PerVoice Brochure 2014 (English)


Embed or link this publication


PerVoice Brochure

Popular Pages

p. 1


p. 2

PERVOICE AUTOMATIC SPEECH RECOGNITION (ASR) You can get full control of your speech recognition processes with advanced ASR technologies and solutions from PerVoice. PerVoice Speech technologies provide excellent recognition results for many languages. The fundamental advantages of PerVoice ASR include: Very large vocabularies Speaker independence (requires no voice training) Spontaneous speech recognition Real-time and/or deferred recognition In addition, PerVoice Speech Technologies provide customizable and powerful solutions that let you: Add words on the fly Create industry specific topics Insert punctuation automatically Manage multiple audio channels Manage keyword lists THE LANGUAGES YOU NEED We currently provide recognition services in 13 languages: Italian, English (USA, UK, Australia), French, Spanish, Arabic (standard), Portuguese (Portugal and Brazil), German, Danish, Greek, and Flemish. We continue to work with our customers to develop new languages, converting their data into language solutions that fit specific needs.


p. 3

MARKET & SOLUTIONS PerVoice has four Business Units that focus on: To Broadcasting (Broadcast and Media monitoring companies). Reporting and Dictation (Courts, Government, HealthCare). Speech Analytics (Call Centers). Technology (EU and local projects). meet the needs of these customers, PerVoice has a number of targeted solution, including: FlyScribe and PASS (Pervoice AudioText Synchro Suite) for transcription services. PSW (Pervoice Subtitling Workstation) for the New Media sector. AudiomaBox and AudiomaRT for partners and customers that require only ASR technology. Find out more information about these products at RESEARCH AND DEVELOPMENT EU-Bridge - an EU funded project – will create a platform for handling subtitles, parliament reporting, transcriptions for university lectures and conference call language facilities. With partners from the most prestigious research centers in Europe and 10 million euros in funding, the project is advancing ASR and machine translation technologies. Wikivoice - funded by local government – aims to create a marketplace for automatic transcription based on user-created language models. COMPANY Pervoice S.p.A., a start-up of the Fondazione Bruno Kessler research center and other companies, was the first Italian service company to provide processing, analysis and exploitation of spoken language. Its main shareholder is Almawave, part of the Almaviva Group, Italy’s ICT leader. Other shareholders are private companies, private and public financial institutions, a research center and company management, all from the Trentino region. With a Company capital of €670,000 and 15 technical staff, PerVoice is rapidly growing in the speech recognition market. PerVoice sales are primarily international, with direct customers in Brazil, Denmark, Switzerland and the USA.


p. 4

BROADCASTING PerVoice provides solutions for traditional TV Media companies (subtitling and indexing for multimedia content) and for Media Monitoring customers (search for TV/radio content). For subtitling and close captioning, PerVoice offers a real time solution for live programs and a batch solution for pre-recorded programs. For real-time subtitling, PerVoice provides the PSW solution (PerVoice Subtitling Workstation). PSW consists of two touch screens that a pair of operators use to create high quality live subtitles/closed captions. PSW allows the operator to correct the automatic transcription from the TV source with touch gestures or to respeak what he/she is hearing, a technique used for passages where simultaneous speech, noise and use of dialect make automatic recognition difficult. Thanks to a second operator, the PSW achieves nearly 100% accuracy, remarkably high quality with respect to other solutions. The subtitling process for prerecorded programs is fully automatic. The only human intervention is in revising the computed transcription. For Media Monitoring market, PerVoice and its partners provide the following solutions: Automated transcription of radio/TV speech to enable keyword searches (Broadcast Monitoring); Automated creation of content indexes of human speech contained in an audio/video clip (Media Indexing). Screenshot of PSW software


p. 5

SPEECH ANALYTICS PerVoice provides two speech analysis solutions that help you manage Call Center quality. Iride-Voice®, developed through co-operation between PerVoice and its main shareholder Almawave, lets you check the quality of interaction between a call-center operator and a customer. Iride-Voice transcribes a phone call with ASR and – thanks to the interpretation of natural language – makes it possible to identify and classify customer needs quickly, easily and accurately by tracking and extracting the topics of interest. The system highlights phone calls containing relevant emotional contents and recognizes immediately and precisely points in a conversation with highly altered values. This type of systems is useful for incentivizing call center teams to improve CRM quality. Screenshot of Iride-Voice software


p. 6

REPORTING & DICTATION Speech recognition systems are often used for reporting activities in settings such as Law Courts or Parliaments. In this sector, PerVoice offers two solutions: Flyscribe is an ASR cloud service that lets professionals create text files from recorded speech. Today, Flyscribe is used by a wide variety of clients working in the transcription of speech both for public administration and conferences. The user uploads an audio file in a web browser and can follow the transcription process. A short time later, the user downloads the results. PASS (Pervoice AudioText Synchro Suite) is a suite of software tools developed around Flyscribe that lets you manage the entire transcription workflow. Users correct transcriptions in Microsoft Word® while listening to the synchronized audio file. The user can then store the synchronized audio and text permanently in PDF format. The solution was developed for courtroom hearings of the Italian Ministry of Justice. - ASR technology can also be used for dictation. Through Audioma-RT (Real Time), PerVoice provides advanced dictation technologies, including: Speaker independence (requires no voice training) Extensive vocabularies Management of spontaneous speech events Audioma-RT can be used in many sectors, including dictating medical reports, respeaking (an alternative to stenography) and transcribing conversations (for example, interrogations). With PASS you can certify human speech


p. 7

TECHNOLOGY & PROFESSIONAL SERVICES PerVoice is constantly working to improve its ASR technologies. These efforts include: Development and maintenance of our ASR products: o Audioma™ Box: a software platform for automated transcription of audio files in batch mode. It is used in several PerVoice solutions, from Broadcast Monitoring to Speech Analytics transcriptions. Audioma™ RT (Real Time): software for dictation. It is used for medical records, respeaking and “live” subtitling. o - Basic and applied research funded by local, national and international bodies. Realization of new transcription languages. A new language can be delivered rapidly with customer involvement. Analysis and realization of new acoustic and language models for special requirements. Support to partners and customers for analysis, integration and customization of technology as well as delivery of technical services. PerVoice can also provide professional services for transcription and subtitling of live and prerecorded TV programs. Example of audio spectrum analyzed by PerVoice technical services


p. 8

Any sufficiently advanced technology is indistinguishable from magic. > Arthur C. Clarke Pervoice SpA Viale Verona, 190/1 38123 Trento - Italy Tel. +39 0461 383960 Fax +39 0461 383910 VAT ID. 02019100227 Legal Reg. Trento n. 194584 Capital € 670.000 Company subject to management and coordination of Almaviva S.p.A.



no comments yet