PerVoice's Speech-To-Text engine has been developed by researchers at Fondazione Bruno Kessler (FBK). It comes in two of-the-shelf products:
Audioma Box, Speech-To-Text solution for specialized or technical vocabulary typically used in fields like broadcasting, media monitoring, call centers, health care and reporting;
Audioma Real Time, a dictation and respeaking solution that provides advanced dictation technologies, including large vocabularies and speaker independence.
Features
New language models can be created.
The PerVoice Speech-To-Text engine uses multicore parallel processing to make the recognition process fast and efficient.
Speaker independent, i.e. there is no need to train the system.
Add new words on-the-fly.
Language Coverage
Arabic
(Arabic),
English
(Latin - Australia),
English
(Latin - Caribbean),
English
(Latin - India),
English
(Latin - Ireland),
English
(Latin - United Kingdom),
English
(Latin - United States),
Danish
(Latin),
Dutch; Flemish
(Latin),
Persian
(Arabic),
French
(Latin),
French
(Latin - Canada),
French
(Latin - Switzerland),
Modern Greek (1453-)
(Greek),
Hindi
(Devanagari; Nagari),
Italian
(Latin),
Italian
(Latin - Switzerland),
Portuguese
(Latin - Portugal),
Portuguese
(Latin - Brazil),
Russian
(Cyrillic),
Spanish; Castilian
(Latin),
Spanish; Castilian
(Latin - Colombia),
Swedish
(Latin),
Turkish
(Latin),
Urdu
(Arabic),
German
(Latin),
German
(Latin - Austria),
German
(Latin - Switzerland)
