Patent US405270A successfully refreshed from Google Patents.
Patent US2701271A successfully refreshed from Google Patents.
US4423291A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Method for operating a speech recognition device
Assignee: Siemens AG
•
Filing: 1981-02-04
•
Granted: 1983-12-27
•
Expires: 2001-02-04
Abstract
Abstract Speech recognition is improved by time-standardization of reference and unknown words of constant bit-length formed by suppressing bits or generating copy bits.
Inventors
- 👤 Eberhard Zwicker
- 👤 Wolfgang Daxer
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-10-28 14:02
CPC Classifications
G; G10; G10L; G10L15/00
Claims Analysis
Total Claims: 12
Claim 1
Independent Claim
183 words
A method for operating a speech recognition device in which a plurality of rectified channel signals of different frequencies are derived from the speech signal of a spoken word, said channel signals being periodically sampled and digitized to produce digital signals which are stored in a digital memory and compared with a plurality of stored groups of corresponding comparison signals, and in which the group of comparison signals producing the best coincidence is employed as the recognition result, comprising the steps of modifying one of said digital signals having a channel-time pattern (P11 . . . Pmn) by suppressing bits belonging to a prescribed group of sampled values or storing multiple copies of bits belonging to said prescribed group, said group being prescribed as a function of the overall number (m) of sampling operations which take place within the acoustical word duration, so that the modified channel-time pattern (P11' . . . Pnm') correspond to a standardized plurality (p) of sampling operations, and comparing the modified channel-time pattern (P11' . . . Pnp') with the stored groups of corresponding patterns of comparison signals.
Claim 2
Independent Claim
183 words
A method for operating a speech recognition device in which a plurality of rectified channel signals of different frequencies are derived from the speech signal of a spoken word, said channel signals being periodically sampled and digitized to produce digital signals which are stored in a digital memory and compared with a plurality of stored groups of corresponding comparison signals, and in which the group of comparison signals producing the best coincidence is employed as the recognition result, comprising the steps of modifying one of said digital signals having a channel-time pattern (P11 . . . Pmn) by suppressing bits belonging to a prescribed group of sampled values or storing multiple copies of bits belonging to said prescribed group, said group being prescribed as a function of the overall number (m) of sampling operations which take place within the acoustical word duration, so that the modified channel-time pattern (P11' . . . Pnm') correspond to a standardized plurality (p) of sampling operations, and comparing the modified channel-time pattern (P11' . . . Pnp') with the stored groups of corresponding patterns of comparison signals.
Claim 3
Dependent Claim
24 words
The method according to claim 1, including the step of logarithmizing said channel signals and filtering said channel signals through individual low pass filters.
Claim 4
Dependent Claim
19 words
The method according to claim 1, including the step of digitizing said sampled channel signals into 8 bit words.
Claim 5
Dependent Claim
63 words
The method according to claim 1, including the step of comparing the modified channel-time pattern (P11' . . . Pp') to the corresponding stored patterns of comparison signals by determining the difference between corresponding signals and summing all of such differences, to produce a comparison signal for each group, and employing the comparison signal allocated to the lowest sum as the recognition result.
Claim 6
Dependent Claim
35 words
The method according to claim 1, including the step of comparing the level of the signal in one of said channels to a theshold value for determining the beginning and end of a spoken word.
Claim 7
Dependent Claim
32 words
The method according to claim 1, including the step of producing an acoustical output of the word corresponding to the one of said groups having the best comparison with said digital signals.
Claim 8
Dependent Claim
21 words
The method according to claim 1, including the step of deriving control signals for electrical switches in response to spoken words.
Claim 9
Dependent Claim
18 words
The method according to claim 7, including the step of using said switching signals for moving artificial limbs.
Claim 10
Dependent Claim
227 words
A speech recognition device comprising in combination, a plurality of band pass filters connected to the output of a microphone for passing a plurality of frequency bands, a plurality of logarithmization stages individually connected to the outputs of said filters, a plurality of rectifier and low pass stages individually connected to said logarithmization stages, a sampling device having a plurality of inputs connected to the outputs of said rectifier and low pass stages, for sequentially selecting said channels in sequence, an analog to digital converter connected to the output of said sampling device, a microcomputer connected to said analog to digital converter, said microcomputer having a storage device for storing digital signals produced by said sampling device during the duration of a spoken word, means for transcribing said digital signals from one location to another in said storage device by omitting preselected bits of said digital signals or multiply recording preselected bits of said digital signals, means for comparing said transcribed digital signals with a plurality of groups of stored digital signals and for producing an output representative of the group of storage signals which best compares with said digital signals, a speech synthesis device connected to receive said signals for producing an audible output of a word corresponding to a stored group comparing favorably with said digital signals, and a loudspeaker connected to said synthesis device.
Claim 11
Dependent Claim
227 words
A speech recognition device comprising in combination, a plurality of band pass filters connected to the output of a microphone for passing a plurality of frequency bands, a plurality of logarithmization stages individually connected to the outputs of said filters, a plurality of rectifier and low pass stages individually connected to said logarithmization stages, a sampling device having a plurality of inputs connected to the outputs of said rectifier and low pass stages, for sequentially selecting said channels in sequence, an analog to digital converter connected to the output of said sampling device, a microcomputer connected to said analog to digital converter, said microcomputer having a storage device for storing digital signals produced by said sampling device during the duration of a spoken word, means for transcribing said digital signals from one location to another in said storage device by omitting preselected bits of said digital signals or multiply recording preselected bits of said digital signals, means for comparing said transcribed digital signals with a plurality of groups of stored digital signals and for producing an output representative of the group of storage signals which best compares with said digital signals, a speech synthesis device connected to receive said signals for producing an audible output of a word corresponding to a stored group comparing favorably with said digital signals, and a loudspeaker connected to said synthesis device.
Claim 12
Dependent Claim
55 words
The speech recognition device according to claim 9, including a further logarithmization stage connected in parallel width said band pass filters, a further rectifier and low pass stage connected to the output of said further logarithmization stage, and means for connecting said further rectifier and low pass stage to an input of said sampling device.
References Cited (Backward Citations)
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 1986-12-26 | 1987-06-27 | 1987-12-27 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 1990-12-26 | 1991-06-27 | 1991-12-27 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 1994-12-26 | 1995-06-27 | 1995-12-27 | $7700.00 | Window Closed / Lapsed |