Device for normalizing voice pitch for voice recognition
AssigneeMatsushita Electric Industrial Co., Ltd.
InventorsMikio Oda, Tomoe Kawane
Life of this patent
Estimating the term…
Abstract
In a voice pitch normalization device equipped in a voice recognition device VRAp for recognizing an incoming command voice Sva uttered by any speaker, and used to normalize the incoming command voice to be in an optimal pitch for voice recognition, a target voice generator produces a target voice signal by changing the incoming command voice Svd on the basis of a predetermined degree. A probability calculator calculates a probability indicating a degree of coincidence among the target voice signal and a plurality of words in sample data. A voice pitch changer repeatedly changes the target voice signal in voice pitch until a maximum probability becomes a predetermined probability or greater.
Claims12
A voice pitch normalization apparatus for recognizing an incoming command voice uttered by a speaker based on sample data for a plurality of words, said voice pitch normalization apparatus comprising:
- a voice analyzer operable to calculate a probability indicating a degree of coincidence among a target voice signal and each of the plurality of words in the sample data;
- and
- a voice pitch normalization device operable to generate the target voice signal by changing the incoming command voice on a predetermined degree basis and operable to change the target voice signal in voice pitch until a maximum of the probabilities reaches a predetermined probability or higher,
- wherein said voice pitch normalization device is operable to increase or decrease the target voice signal on the predetermined degree basis when the maximum of the probabilities is smaller than the predetermined probability,
- wherein said voice pitch normalization device is operable to increase the target voice signal in voice pitch on the predetermined degree basis starting from a pitch level of the incoming command voice, and
- wherein the target voice signal is limited in voice pitch up to a first predetermined pitch, and when the maximum of the probabilities fails to reach the predetermined probability or higher before the target voice signal reaches the first predetermined pitch, said voice pitch normalization device decreases the target voice signal in voice pitch on the predetermined degree basis starting from the pitch level of the incoming command voice.
A voice pitch normalization apparatus for recognizing an incoming command voice uttered by a speaker based on sample data for a plurality of words, said voice pitch normalization apparatus comprising:
- a voice analyzer operable to calculate a probability indicating a degree of coincidence among a target voice signal and each of the plurality of words in the sample data;
- and
- a voice pitch normalization device operable to generate the target voice signal by changing the incoming command voice on a predetermined degree basis and operable to change the target voice signal in voice pitch until a maximum of the probabilities reaches a predetermined probability or higher,
- wherein said voice pitch normalization device is operable to increase or decrease the target voice signal on the predetermined degree basis when the maximum of the probabilities is smaller than the predetermined probability,
- wherein said voice pitch normalization device is operable to decrease the target voice signal in voice pitch on the predetermined degree basis starting from a pitch level of the incoming command voice, and
- wherein the target voice signal is limited in voice pitch down to a third predetermined pitch, and when the maximum of the probabilities fails to reach the predetermined probability or higher before the target voice signal reaches the third predetermined pitch, said voice pitch normalization device is operable to increase the target voice signal in voice pitch on the predetermined degree basis starting from the pitch level of the incoming command voice.
A voice recognition device for recognizing an incoming command voice optimally normalized for voice recognition based on sample data for a plurality of words, said voice recognition device comprising:
- a voice analyzer operable to calculate a probability indicating a degree of coincidence among a target voice signal and each of the plurality of words in the sample data;
- and
- a voice pitch normalization device operable to generate the target voice signal by changing the incoming command voice on a predetermined degree basis and changing the target voice signal in voice pitch until a maximum of the probabilities reaches a predetermined probability or higher,
- wherein said voice pitch normalization device is operable to increase or decrease the target voice signal on the predetermined degree basis when the maximum of the probabilities is smaller than the predetermined probability,
- wherein said voice pitch normalization device is operable to increase the target voice signal in voice pitch on the predetermined degree basis starting from a pitch level of the incoming command voice, and
- wherein the target voice signal is limited in voice pitch up to a first predetermined pitch, and when the maximum of the probabilities fails to reach the predetermined probability or higher before the target voice signal reaches the first predetermined pitch, said voice pitch normalization device is operable to decrease the target voice signal in voice pitch on the predetermined degree basis starting from the pitch level of the incoming command voice.
A voice recognition device for recognizing an incoming command voice optimally normalized for voice recognition based on sample data for a plurality of words, said voice recognition device comprising:
- a voice analyzer operable to calculate a probability indicating a degree of coincidence among a target voice signal and each of the plurality of words in the sample data;
- and
- a voice pitch normalization device operable to generate the target voice signal by changing the incoming command voice on a predetermined degree basis and changing the target voice signal in voice pitch until a maximum of the probabilities reaches a predetermined probability or higher,
- wherein said voice pitch normalization device is operable to increase or decrease the target voice signal on the predetermined degree basis when the maximum of the probabilities is smaller than the predetermined probability,
- wherein said voice pitch normalization device is operable to decrease the target voice signal in voice pitch on the predetermined degree basis started from a pitch level of the incoming command voice, and
- wherein the target voice signal is limited in voice pitch down to a third predetermined pitch, and when the maximum of the probabilities fails to reach the predetermined probability or higher before the target voice signal reaches the third predetermined pitch, said voice pitch normalization device is operable to increase the target voice signal in voice pitch on the predetermined degree basis starting from the pitch level of the incoming command voice.
A voice pitch normalization method for recognizing an incoming command voice uttered by a speaker based on sample data for a plurality of words, said voice pitch normalization method comprising:
- generating a target voice signal by changing the incoming command voice on a predetermined degree basis;
- calculating a probability indicating a degree of coincidence among the target voice signal and each of the plurality of words in the sample data;
- changing the target voice signal in voice pitch until a maximum of the probabilities reaches a predetermined probability or higher;
- when the maximum of the probabilities is smaller than the predetermined probability, one of increasing and decreasing the target voice signal on the predetermined degree basis;
- and
- increasing the target voice signal in voice pitch on the predetermined degree basis starting from a pitch level of the incoming command voice,
- wherein the target voice signal is limited in voice pitch up to a first predetermined pitch, and
- said voice pitch normalization method further comprises, when the maximum of the probabilities fails to reach the predetermined probability or higher before the target voice signal reaches the first predetermined pitch, decreasing the target voice signal in voice pitch on the predetermined degree basis starting from the pitch level of the incoming command voice.
The voice pitch normalization method for recognizing an incoming command voice uttered by a speaker based on sample data for a plurality of words, said voice pitch normalization method comprising:
- generating a target voice signal by changing the incoming command voice on a predetermined degree basis;
- calculating a probability indicating a degree of coincidence among the target voice signal and each of the plurality of words in the sample data;
- changing the target voice signal in voice pitch until a maximum of the probabilities reaches a predetermined probability or higher;
- when the maximum of the probabilities is smaller than the predetermined probability, one of increasing and decreasing the target voice signal on the predetermined degree basis;
- and
- decreasing the target voice signal in voice pitch on the predetermined degree basis starting from a pitch level of the incoming command voice,
- wherein the target voice signal is limited in voice pitch down to a third predetermined pitch, and
- said voice pitch normalization method further comprises, when the maximum of the probabilities fails to reach the predetermined probability or higher before the target voice signal reaches the third predetermined pitch, increasing the target voice signal in voice pitch on the predetermined degree basis starting from the pitch level of the incoming command voice.
Dependencies are inferred from the claim text.