EP0140249A1
Expired / Lapsed
EP
⏱️ 0.0 years remaining
Speech analysis/synthesis with energy normalization
Assignee: Texas Instruments Inc
•
Filing: 1984-10-12
•
Granted: —
•
Expires: 2004-10-12
Abstract
Abstract Energy normalization in speech synthesis systems is achieved by a look-ahead adaptive normalization procedure, wherein energy is adaptively tracked, and the adaptive energy-tracking value is used to normalize a much earlier frame's energy value. In another aspect, silence suppression in speech synthesis systems is achieved by detecting and processing only segments of voice activity. A segment is classified as "speech" if the energy of the signal is greater than an adaptively adjusted threshold. The adaptively adjusted threshold is preferably defined as the maximum of scaled values of two separate envelope parameters, which both track the variation in energy over the sequence of frames of speech data. One contour is a slow-rising fast-falling value, which is updated only during unvoiced speech frames, and therefore tracks a lower envelope of the energy contour. This parameter in effect tracks an ambient noise level. The other parameter is a fast-rising slow-falling parameter, which is updated only during voiced speech frames, and thus tracks an upper envelope of the energy contour. (This in effect tracks the average speech level.)
Inventors
- 👤 George R. Doddington
- 👤 Panos E. Papamichalis
Bibliographic Details
Application Number
—
Jurisdiction
EP
Data Source
google_patents
Last Synchronized
2026-11-02 14:35
CPC Classifications
G; G10; G10L; G10L25/00; G10L25/78; G10L21/00; G10L21/02; G10L21/0316; G10L21/0364; G10L2025/783; G10L2025/786; G10L21/0208; G10L21/0216; G10L21/0232
Claims Analysis
Total Claims: 25
Claim 1
Independent Claim
80 words
A speech coding system, comprising: An analyzer connected to receive a digital speech signal and to generate therefrom a sequence of frames of speech parameters, said parameters of each frame including an energy value, and means for normalizing the energy value of each said speech frame with respect to energy values of subsequent frames; and output means for loading said parameters for each said speech frame including said normalized energy parameter of each said speech frame into a data channel.
Claim 2
Dependent Claim
80 words
A speech coding system, comprising: An analyzer connected to receive a digital speech signal and to generate therefrom a sequence of frames of speech parameters, said parameters of each frame including an energy value, and means for normalizing the energy value of each said speech frame with respect to energy values of subsequent frames; and output means for loading said parameters for each said speech frame including said normalized energy parameter of each said speech frame into a data channel.
Claim 3
Dependent Claim
158 words
A voice mail system, comprising: An analyzer connected to receive a digital speech signal and to generate therefrom a sequence of frames of speech parameters, said parameters at each frame including an energy value; means for normalizing the energy value of each said speech frame with respect to energy values of subsequent frames; output means for loading said parameters for each said speech frame including said normalized energy parameter of each said speech frame into a data channel; input means for receivinq a sequence of frames of speech parameters, said parameters including linear predictive coding parameters and excitation parameters; means for configuring a lattice filter in accordance with said linear predictive coding parameters; means for generating an excitation signal in accordance with said excitation parameters, said excitation signal being provided as input to said lattice filter; and means for modulating the output of said lattice filter in accordance with an energy parameter, to provide a speech signal output.
Claim 4
Dependent Claim
158 words
A voice mail system, comprising: An analyzer connected to receive a digital speech signal and to generate therefrom a sequence of frames of speech parameters, said parameters at each frame including an energy value; means for normalizing the energy value of each said speech frame with respect to energy values of subsequent frames; output means for loading said parameters for each said speech frame including said normalized energy parameter of each said speech frame into a data channel; input means for receivinq a sequence of frames of speech parameters, said parameters including linear predictive coding parameters and excitation parameters; means for configuring a lattice filter in accordance with said linear predictive coding parameters; means for generating an excitation signal in accordance with said excitation parameters, said excitation signal being provided as input to said lattice filter; and means for modulating the output of said lattice filter in accordance with an energy parameter, to provide a speech signal output.
Claim 5
Independent Claim
44 words
The system of either of Claims 1 or 2, wherein said energy value of each said speech frame is normalized with respect to said energy values primarily of those of said frames which are later than said respective frame by at least 0.1 second.
Claim 6
Independent Claim
47 words
The system of any of Claims 1-3, wherein said energy value of each said speech frame is normalized with respect to a peak-tracking parameter of said subsequent frames, said peak-tracking parameter corresponding generally to an upper envelope of the sequence of said energy values of said frames.
Claim 7
Independent Claim
27 words
The system of either of Claims 1 or 2, wherein said speech parameters of each of said frame also indicate the voiced/unvoiced status of said respective frame.
Claim 8
Dependent Claim
46 words
The system of Claim 5, wherein said parameters also include pitch information for each of said speech frames, and wherein said analyzer jointly determines pitch and voicing of each frame, so that the said pitch and voicing decisions vary as smoothly as possible across adjacent frames.
Claim 9
Independent Claim
73 words
A method of encoding speech, comprising the steps of: Analyzing a speech signal to provide a sequence of frames as speech parameters, each said frame of said sequence of parameters including an energy value; normalizing said energy values of each of said speech frames with respect to energy values of subsequent ones of said speech frames; and encoding said speech parameters including said normalized ones of said energy values into a data channel.
Claim 10
Independent Claim
73 words
A method of encoding speech, comprising the steps of: Analyzing a speech signal to provide a sequence of frames as speech parameters, each said frame of said sequence of parameters including an energy value; normalizing said energy values of each of said speech frames with respect to energy values of subsequent ones of said speech frames; and encoding said speech parameters including said normalized ones of said energy values into a data channel.
Claim 11
Dependent Claim
41 words
The method of Claim 7, wherein said energy value of each said speech frame is normalized with respect to said energy value of only those of said frames which are later than each said respective frame by at least 0.1 second.
Claim 12
Independent Claim
49 words
The method of either of Claims 7 or 8, wherein said energy value of each said speech frame is normalized with respect to a peak-tracking parameter of said subsequent frames, said peak-tracking parameter corresponding generally to an upper envelope of the sequence of said energy values of said frame.
Claim 13
Dependent Claim
184 words
A speech coding system, comprising: an analyzer connected to receive speech input data and to generate therefrom a sequence of frames of speech parameters, said frames being provided at a predetermined frame rate, said frames each comprising plural parameters including an energy value; an encoder, for encoding said successive speech frames as digital values; silence suppression means, connected to said encoding means, said silence suppression means preventing said encoder from encoding ones of said sequence of frames which do not correspond to an actual speech signal; wherein said silence suppression means identifies each said frame as silent or nonsilent by steps which include comparing the energy value of each successive one of said frames against a function of first and second adaptively updated threshold values, said first adaptively updated threshold value corresponding to a multiple of an upper envelope of said successive energy values of successive ones of said frames and said second threshold value corresponding to a multiple of a lower envelope of said successive updated values of said frames; and output means for loading said encoded digital values into a data channel.
Claim 14
Dependent Claim
184 words
A speech coding system, comprising: an analyzer connected to receive speech input data and to generate therefrom a sequence of frames of speech parameters, said frames being provided at a predetermined frame rate, said frames each comprising plural parameters including an energy value; an encoder, for encoding said successive speech frames as digital values; silence suppression means, connected to said encoding means, said silence suppression means preventing said encoder from encoding ones of said sequence of frames which do not correspond to an actual speech signal; wherein said silence suppression means identifies each said frame as silent or nonsilent by steps which include comparing the energy value of each successive one of said frames against a function of first and second adaptively updated threshold values, said first adaptively updated threshold value corresponding to a multiple of an upper envelope of said successive energy values of successive ones of said frames and said second threshold value corresponding to a multiple of a lower envelope of said successive updated values of said frames; and output means for loading said encoded digital values into a data channel.
Claim 15
Dependent Claim
183 words
A speech coding system, comprising: An analyzer connected to receive a digital speech signal and to generate therefrom a sequence of frames of speech parameters, said parameters of each frame including an energy value; means for normalizing the energy value of each said speech frame with respect to energy values of subsequent frames; silence suppression means, connected to said encoding means, said silence suppression means preventing said encoder from encoding ones of said sequence of frames which do not correspond to an actual speech signal; wherein said silence supression means identifies each said frame as silent or nonsilent by steps which include comparing the energy value of each successive one of said frames against a function of first and second adaptively updated threshold values, said first adaptively updated threshold value corresponding to a multiple of an upper envelope of said successive energy values of successive ones of said frames and said second threshold value corresponding to a multiple of a lower envelope of said successive updated values of said frames; and output means for loading said encoded digital values into a data channel.
Claim 16
Dependent Claim
183 words
A speech coding system, comprising: An analyzer connected to receive a digital speech signal and to generate therefrom a sequence of frames of speech parameters, said parameters of each frame including an energy value; means for normalizing the energy value of each said speech frame with respect to energy values of subsequent frames; silence suppression means, connected to said encoding means, said silence suppression means preventing said encoder from encoding ones of said sequence of frames which do not correspond to an actual speech signal; wherein said silence supression means identifies each said frame as silent or nonsilent by steps which include comparing the energy value of each successive one of said frames against a function of first and second adaptively updated threshold values, said first adaptively updated threshold value corresponding to a multiple of an upper envelope of said successive energy values of successive ones of said frames and said second threshold value corresponding to a multiple of a lower envelope of said successive updated values of said frames; and output means for loading said encoded digital values into a data channel.
Claim 17
Dependent Claim
275 words
A voice mail system, comprising: An analyzer connected to receive a digital speech signal and to generate therefrom a sequence of frames of speech parameters, said parameters at each frame including an energy value; means for normalizing the energy value of each said speech frame with respect to energy values of subsequent frames; silence suppression means, connected to said encoding means, said silence suppression means preventing said encoder from encoding ones of said sequence of frames which do not correspond to an actual speech signal; wherein said silence suppression means identifies each said frame as silent or nonsilent by steps which include comparing the energy value of each successive one of said frames against a function of first and second adaptively updated threshold values, said first adaptively updated threshold value corresponding to a multiple of an upper envelope of said successive energy values of successive ones of said frames and said second threshold value corresponding to a multiple of a lower envelope of said successive updated values of said frames; and output means for loading said parameters for each said speech frame including said normalized energy parameter of each said speech frame into a data channel; input means for receiving a sequence of frames of speech parameters, said parameters including linear predictive coding parameters and excitation prarmeters; means for configuring a lattice filter in accordance with said linear predictive coding parameters; means for generating an excitation signal in accordance with said excitation parameters, said excitation signal being provided as input to said lattice filter; and means for modulating the output of said lattice filter-in accordance with an energy parameter, to provide a speech signal output.
Claim 18
Dependent Claim
275 words
A voice mail system, comprising: An analyzer connected to receive a digital speech signal and to generate therefrom a sequence of frames of speech parameters, said parameters at each frame including an energy value; means for normalizing the energy value of each said speech frame with respect to energy values of subsequent frames; silence suppression means, connected to said encoding means, said silence suppression means preventing said encoder from encoding ones of said sequence of frames which do not correspond to an actual speech signal; wherein said silence suppression means identifies each said frame as silent or nonsilent by steps which include comparing the energy value of each successive one of said frames against a function of first and second adaptively updated threshold values, said first adaptively updated threshold value corresponding to a multiple of an upper envelope of said successive energy values of successive ones of said frames and said second threshold value corresponding to a multiple of a lower envelope of said successive updated values of said frames; and output means for loading said parameters for each said speech frame including said normalized energy parameter of each said speech frame into a data channel; input means for receiving a sequence of frames of speech parameters, said parameters including linear predictive coding parameters and excitation prarmeters; means for configuring a lattice filter in accordance with said linear predictive coding parameters; means for generating an excitation signal in accordance with said excitation parameters, said excitation signal being provided as input to said lattice filter; and means for modulating the output of said lattice filter-in accordance with an energy parameter, to provide a speech signal output.
Claim 19
Independent Claim
51 words
The system of either of Claims 10 or 11, wherein said analyzer provides a voicing decision for each of said speech frames, and wherein said silence suppression means updates said first threshold only during voiced ones of said frames and updates said second threshold only during unvoiced ones of said frames.
Claim 20
Independent Claim
80 words
The system of either of Claims 10 or 11, wherein said silence suppression means, once a silent frame has been identified, does not identify a nonsilent frame thereafter until a voiced super-threshold-energy frame is detected, in which case said voiced super-threshold-energy frame and all preceding unvoiced super-threshold-energy speech frames which are not separated from said super-threshold-energy voiced frame by at least a predetermined number of successive frames each having an energy level below said threshold level, are identified as nonsilent.
Claim 21
Independent Claim
41 words
The system of either of Claims 10 or 11, wherein said silence suppression means, once a nonsilent frame has been identified, identifies a silent frame only when a continuous succession of subthreshold-energy frames has been identified over a predetermined time interval.
Claim 22
Independent Claim
20 words
The system of either of Claims 14 or 15, wherein said predetermined time interval is between 0.2 and 0.8 seconds.
Claim 23
Independent Claim
44 words
The system of either of Claims 11 or 12, wherein said energy value of each said speech frame is normalized with respect to said energy values primarily of those of said frames which are later than said respective frame by at least 0.1 second.
Claim 24
Independent Claim
50 words
The system of any of Claims 11, 12 and 17, wherein said energy value of each said speech frame is normalized with respect to a peak-tracking parameter of said subsequent frames, said peak-tracking parameter corresponding generally to an upper envelope of the sequence of said energy values of said frames.
Claim 25
Dependent Claim
43 words
The system of Claim 15, wherein said silence suppression means, once a nonsilent frame has been identified, identifies a silent frame only if said continuous succession of subthreshold energy frames over said predetermined time interval is found after a voiced subthreshold energy frame.
References Cited (Backward Citations)
FR2308248A1
Cited
US4071695A
Cited
US4280192A
Cited
FR2380612A1
Cited
US4351983A
Cited
FR2451680A1
Cited
EP0027066A1
Cited
EP0047589A1
Cited
USPTO Annuity & Maintenance Windows
Maintenance schedule is calculated for granted US utility patents.