovative neural architectures of automatic translation, integrated with its own voice recognition systems.
Billions of words training guarantees an excellent quality of translation of PerVoice vocal transcriptions, regardless of the topic covered and the reference domain.
PerVoice Machine Translation solution is available with integration in Audioma® and FlyScribe®, or via REST API.
ASR for IVR
PerVoice‘s voice recognition technology enables Contact Center IVR to manage automatic customer-company interaction on the voice channel.
The spontaneous language on phone calls is transcribed in real time, activating the tasks for analysis and interpretation of the caller’s request, in order to efficiently process and resolve it.
PerVoice‘s technology for IVR offers the opportunity to implement innovative self-caring applications, with On Premises and Cloud installations.
WHY CHOOSE PERVOICE’S ASR TEchnology for ivr?
Recognition precision and state-of-the-art acoustic models
Easy integration with NLU and Chatbot engines
Command and control
PerVoice voice command and control solution dedicated to enterprises, is based on TensorFlow neural frameworks which, among the various advantages, allow on-device operation.
PerVoice uses neural network adaptation techniques that allow, with little data in the domain, to create customized command and control systems.
The areas of application and interest are always wider; include, for example, industrial machinery, tools for the disabled, warehouse picking operations, …
In general, PerVoice command and control solution can be used in all contexts where user experience is remodelled and improved through the voice digitization.
PerVoice‘s consolidated experience in subtitling is the result of the numerous use cases implemented, mixing business and social inclusion needs.
Accessibility to information for people with hearing impairments, the best video content usability on mobility, audio translation into subtitle, customer’s production systems integration, are just a few examples of PerVoice solutions use.
PerVoice subtitling solutions are available through PSW (real time mode) or integrated in FlyScribe® and via REST API (batch mode).
Languages and models
To face the needs of an increasingly global market, over the years PerVoice has constantly increased the number of supported languages and specific language models of the speech domain (News, Parliamentarian, Sport, Finance, etc.)
To date, there are over 30 languages available and over 50 language models, developed to support types of wideband (radio/TV/conferences) and narrowband (phone) audio.
PerVoice linguistic models are distinguished by Speaker Independent technology which does not require any speaker training and include features for managing accents and pronunciation variations: e.g. British, American, Australian and Indian English.
The methodology used by PerVoice for the preparation of the linguistic models, it can be quickly extended to any new language.
PerVoice Automatic Speech Recognition technology is able to recognize professional jargon and technical terms, as well as toponyms, acronyms, proper names, acronyms, etc.