Dual-band speech encoding
View Patent ↗This document describes various techniques for dual-band speech encoding. In some embodiments, a first type of speech feature is received from a remote entity, an estimate of a second type of speech feature is determined based on the first type of speech feature, the estimate of the second type of speech feature is provided to a speech recognizer, speech-recognition results based on the estimate of the second type of speech feature are received from the speech recognizer, and the speech-recognition results are transmitted to the remote entity.
1. A method comprising:
determining, based on one speech waveform, a wideband speech feature and a narrowband speech feature;
determining, based on the wideband speech feature, an estimate of the narrowband speech feature;
determining, based on the narrowband speech feature and the estimate of the narrowband speech feature, an estimation error of the estimate of the narrowband speech feature;
transmitting the wideband speech feature and the estimation error to a remote entity; and
receiving, from the remote entity, data associated with a speech-based service based on the wideband speech feature.
2. The method as described in claim 1 , further comprising determining if sufficient bandwidth is available to transmit the estimation error of the narrowband speech feature and, when the sufficient bandwidth is available, transmitting the estimation error of the narrowband speech feature estimate.
3. The method as recited in claim 1 , further comprising encoding the wideband speech feature or the estimation error of the narrowband speech feature estimate using codebook-free encoding.
4. The method as recited in claim 1 , wherein determining the narrowband speech feature estimate uses an affine transform.
5. The method as recited in claim 1 , further comprising encoding the wideband speech feature or the estimation error of the narrowband speech feature estimate using adaptive differential pulse-code modulation.
6. The method as recited in claim 1 , further comprising encoding the wideband speech feature with dynamic mean normalization.
7. A method comprising:
receiving a wideband speech feature from a remote entity;
determining an estimate of a narrowband speech feature based on the wideband speech feature;
providing the estimate of the narrowband speech feature to a speech recognizer trained on the narrowband speech features;
receiving, from the speech recognizer, speech-recognition results based on the estimate of the narrowband speech feature; and
transmitting, to the remote entity, the speech-recognition results based on the estimate of the narrowband speech feature.
8. The method as recited in claim 7 , wherein the wideband speech feature is encoded with adaptive differential pulse-code modulation and further comprising decoding the wideband speech feature.
9. The method as recited in claim 7 , wherein determining the estimate of the narrowband speech feature uses an affine transform.
10. The method as recited in claim 9 , wherein parameters of the affine transform are based on a parallel set of the wideband speech features and the narrowband speech features.
11. The method as recited in claim 7 , further comprising providing the speech-recognition results of the speech recognizer trained on narrowband speech features to a search engine, receiving search results from the search engine, and transmitting the search results to the remote entity.
12. The method as recited in claim 7 , further comprising storing the wideband speech feature for training a second speech recognizer based on wideband speech features.
13. A method comprising:
receiving, from a remote entity, a wideband speech feature and an estimation error of a narrowband speech feature;
determining an estimate of a narrowband speech feature based on the wideband speech feature and the estimation error of the narrowband speech feature;
providing the estimate of the narrowband speech feature to a speech recognizer trained on narrowband speech features;
receiving, from the speech recognizer trained on narrowband speech features, speech-recognition results based on the estimate of the narrowband speech feature; and
transmitting, to the remote entity, the speech-recognition results based on the estimate of the narrowband speech feature.
14. The method as recited in claim 13 , wherein determining the estimate of the narrowband speech feature uses an affine transform.
15. The method as recited in claim 14 , wherein parameters of the affine transform are based on a parallel set of the wideband speech features and the narrowband speech features.
16. The method as recited in claim 14 , wherein the parameters of the affine transform are based on a minimum mean squared error model.
17. The method as recited in claim 13 , wherein determining the estimate of the narrowband speech feature uses a pseudo-inverse derivation model.
18. The method as recited in claim 13 , further comprising providing the speech-recognition results of the speech recognizer trained on the narrowband speech feature to a search engine, receiving, from the search engine, search results based on the speech-recognition results, and transmitting the search results to the remote entity.
19. The method as recited in claim 13 , further comprising storing the wideband speech feature for training a second speech recognizer based on the wideband speech feature.