Speech-processing system and method

• US 7,286,989 B1
• Filed: 09/02/1997
• Issued: 10/23/2007
• Est. Priority Date: 09/03/1996
• Status: Expired due to Fees

• Current Assignee: Siemens AG
• Original Assignee: Siemens AG
• Inventors: Harengel, Steffen, Murr, Florian, Niedermair, Gerhard, Karsten, Peter
• Primary Examiner(s): Opsasnick, Michael N
• Application Number:
  US09/254,242
• Time in Patent Office: 3,703 Days
• Field of Search: 704/254, 704/231, 704/251, 704/270, 704/275, 379/88.01, 379/88.02
• US Class Current: 704/275
• CPC Class Codes: G10L 13/00 Speech synthesis; Text to speech systems
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence

Abstract

A speech processing system has an arbitrary number of speech recognition modules (Ei, i=1 . . . n) and speech output modules (Aj, j−1 . . . m). The modules provided respectively for a particular type of speech recognition or, respectively, speech output are selected and activated and parameterized according to the situation of application, dependent on an input signal (ES), by a device (MA) for selecting the modules, so that the selected modules are applied for speech recognition or, respectively, speech output corresponding to their construction.

Claims

1. A speech processing system comprising:

a plurality of modules of respectively different types, said plurality of modules comprising a plurality of speech recognition modules for respectively different types of speech recognition and a plurality of speech output modules for respectively different types of speech output, each speech output module being capable of outputting oral speech, each of the speech recognition modules and each of the speech output modules being respectively constructed specifically for a different type of speech recognition or speech output; and

a selector supplied with an input signal containing information identifying a type needed to process said input signal, said type relating to type of speech recognition and said type of speech output needed to process said input signal, said selector being connected to said plurality of modules and routing said input signal to one of said modules while excluding said input signal from another of said modules, the selector routing said signal based on the type identified in said input signal for processing said input signal only in said one of said modules to which said input signal is routed.

2. A speech processing system as claimed in claim 1 wherein said selector controls processing of said input signal in said one of said modules to which said input signal is routed.

3. A speech processing system as claimed in claim 1 wherein a group of said speech recognition modules in said plurality of speech recognition modules share common pre-processing features, and wherein said plurality of modules further includes a pre-processing module, connected to said speech recognition modules in said group, and wherein said selector, if said at of said modules to which said input signal is routed is a speech recognition module in said group, routes said input signal to said one of said speech recognition modules in said group through said pre-processing module.

4. A speech processing system as claimed in claim 1 wherein a group of said speech output module in said plurality of speech output modules share common post-processing features, and wherein said plurality of modules further includes a post-processing module, connected to said speech output modules in said group, and wherein said selector, if said one of said modules to which said input signal is routed is a speech output module in said group, routes said input signal through said post-processing module.

5. A speech processing system as claimed in claim 1 wherein said plurality of speech recognition modules comprise a speech recognition module for individual numerical recognition, a speech recognition module for recognition of chains of numerals, a speech recognition module for recognition of words from a limited vocabulary, a speech recognition module for recognition of an individual word with an unlimited vocabulary, a speech recognition of an individual word with an unlimited vocabulary, a speech recognition of speech spoken in a flowing manner with an unlimited vocabulary, a speech recognition module for recognition of predetermined word combinations, a speech recognition module for key word recognition, a speech recognition module for alphabet recognition, a speech recognition module for sound sequence recognition, a speech recognition module for speech recognition module for speech recognition, and a speech recognition module for DTMF recognition.

6. A speech processing system as claimed in claim 1 wherein said plurality of speech output modules comprise a speech output module for output of predetermined stored speech components, a speech output module for output of predetermined stored speech components, a speech output module for output of combined individual predetermined stored speech components, a speech output module for output of words synthesized from stored phone names, and a speech output module for output of DTMF tones.

7. A speech processing system as claimed in claim 1 further comprising a dialog sequence control connected to said selector for controlling a dialog between said selector and a user.

8. A speech processing system as claimed in claim 7 wherein said dialog sequence control generates said input signal.

9. A speech processing method comprising:

providing a plurality of modules of respectively different types including speech recognition modules for respectively different types of speech recognition and speech output modules for respectively different types of speech output, each of the speech recognition modules and each of the speech output modules being respectively constructed specifically for a different type of speech recognition or speech output;

analyzing an input signal to be processed to identify a type of speech recognition and a type of speech output needed to communicate with a user;

selecting one of the speech recognition modules and one of the speech output modules according to the analysis of the input signal;

routing said input signal to one of said modules according to the analysis of the input signal;

recognizing speech from the user using only the selected speech recognition module; and

outputting speech to the user using only the selected speech output module such that two-way oral communication is performed with the user using the selected speech recognition module and the selected speech output module.

10. A speech processing method as claimed in claim 9 comprising the additional step of:

If said input signal is routed to at least one of said speech recognition modules in said plurality of modules, pre-processing said input signal before processing said input signal in said at least one of said speech recognition modules.

11. A speech processing method as claimed in claim 9 comprising the additional step of:

if said input signal is routed to at least one of said speech output modules in said plurality of modules, post-processing said input signal after processing said input signal in said at least one of said speech output modules.