Method and device for providing speech-to-text encoding and telephony service
View Patent ↗A method and an apparatus for providing automated speech-to-text encoding and decoding for hearing-impaired persons. A broadband subscriber terminal interfaces to: (a) a network to convey speech packets thereover, (b) a telephone to convey speech information, and (c) a display device to display textual information of spoken words. A speech buffer in the subscriber terminal receives speech data and a processor decodes and displays textual representations of speech on the display device. A database stores voice and/or speech patterns that are used by a speech analyzer to recognize an incoming caller and to associate a name or characteristic (e.g., male or female) with the incoming call. A tonal and inflection analyzer analyzes speech to add punctuation to the displayed text. A detector, such as a DTMF detector, responds to subscriber inputs to activate/deactivate speech recognition or other functions.
1. A device for use in a network comprising:
a modem that connects with the network to convey information to, and receive information from the network;
a subscriber terminal having an interface that enables communication with the modem, a display interface that communicates with a visual display device to display information, a telephone interface that enables communication with a telephone to convey voice information of a user, and a buffer that receives and stores speech information; and
a processor to decode and display on the display device speech information as text in a form of words upon receipt of speech information from the network, wherein said processor includes a detector that responds to subscriber inputs in a form of at least one DTMF tone to activate and deactivate speech recognition.
2. The device as recited in claim 1 , wherein the subscriber terminal further includes:
a memory that stores voice patterns, and wherein
said processor further includes a speech analyzer that recognizes an incoming voice pattern based on information stored in the memory.
3. The device as recited in claim 1 , wherein said subscriber terminal includes a speech database for storing speech segments identified with certain users, and said processor accesses said database to identify and display an identity of users according to matches between speech segments received in real time and stored in the database.
4. The device as recited in claim 1 , wherein said detector comprises a DTMF tone detector and said subscriber inputs comprise DTMF tones of a telephone.
5. A method of providing automated speech-to-text translation, the method comprising:
receiving at a broadband telephony interface speech packets destined for an individual;
storing the speech packets in a buffer;
processing the speech packets to display textual representations thereof as words on a display device; and
responding to a command, in a form of at least one DTMF tone, from a subscriber to activate and deactivate speech processing.
6. The method as recited in claim 5 , further comprising:
storing speech patterns in a database, and
analyzing and comparing incoming speech obtained by processing the speech packets with speech patterns stored in the database in order to provide speaker identification capability.